{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9230059226213368,
  "eval_steps": 1000,
  "global_step": 12000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 7.691716021844474e-05,
      "grad_norm": 193.0,
      "learning_rate": 2.9998846242596723e-06,
      "loss": 5.4644,
      "mean_token_accuracy": 0.47491639852523804,
      "num_tokens": 300.0,
      "step": 1
    },
    {
      "epoch": 0.00015383432043688947,
      "grad_norm": 116.0,
      "learning_rate": 2.999769248519345e-06,
      "loss": 3.8241,
      "mean_token_accuracy": 0.55064457654953,
      "num_tokens": 844.0,
      "step": 2
    },
    {
      "epoch": 0.0002307514806553342,
      "grad_norm": 164.0,
      "learning_rate": 2.999653872779017e-06,
      "loss": 4.6813,
      "mean_token_accuracy": 0.5,
      "num_tokens": 1176.0,
      "step": 3
    },
    {
      "epoch": 0.00030766864087377894,
      "grad_norm": 160.0,
      "learning_rate": 2.9995384970386895e-06,
      "loss": 5.1307,
      "mean_token_accuracy": 0.4816053509712219,
      "num_tokens": 1476.0,
      "step": 4
    },
    {
      "epoch": 0.00038458580109222365,
      "grad_norm": 334.0,
      "learning_rate": 2.999423121298362e-06,
      "loss": 5.2475,
      "mean_token_accuracy": 0.4615384638309479,
      "num_tokens": 1764.0,
      "step": 5
    },
    {
      "epoch": 0.0004615029613106684,
      "grad_norm": 167.0,
      "learning_rate": 2.999307745558034e-06,
      "loss": 4.4516,
      "mean_token_accuracy": 0.5275080800056458,
      "num_tokens": 2074.0,
      "step": 6
    },
    {
      "epoch": 0.0005384201215291131,
      "grad_norm": 300.0,
      "learning_rate": 2.9991923698177063e-06,
      "loss": 4.5944,
      "mean_token_accuracy": 0.5302491188049316,
      "num_tokens": 2356.0,
      "step": 7
    },
    {
      "epoch": 0.0006153372817475579,
      "grad_norm": 118.0,
      "learning_rate": 2.999076994077379e-06,
      "loss": 3.1606,
      "mean_token_accuracy": 0.5819819569587708,
      "num_tokens": 2912.0,
      "step": 8
    },
    {
      "epoch": 0.0006922544419660026,
      "grad_norm": 106.0,
      "learning_rate": 2.998961618337051e-06,
      "loss": 4.0391,
      "mean_token_accuracy": 0.5351170301437378,
      "num_tokens": 3214.0,
      "step": 9
    },
    {
      "epoch": 0.0007691716021844473,
      "grad_norm": 92.0,
      "learning_rate": 2.9988462425967235e-06,
      "loss": 2.6016,
      "mean_token_accuracy": 0.6216545104980469,
      "num_tokens": 4038.0,
      "step": 10
    },
    {
      "epoch": 0.0008460887624028921,
      "grad_norm": 157.0,
      "learning_rate": 2.9987308668563958e-06,
      "loss": 4.265,
      "mean_token_accuracy": 0.5675675868988037,
      "num_tokens": 4298.0,
      "step": 11
    },
    {
      "epoch": 0.0009230059226213368,
      "grad_norm": 71.0,
      "learning_rate": 2.998615491116068e-06,
      "loss": 3.2769,
      "mean_token_accuracy": 0.589385449886322,
      "num_tokens": 4658.0,
      "step": 12
    },
    {
      "epoch": 0.0009999230828397815,
      "grad_norm": 96.5,
      "learning_rate": 2.9985001153757407e-06,
      "loss": 3.8583,
      "mean_token_accuracy": 0.5953177213668823,
      "num_tokens": 4958.0,
      "step": 13
    },
    {
      "epoch": 0.0010768402430582262,
      "grad_norm": 70.0,
      "learning_rate": 2.9983847396354125e-06,
      "loss": 3.5853,
      "mean_token_accuracy": 0.5681818127632141,
      "num_tokens": 5268.0,
      "step": 14
    },
    {
      "epoch": 0.001153757403276671,
      "grad_norm": 89.0,
      "learning_rate": 2.998269363895085e-06,
      "loss": 3.5773,
      "mean_token_accuracy": 0.577049195766449,
      "num_tokens": 5574.0,
      "step": 15
    },
    {
      "epoch": 0.0012306745634951158,
      "grad_norm": 74.0,
      "learning_rate": 2.9981539881547575e-06,
      "loss": 3.4034,
      "mean_token_accuracy": 0.5653495192527771,
      "num_tokens": 5904.0,
      "step": 16
    },
    {
      "epoch": 0.0013075917237135605,
      "grad_norm": 75.5,
      "learning_rate": 2.9980386124144297e-06,
      "loss": 2.6048,
      "mean_token_accuracy": 0.6387434601783752,
      "num_tokens": 6478.0,
      "step": 17
    },
    {
      "epoch": 0.0013845088839320053,
      "grad_norm": 70.5,
      "learning_rate": 2.997923236674102e-06,
      "loss": 3.2002,
      "mean_token_accuracy": 0.5954415798187256,
      "num_tokens": 6830.0,
      "step": 18
    },
    {
      "epoch": 0.00146142604415045,
      "grad_norm": 52.75,
      "learning_rate": 2.9978078609337743e-06,
      "loss": 2.5099,
      "mean_token_accuracy": 0.6377079486846924,
      "num_tokens": 7372.0,
      "step": 19
    },
    {
      "epoch": 0.0015383432043688946,
      "grad_norm": 95.5,
      "learning_rate": 2.9976924851934465e-06,
      "loss": 3.3634,
      "mean_token_accuracy": 0.6420664191246033,
      "num_tokens": 7644.0,
      "step": 20
    },
    {
      "epoch": 0.0016152603645873394,
      "grad_norm": 56.5,
      "learning_rate": 2.997577109453119e-06,
      "loss": 3.0766,
      "mean_token_accuracy": 0.6302250623703003,
      "num_tokens": 7956.0,
      "step": 21
    },
    {
      "epoch": 0.0016921775248057841,
      "grad_norm": 49.0,
      "learning_rate": 2.9974617337127914e-06,
      "loss": 2.8693,
      "mean_token_accuracy": 0.6246246099472046,
      "num_tokens": 8290.0,
      "step": 22
    },
    {
      "epoch": 0.0017690946850242289,
      "grad_norm": 50.0,
      "learning_rate": 2.9973463579724637e-06,
      "loss": 2.7224,
      "mean_token_accuracy": 0.6697819232940674,
      "num_tokens": 8612.0,
      "step": 23
    },
    {
      "epoch": 0.0018460118452426736,
      "grad_norm": 42.25,
      "learning_rate": 2.9972309822321364e-06,
      "loss": 2.2681,
      "mean_token_accuracy": 0.6654343605041504,
      "num_tokens": 9154.0,
      "step": 24
    },
    {
      "epoch": 0.0019229290054611184,
      "grad_norm": 48.75,
      "learning_rate": 2.9971156064918082e-06,
      "loss": 2.3757,
      "mean_token_accuracy": 0.7151514887809753,
      "num_tokens": 9486.0,
      "step": 25
    },
    {
      "epoch": 0.001999846165679563,
      "grad_norm": 61.0,
      "learning_rate": 2.9970002307514805e-06,
      "loss": 3.0606,
      "mean_token_accuracy": 0.6653386354446411,
      "num_tokens": 9738.0,
      "step": 26
    },
    {
      "epoch": 0.002076763325898008,
      "grad_norm": 41.5,
      "learning_rate": 2.996884855011153e-06,
      "loss": 2.1466,
      "mean_token_accuracy": 0.6944937705993652,
      "num_tokens": 10302.0,
      "step": 27
    },
    {
      "epoch": 0.0021536804861164525,
      "grad_norm": 40.0,
      "learning_rate": 2.9967694792708254e-06,
      "loss": 2.2604,
      "mean_token_accuracy": 0.6949495077133179,
      "num_tokens": 10798.0,
      "step": 28
    },
    {
      "epoch": 0.0022305976463348975,
      "grad_norm": 43.75,
      "learning_rate": 2.9966541035304977e-06,
      "loss": 2.4389,
      "mean_token_accuracy": 0.6850152611732483,
      "num_tokens": 11126.0,
      "step": 29
    },
    {
      "epoch": 0.002307514806553342,
      "grad_norm": 36.25,
      "learning_rate": 2.99653872779017e-06,
      "loss": 2.1034,
      "mean_token_accuracy": 0.6873857378959656,
      "num_tokens": 11674.0,
      "step": 30
    },
    {
      "epoch": 0.002384431966771787,
      "grad_norm": 48.0,
      "learning_rate": 2.996423352049842e-06,
      "loss": 2.4718,
      "mean_token_accuracy": 0.6842105388641357,
      "num_tokens": 12018.0,
      "step": 31
    },
    {
      "epoch": 0.0024613491269902315,
      "grad_norm": 44.5,
      "learning_rate": 2.996307976309515e-06,
      "loss": 2.3089,
      "mean_token_accuracy": 0.715976357460022,
      "num_tokens": 12358.0,
      "step": 32
    },
    {
      "epoch": 0.002538266287208676,
      "grad_norm": 52.0,
      "learning_rate": 2.996192600569187e-06,
      "loss": 2.6736,
      "mean_token_accuracy": 0.6996047496795654,
      "num_tokens": 12612.0,
      "step": 33
    },
    {
      "epoch": 0.002615183447427121,
      "grad_norm": 50.5,
      "learning_rate": 2.9960772248288594e-06,
      "loss": 2.4471,
      "mean_token_accuracy": 0.6861313581466675,
      "num_tokens": 12888.0,
      "step": 34
    },
    {
      "epoch": 0.0026921006076455656,
      "grad_norm": 40.0,
      "learning_rate": 2.995961849088532e-06,
      "loss": 2.1065,
      "mean_token_accuracy": 0.7570977807044983,
      "num_tokens": 13206.0,
      "step": 35
    },
    {
      "epoch": 0.0027690177678640106,
      "grad_norm": 33.25,
      "learning_rate": 2.995846473348204e-06,
      "loss": 1.9106,
      "mean_token_accuracy": 0.7246654033660889,
      "num_tokens": 13730.0,
      "step": 36
    },
    {
      "epoch": 0.002845934928082455,
      "grad_norm": 30.875,
      "learning_rate": 2.995731097607876e-06,
      "loss": 1.6773,
      "mean_token_accuracy": 0.7216748595237732,
      "num_tokens": 14544.0,
      "step": 37
    },
    {
      "epoch": 0.0029228520883009,
      "grad_norm": 30.875,
      "learning_rate": 2.995615721867549e-06,
      "loss": 1.667,
      "mean_token_accuracy": 0.7341269850730896,
      "num_tokens": 15302.0,
      "step": 38
    },
    {
      "epoch": 0.0029997692485193446,
      "grad_norm": 30.0,
      "learning_rate": 2.995500346127221e-06,
      "loss": 1.5991,
      "mean_token_accuracy": 0.7313624620437622,
      "num_tokens": 16082.0,
      "step": 39
    },
    {
      "epoch": 0.003076686408737789,
      "grad_norm": 32.5,
      "learning_rate": 2.9953849703868934e-06,
      "loss": 1.5239,
      "mean_token_accuracy": 0.7475728392601013,
      "num_tokens": 16908.0,
      "step": 40
    },
    {
      "epoch": 0.003153603568956234,
      "grad_norm": 36.25,
      "learning_rate": 2.9952695946465656e-06,
      "loss": 1.8676,
      "mean_token_accuracy": 0.7784430980682373,
      "num_tokens": 17244.0,
      "step": 41
    },
    {
      "epoch": 0.0032305207291746787,
      "grad_norm": 46.5,
      "learning_rate": 2.995154218906238e-06,
      "loss": 1.8907,
      "mean_token_accuracy": 0.7737226486206055,
      "num_tokens": 17520.0,
      "step": 42
    },
    {
      "epoch": 0.0033074378893931237,
      "grad_norm": 49.5,
      "learning_rate": 2.9950388431659106e-06,
      "loss": 2.0743,
      "mean_token_accuracy": 0.7050359845161438,
      "num_tokens": 17800.0,
      "step": 43
    },
    {
      "epoch": 0.0033843550496115682,
      "grad_norm": 45.25,
      "learning_rate": 2.994923467425583e-06,
      "loss": 1.6687,
      "mean_token_accuracy": 0.7951807379722595,
      "num_tokens": 18134.0,
      "step": 44
    },
    {
      "epoch": 0.0034612722098300132,
      "grad_norm": 38.0,
      "learning_rate": 2.994808091685255e-06,
      "loss": 1.4949,
      "mean_token_accuracy": 0.7614678740501404,
      "num_tokens": 18680.0,
      "step": 45
    },
    {
      "epoch": 0.0035381893700484578,
      "grad_norm": 40.0,
      "learning_rate": 2.9946927159449278e-06,
      "loss": 1.8658,
      "mean_token_accuracy": 0.717717707157135,
      "num_tokens": 19014.0,
      "step": 46
    },
    {
      "epoch": 0.0036151065302669027,
      "grad_norm": 48.75,
      "learning_rate": 2.9945773402045996e-06,
      "loss": 1.6934,
      "mean_token_accuracy": 0.773809552192688,
      "num_tokens": 19352.0,
      "step": 47
    },
    {
      "epoch": 0.0036920236904853473,
      "grad_norm": 44.0,
      "learning_rate": 2.994461964464272e-06,
      "loss": 1.4335,
      "mean_token_accuracy": 0.7425044178962708,
      "num_tokens": 19920.0,
      "step": 48
    },
    {
      "epoch": 0.003768940850703792,
      "grad_norm": 54.75,
      "learning_rate": 2.9943465887239445e-06,
      "loss": 1.8076,
      "mean_token_accuracy": 0.7728813290596008,
      "num_tokens": 20216.0,
      "step": 49
    },
    {
      "epoch": 0.003845858010922237,
      "grad_norm": 48.5,
      "learning_rate": 2.994231212983617e-06,
      "loss": 1.5646,
      "mean_token_accuracy": 0.7846607565879822,
      "num_tokens": 20558.0,
      "step": 50
    },
    {
      "epoch": 0.003922775171140681,
      "grad_norm": 53.75,
      "learning_rate": 2.994115837243289e-06,
      "loss": 1.674,
      "mean_token_accuracy": 0.7798165082931519,
      "num_tokens": 20886.0,
      "step": 51
    },
    {
      "epoch": 0.003999692331359126,
      "grad_norm": 48.0,
      "learning_rate": 2.9940004615029613e-06,
      "loss": 1.2712,
      "mean_token_accuracy": 0.7846153974533081,
      "num_tokens": 21472.0,
      "step": 52
    },
    {
      "epoch": 0.004076609491577571,
      "grad_norm": 46.0,
      "learning_rate": 2.9938850857626336e-06,
      "loss": 1.499,
      "mean_token_accuracy": 0.7841726541519165,
      "num_tokens": 21752.0,
      "step": 53
    },
    {
      "epoch": 0.004153526651796016,
      "grad_norm": 35.0,
      "learning_rate": 2.9937697100223062e-06,
      "loss": 1.4662,
      "mean_token_accuracy": 0.8037382960319519,
      "num_tokens": 22074.0,
      "step": 54
    },
    {
      "epoch": 0.00423044381201446,
      "grad_norm": 35.5,
      "learning_rate": 2.9936543342819785e-06,
      "loss": 1.4637,
      "mean_token_accuracy": 0.7928994297981262,
      "num_tokens": 22414.0,
      "step": 55
    },
    {
      "epoch": 0.004307360972232905,
      "grad_norm": 28.0,
      "learning_rate": 2.9935389585416508e-06,
      "loss": 1.1755,
      "mean_token_accuracy": 0.7884615659713745,
      "num_tokens": 23196.0,
      "step": 56
    },
    {
      "epoch": 0.0043842781324513495,
      "grad_norm": 35.0,
      "learning_rate": 2.993423582801323e-06,
      "loss": 1.2011,
      "mean_token_accuracy": 0.803056001663208,
      "num_tokens": 23786.0,
      "step": 57
    },
    {
      "epoch": 0.004461195292669795,
      "grad_norm": 33.75,
      "learning_rate": 2.9933082070609953e-06,
      "loss": 1.2322,
      "mean_token_accuracy": 0.7984644770622253,
      "num_tokens": 24308.0,
      "step": 58
    },
    {
      "epoch": 0.0045381124528882395,
      "grad_norm": 60.0,
      "learning_rate": 2.9931928313206675e-06,
      "loss": 1.392,
      "mean_token_accuracy": 0.8080495595932007,
      "num_tokens": 24632.0,
      "step": 59
    },
    {
      "epoch": 0.004615029613106684,
      "grad_norm": 21.125,
      "learning_rate": 2.9930774555803402e-06,
      "loss": 1.0966,
      "mean_token_accuracy": 0.8101933002471924,
      "num_tokens": 25202.0,
      "step": 60
    },
    {
      "epoch": 0.0046919467733251286,
      "grad_norm": 43.75,
      "learning_rate": 2.9929620798400125e-06,
      "loss": 1.1938,
      "mean_token_accuracy": 0.840624988079071,
      "num_tokens": 25526.0,
      "step": 61
    },
    {
      "epoch": 0.004768863933543574,
      "grad_norm": 33.25,
      "learning_rate": 2.9928467040996847e-06,
      "loss": 1.2119,
      "mean_token_accuracy": 0.8318583965301514,
      "num_tokens": 25866.0,
      "step": 62
    },
    {
      "epoch": 0.0048457810937620185,
      "grad_norm": 20.625,
      "learning_rate": 2.992731328359357e-06,
      "loss": 1.0683,
      "mean_token_accuracy": 0.8288288116455078,
      "num_tokens": 26422.0,
      "step": 63
    },
    {
      "epoch": 0.004922698253980463,
      "grad_norm": 40.0,
      "learning_rate": 2.9926159526190293e-06,
      "loss": 1.0679,
      "mean_token_accuracy": 0.8494983315467834,
      "num_tokens": 26724.0,
      "step": 64
    },
    {
      "epoch": 0.004999615414198908,
      "grad_norm": 19.25,
      "learning_rate": 2.9925005768787015e-06,
      "loss": 1.0485,
      "mean_token_accuracy": 0.8370118737220764,
      "num_tokens": 27314.0,
      "step": 65
    },
    {
      "epoch": 0.005076532574417352,
      "grad_norm": 34.0,
      "learning_rate": 2.992385201138374e-06,
      "loss": 1.1364,
      "mean_token_accuracy": 0.8585858345031738,
      "num_tokens": 27614.0,
      "step": 66
    },
    {
      "epoch": 0.0051534497346357976,
      "grad_norm": 22.5,
      "learning_rate": 2.9922698253980465e-06,
      "loss": 1.0561,
      "mean_token_accuracy": 0.8323809504508972,
      "num_tokens": 28140.0,
      "step": 67
    },
    {
      "epoch": 0.005230366894854242,
      "grad_norm": 21.625,
      "learning_rate": 2.9921544496577187e-06,
      "loss": 0.8978,
      "mean_token_accuracy": 0.8465228080749512,
      "num_tokens": 28976.0,
      "step": 68
    },
    {
      "epoch": 0.005307284055072687,
      "grad_norm": 23.25,
      "learning_rate": 2.992039073917391e-06,
      "loss": 0.9425,
      "mean_token_accuracy": 0.8390501141548157,
      "num_tokens": 29736.0,
      "step": 69
    },
    {
      "epoch": 0.005384201215291131,
      "grad_norm": 34.0,
      "learning_rate": 2.9919236981770632e-06,
      "loss": 1.1814,
      "mean_token_accuracy": 0.8590164184570312,
      "num_tokens": 30042.0,
      "step": 70
    },
    {
      "epoch": 0.005461118375509577,
      "grad_norm": 57.75,
      "learning_rate": 2.991808322436736e-06,
      "loss": 1.0604,
      "mean_token_accuracy": 0.8674699068069458,
      "num_tokens": 30376.0,
      "step": 71
    },
    {
      "epoch": 0.005538035535728021,
      "grad_norm": 66.0,
      "learning_rate": 2.991692946696408e-06,
      "loss": 1.0717,
      "mean_token_accuracy": 0.8637771010398865,
      "num_tokens": 30700.0,
      "step": 72
    },
    {
      "epoch": 0.005614952695946466,
      "grad_norm": 74.0,
      "learning_rate": 2.9915775709560804e-06,
      "loss": 1.1669,
      "mean_token_accuracy": 0.8532423377037048,
      "num_tokens": 30994.0,
      "step": 73
    },
    {
      "epoch": 0.00569186985616491,
      "grad_norm": 45.5,
      "learning_rate": 2.9914621952157527e-06,
      "loss": 1.3385,
      "mean_token_accuracy": 0.830670952796936,
      "num_tokens": 31308.0,
      "step": 74
    },
    {
      "epoch": 0.005768787016383355,
      "grad_norm": 30.5,
      "learning_rate": 2.991346819475425e-06,
      "loss": 0.8953,
      "mean_token_accuracy": 0.8855218887329102,
      "num_tokens": 31608.0,
      "step": 75
    },
    {
      "epoch": 0.0058457041766018,
      "grad_norm": 22.5,
      "learning_rate": 2.991231443735097e-06,
      "loss": 0.9118,
      "mean_token_accuracy": 0.8561019897460938,
      "num_tokens": 32158.0,
      "step": 76
    },
    {
      "epoch": 0.005922621336820245,
      "grad_norm": 47.0,
      "learning_rate": 2.99111606799477e-06,
      "loss": 0.9885,
      "mean_token_accuracy": 0.8668639063835144,
      "num_tokens": 32498.0,
      "step": 77
    },
    {
      "epoch": 0.005999538497038689,
      "grad_norm": 17.75,
      "learning_rate": 2.991000692254442e-06,
      "loss": 0.8274,
      "mean_token_accuracy": 0.8747856020927429,
      "num_tokens": 33082.0,
      "step": 78
    },
    {
      "epoch": 0.006076455657257134,
      "grad_norm": 29.75,
      "learning_rate": 2.9908853165141144e-06,
      "loss": 1.0934,
      "mean_token_accuracy": 0.8674351572990417,
      "num_tokens": 33430.0,
      "step": 79
    },
    {
      "epoch": 0.006153372817475578,
      "grad_norm": 75.5,
      "learning_rate": 2.9907699407737867e-06,
      "loss": 1.158,
      "mean_token_accuracy": 0.8551236987113953,
      "num_tokens": 33714.0,
      "step": 80
    },
    {
      "epoch": 0.006230289977694024,
      "grad_norm": 33.0,
      "learning_rate": 2.990654565033459e-06,
      "loss": 0.9992,
      "mean_token_accuracy": 0.867986798286438,
      "num_tokens": 34018.0,
      "step": 81
    },
    {
      "epoch": 0.006307207137912468,
      "grad_norm": 31.25,
      "learning_rate": 2.9905391892931316e-06,
      "loss": 0.8155,
      "mean_token_accuracy": 0.9057239294052124,
      "num_tokens": 34318.0,
      "step": 82
    },
    {
      "epoch": 0.006384124298130913,
      "grad_norm": 21.125,
      "learning_rate": 2.990423813552804e-06,
      "loss": 0.7986,
      "mean_token_accuracy": 0.87613844871521,
      "num_tokens": 34868.0,
      "step": 83
    },
    {
      "epoch": 0.006461041458349357,
      "grad_norm": 20.0,
      "learning_rate": 2.990308437812476e-06,
      "loss": 0.7261,
      "mean_token_accuracy": 0.8848167657852173,
      "num_tokens": 35442.0,
      "step": 84
    },
    {
      "epoch": 0.006537958618567803,
      "grad_norm": 49.75,
      "learning_rate": 2.9901930620721484e-06,
      "loss": 0.9856,
      "mean_token_accuracy": 0.8753893971443176,
      "num_tokens": 35764.0,
      "step": 85
    },
    {
      "epoch": 0.006614875778786247,
      "grad_norm": 47.25,
      "learning_rate": 2.9900776863318206e-06,
      "loss": 0.7904,
      "mean_token_accuracy": 0.8920634984970093,
      "num_tokens": 36080.0,
      "step": 86
    },
    {
      "epoch": 0.006691792939004692,
      "grad_norm": 62.25,
      "learning_rate": 2.989962310591493e-06,
      "loss": 0.8587,
      "mean_token_accuracy": 0.8876811861991882,
      "num_tokens": 36358.0,
      "step": 87
    },
    {
      "epoch": 0.0067687100992231365,
      "grad_norm": 45.0,
      "learning_rate": 2.9898469348511656e-06,
      "loss": 1.0437,
      "mean_token_accuracy": 0.8603988885879517,
      "num_tokens": 36710.0,
      "step": 88
    },
    {
      "epoch": 0.006845627259441581,
      "grad_norm": 20.25,
      "learning_rate": 2.989731559110838e-06,
      "loss": 0.6288,
      "mean_token_accuracy": 0.8998144865036011,
      "num_tokens": 37250.0,
      "step": 89
    },
    {
      "epoch": 0.0069225444196600264,
      "grad_norm": 66.0,
      "learning_rate": 2.98961618337051e-06,
      "loss": 0.874,
      "mean_token_accuracy": 0.8753893971443176,
      "num_tokens": 37572.0,
      "step": 90
    },
    {
      "epoch": 0.006999461579878471,
      "grad_norm": 55.25,
      "learning_rate": 2.9895008076301823e-06,
      "loss": 1.0228,
      "mean_token_accuracy": 0.8507936596870422,
      "num_tokens": 37888.0,
      "step": 91
    },
    {
      "epoch": 0.0070763787400969155,
      "grad_norm": 50.0,
      "learning_rate": 2.9893854318898546e-06,
      "loss": 0.7468,
      "mean_token_accuracy": 0.9057971239089966,
      "num_tokens": 38166.0,
      "step": 92
    },
    {
      "epoch": 0.00715329590031536,
      "grad_norm": 30.375,
      "learning_rate": 2.9892700561495273e-06,
      "loss": 0.6356,
      "mean_token_accuracy": 0.8892921805381775,
      "num_tokens": 38718.0,
      "step": 93
    },
    {
      "epoch": 0.0072302130605338055,
      "grad_norm": 59.25,
      "learning_rate": 2.9891546804091995e-06,
      "loss": 0.7689,
      "mean_token_accuracy": 0.8878504633903503,
      "num_tokens": 39040.0,
      "step": 94
    },
    {
      "epoch": 0.00730713022075225,
      "grad_norm": 49.25,
      "learning_rate": 2.9890393046688714e-06,
      "loss": 0.5904,
      "mean_token_accuracy": 0.925000011920929,
      "num_tokens": 39364.0,
      "step": 95
    },
    {
      "epoch": 0.007384047380970695,
      "grad_norm": 21.5,
      "learning_rate": 2.988923928928544e-06,
      "loss": 0.5887,
      "mean_token_accuracy": 0.9005235433578491,
      "num_tokens": 39938.0,
      "step": 96
    },
    {
      "epoch": 0.007460964541189139,
      "grad_norm": 40.25,
      "learning_rate": 2.9888085531882163e-06,
      "loss": 0.6436,
      "mean_token_accuracy": 0.9191616773605347,
      "num_tokens": 40274.0,
      "step": 97
    },
    {
      "epoch": 0.007537881701407584,
      "grad_norm": 39.75,
      "learning_rate": 2.9886931774478886e-06,
      "loss": 0.6935,
      "mean_token_accuracy": 0.908196747303009,
      "num_tokens": 40580.0,
      "step": 98
    },
    {
      "epoch": 0.007614798861626029,
      "grad_norm": 74.5,
      "learning_rate": 2.9885778017075613e-06,
      "loss": 0.6936,
      "mean_token_accuracy": 0.9034267663955688,
      "num_tokens": 40902.0,
      "step": 99
    },
    {
      "epoch": 0.007691716021844474,
      "grad_norm": 25.75,
      "learning_rate": 2.9884624259672335e-06,
      "loss": 0.5106,
      "mean_token_accuracy": 0.9087779521942139,
      "num_tokens": 41484.0,
      "step": 100
    },
    {
      "epoch": 0.007768633182062918,
      "grad_norm": 163.0,
      "learning_rate": 2.9883470502269058e-06,
      "loss": 0.7401,
      "mean_token_accuracy": 0.8976109027862549,
      "num_tokens": 41778.0,
      "step": 101
    },
    {
      "epoch": 0.007845550342281363,
      "grad_norm": 79.5,
      "learning_rate": 2.988231674486578e-06,
      "loss": 0.6628,
      "mean_token_accuracy": 0.9153094291687012,
      "num_tokens": 42086.0,
      "step": 102
    },
    {
      "epoch": 0.007922467502499808,
      "grad_norm": 50.0,
      "learning_rate": 2.9881162987462503e-06,
      "loss": 0.5841,
      "mean_token_accuracy": 0.9065420627593994,
      "num_tokens": 42622.0,
      "step": 103
    },
    {
      "epoch": 0.007999384662718252,
      "grad_norm": 28.75,
      "learning_rate": 2.988000923005923e-06,
      "loss": 0.4983,
      "mean_token_accuracy": 0.9153153300285339,
      "num_tokens": 43178.0,
      "step": 104
    },
    {
      "epoch": 0.008076301822936697,
      "grad_norm": 22.875,
      "learning_rate": 2.9878855472655952e-06,
      "loss": 0.4675,
      "mean_token_accuracy": 0.9132652878761292,
      "num_tokens": 43964.0,
      "step": 105
    },
    {
      "epoch": 0.008153218983155143,
      "grad_norm": 20.5,
      "learning_rate": 2.987770171525267e-06,
      "loss": 0.4824,
      "mean_token_accuracy": 0.922398567199707,
      "num_tokens": 44532.0,
      "step": 106
    },
    {
      "epoch": 0.008230136143373586,
      "grad_norm": 32.25,
      "learning_rate": 2.9876547957849397e-06,
      "loss": 0.5606,
      "mean_token_accuracy": 0.9304347634315491,
      "num_tokens": 44878.0,
      "step": 107
    },
    {
      "epoch": 0.008307053303592032,
      "grad_norm": 63.0,
      "learning_rate": 2.987539420044612e-06,
      "loss": 0.6075,
      "mean_token_accuracy": 0.9311594367027283,
      "num_tokens": 45156.0,
      "step": 108
    },
    {
      "epoch": 0.008383970463810475,
      "grad_norm": 26.5,
      "learning_rate": 2.9874240443042843e-06,
      "loss": 0.4876,
      "mean_token_accuracy": 0.9176672101020813,
      "num_tokens": 45740.0,
      "step": 109
    },
    {
      "epoch": 0.00846088762402892,
      "grad_norm": 130.0,
      "learning_rate": 2.987308668563957e-06,
      "loss": 0.8474,
      "mean_token_accuracy": 0.8880308866500854,
      "num_tokens": 46000.0,
      "step": 110
    },
    {
      "epoch": 0.008537804784247366,
      "grad_norm": 21.0,
      "learning_rate": 2.987193292823629e-06,
      "loss": 0.4806,
      "mean_token_accuracy": 0.9234972596168518,
      "num_tokens": 46550.0,
      "step": 111
    },
    {
      "epoch": 0.00861472194446581,
      "grad_norm": 65.5,
      "learning_rate": 2.9870779170833015e-06,
      "loss": 0.7366,
      "mean_token_accuracy": 0.9058441519737244,
      "num_tokens": 46860.0,
      "step": 112
    },
    {
      "epoch": 0.008691639104684255,
      "grad_norm": 28.875,
      "learning_rate": 2.9869625413429737e-06,
      "loss": 0.4343,
      "mean_token_accuracy": 0.9347442388534546,
      "num_tokens": 47428.0,
      "step": 113
    },
    {
      "epoch": 0.008768556264902699,
      "grad_norm": 47.25,
      "learning_rate": 2.986847165602646e-06,
      "loss": 0.5997,
      "mean_token_accuracy": 0.919571042060852,
      "num_tokens": 47804.0,
      "step": 114
    },
    {
      "epoch": 0.008845473425121144,
      "grad_norm": 21.125,
      "learning_rate": 2.9867317898623182e-06,
      "loss": 0.4374,
      "mean_token_accuracy": 0.9354838728904724,
      "num_tokens": 48394.0,
      "step": 115
    },
    {
      "epoch": 0.00892239058533959,
      "grad_norm": 21.5,
      "learning_rate": 2.986616414121991e-06,
      "loss": 0.3905,
      "mean_token_accuracy": 0.9296675324440002,
      "num_tokens": 49178.0,
      "step": 116
    },
    {
      "epoch": 0.008999307745558034,
      "grad_norm": 45.25,
      "learning_rate": 2.9865010383816627e-06,
      "loss": 0.5187,
      "mean_token_accuracy": 0.9453551769256592,
      "num_tokens": 49546.0,
      "step": 117
    },
    {
      "epoch": 0.009076224905776479,
      "grad_norm": 18.375,
      "learning_rate": 2.9863856626413354e-06,
      "loss": 0.349,
      "mean_token_accuracy": 0.9395465850830078,
      "num_tokens": 50342.0,
      "step": 118
    },
    {
      "epoch": 0.009153142065994924,
      "grad_norm": 187.0,
      "learning_rate": 2.9862702869010077e-06,
      "loss": 0.6028,
      "mean_token_accuracy": 0.9317406415939331,
      "num_tokens": 50636.0,
      "step": 119
    },
    {
      "epoch": 0.009230059226213368,
      "grad_norm": 95.5,
      "learning_rate": 2.98615491116068e-06,
      "loss": 0.5634,
      "mean_token_accuracy": 0.9395017623901367,
      "num_tokens": 50918.0,
      "step": 120
    },
    {
      "epoch": 0.009306976386431813,
      "grad_norm": 25.875,
      "learning_rate": 2.9860395354203526e-06,
      "loss": 0.3447,
      "mean_token_accuracy": 0.9467889666557312,
      "num_tokens": 51464.0,
      "step": 121
    },
    {
      "epoch": 0.009383893546650257,
      "grad_norm": 50.25,
      "learning_rate": 2.9859241596800245e-06,
      "loss": 0.5715,
      "mean_token_accuracy": 0.9241982698440552,
      "num_tokens": 51808.0,
      "step": 122
    },
    {
      "epoch": 0.009460810706868703,
      "grad_norm": 42.0,
      "learning_rate": 2.9858087839396967e-06,
      "loss": 0.4701,
      "mean_token_accuracy": 0.9430894255638123,
      "num_tokens": 52180.0,
      "step": 123
    },
    {
      "epoch": 0.009537727867087148,
      "grad_norm": 292.0,
      "learning_rate": 2.9856934081993694e-06,
      "loss": 0.5359,
      "mean_token_accuracy": 0.938144326210022,
      "num_tokens": 52472.0,
      "step": 124
    },
    {
      "epoch": 0.009614645027305592,
      "grad_norm": 57.75,
      "learning_rate": 2.9855780324590417e-06,
      "loss": 0.414,
      "mean_token_accuracy": 0.9383177757263184,
      "num_tokens": 53008.0,
      "step": 125
    },
    {
      "epoch": 0.009691562187524037,
      "grad_norm": 19.375,
      "learning_rate": 2.985462656718714e-06,
      "loss": 0.2753,
      "mean_token_accuracy": 0.9611650705337524,
      "num_tokens": 53834.0,
      "step": 126
    },
    {
      "epoch": 0.00976847934774248,
      "grad_norm": 133.0,
      "learning_rate": 2.9853472809783866e-06,
      "loss": 0.5862,
      "mean_token_accuracy": 0.9314641952514648,
      "num_tokens": 54156.0,
      "step": 127
    },
    {
      "epoch": 0.009845396507960926,
      "grad_norm": 21.5,
      "learning_rate": 2.9852319052380584e-06,
      "loss": 0.3396,
      "mean_token_accuracy": 0.9551166892051697,
      "num_tokens": 54714.0,
      "step": 128
    },
    {
      "epoch": 0.009922313668179372,
      "grad_norm": 33.75,
      "learning_rate": 2.985116529497731e-06,
      "loss": 0.4657,
      "mean_token_accuracy": 0.9526627063751221,
      "num_tokens": 55054.0,
      "step": 129
    },
    {
      "epoch": 0.009999230828397815,
      "grad_norm": 53.5,
      "learning_rate": 2.9850011537574034e-06,
      "loss": 0.5524,
      "mean_token_accuracy": 0.9372937083244324,
      "num_tokens": 55358.0,
      "step": 130
    },
    {
      "epoch": 0.01007614798861626,
      "grad_norm": 14.4375,
      "learning_rate": 2.9848857780170756e-06,
      "loss": 0.2812,
      "mean_token_accuracy": 0.9654576778411865,
      "num_tokens": 55938.0,
      "step": 131
    },
    {
      "epoch": 0.010153065148834704,
      "grad_norm": 171.0,
      "learning_rate": 2.9847704022767483e-06,
      "loss": 0.6857,
      "mean_token_accuracy": 0.9066147804260254,
      "num_tokens": 56196.0,
      "step": 132
    },
    {
      "epoch": 0.01022998230905315,
      "grad_norm": 20.75,
      "learning_rate": 2.98465502653642e-06,
      "loss": 0.3356,
      "mean_token_accuracy": 0.9549718499183655,
      "num_tokens": 56730.0,
      "step": 133
    },
    {
      "epoch": 0.010306899469271595,
      "grad_norm": 20.25,
      "learning_rate": 2.9845396507960924e-06,
      "loss": 0.2942,
      "mean_token_accuracy": 0.9634369015693665,
      "num_tokens": 57278.0,
      "step": 134
    },
    {
      "epoch": 0.010383816629490039,
      "grad_norm": 20.125,
      "learning_rate": 2.984424275055765e-06,
      "loss": 0.3327,
      "mean_token_accuracy": 0.959114134311676,
      "num_tokens": 57866.0,
      "step": 135
    },
    {
      "epoch": 0.010460733789708484,
      "grad_norm": 19.125,
      "learning_rate": 2.9843088993154373e-06,
      "loss": 0.317,
      "mean_token_accuracy": 0.9587242007255554,
      "num_tokens": 58400.0,
      "step": 136
    },
    {
      "epoch": 0.010537650949926928,
      "grad_norm": 21.75,
      "learning_rate": 2.9841935235751096e-06,
      "loss": 0.2977,
      "mean_token_accuracy": 0.9613152742385864,
      "num_tokens": 58918.0,
      "step": 137
    },
    {
      "epoch": 0.010614568110145373,
      "grad_norm": 19.625,
      "learning_rate": 2.9840781478347823e-06,
      "loss": 0.2865,
      "mean_token_accuracy": 0.9617486596107483,
      "num_tokens": 59468.0,
      "step": 138
    },
    {
      "epoch": 0.010691485270363819,
      "grad_norm": 16.25,
      "learning_rate": 2.983962772094454e-06,
      "loss": 0.1873,
      "mean_token_accuracy": 0.9826732873916626,
      "num_tokens": 60278.0,
      "step": 139
    },
    {
      "epoch": 0.010768402430582262,
      "grad_norm": 62.5,
      "learning_rate": 2.983847396354127e-06,
      "loss": 0.5019,
      "mean_token_accuracy": 0.9417475461959839,
      "num_tokens": 60588.0,
      "step": 140
    },
    {
      "epoch": 0.010845319590800708,
      "grad_norm": 16.0,
      "learning_rate": 2.983732020613799e-06,
      "loss": 0.1527,
      "mean_token_accuracy": 0.9905392527580261,
      "num_tokens": 61648.0,
      "step": 141
    },
    {
      "epoch": 0.010922236751019153,
      "grad_norm": 95.0,
      "learning_rate": 2.9836166448734713e-06,
      "loss": 0.3292,
      "mean_token_accuracy": 0.9750000238418579,
      "num_tokens": 61972.0,
      "step": 142
    },
    {
      "epoch": 0.010999153911237597,
      "grad_norm": 109.5,
      "learning_rate": 2.983501269133144e-06,
      "loss": 0.4681,
      "mean_token_accuracy": 0.9492753744125366,
      "num_tokens": 62250.0,
      "step": 143
    },
    {
      "epoch": 0.011076071071456042,
      "grad_norm": 17.375,
      "learning_rate": 2.983385893392816e-06,
      "loss": 0.1863,
      "mean_token_accuracy": 0.9833759665489197,
      "num_tokens": 63034.0,
      "step": 144
    },
    {
      "epoch": 0.011152988231674486,
      "grad_norm": 30.375,
      "learning_rate": 2.983270517652488e-06,
      "loss": 0.2768,
      "mean_token_accuracy": 0.9706422090530396,
      "num_tokens": 63580.0,
      "step": 145
    },
    {
      "epoch": 0.011229905391892931,
      "grad_norm": 21.875,
      "learning_rate": 2.9831551419121608e-06,
      "loss": 0.161,
      "mean_token_accuracy": 0.9840849041938782,
      "num_tokens": 64336.0,
      "step": 146
    },
    {
      "epoch": 0.011306822552111377,
      "grad_norm": 31.25,
      "learning_rate": 2.983039766171833e-06,
      "loss": 0.2929,
      "mean_token_accuracy": 0.9672130942344666,
      "num_tokens": 64886.0,
      "step": 147
    },
    {
      "epoch": 0.01138373971232982,
      "grad_norm": 18.625,
      "learning_rate": 2.9829243904315053e-06,
      "loss": 0.2357,
      "mean_token_accuracy": 0.9730700254440308,
      "num_tokens": 65444.0,
      "step": 148
    },
    {
      "epoch": 0.011460656872548266,
      "grad_norm": 14.125,
      "learning_rate": 2.982809014691178e-06,
      "loss": 0.2381,
      "mean_token_accuracy": 0.9767441749572754,
      "num_tokens": 66004.0,
      "step": 149
    },
    {
      "epoch": 0.01153757403276671,
      "grad_norm": 13.3125,
      "learning_rate": 2.98269363895085e-06,
      "loss": 0.1381,
      "mean_token_accuracy": 0.9900249242782593,
      "num_tokens": 66808.0,
      "step": 150
    },
    {
      "epoch": 0.011614491192985155,
      "grad_norm": 241.0,
      "learning_rate": 2.9825782632105225e-06,
      "loss": 0.5469,
      "mean_token_accuracy": 0.9333333373069763,
      "num_tokens": 67064.0,
      "step": 151
    },
    {
      "epoch": 0.0116914083532036,
      "grad_norm": 172.0,
      "learning_rate": 2.9824628874701947e-06,
      "loss": 0.4805,
      "mean_token_accuracy": 0.9421221613883972,
      "num_tokens": 67376.0,
      "step": 152
    },
    {
      "epoch": 0.011768325513422044,
      "grad_norm": 24.625,
      "learning_rate": 2.982347511729867e-06,
      "loss": 0.225,
      "mean_token_accuracy": 0.9773123860359192,
      "num_tokens": 67950.0,
      "step": 153
    },
    {
      "epoch": 0.01184524267364049,
      "grad_norm": 26.25,
      "learning_rate": 2.9822321359895397e-06,
      "loss": 0.2129,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 68482.0,
      "step": 154
    },
    {
      "epoch": 0.011922159833858933,
      "grad_norm": 92.0,
      "learning_rate": 2.9821167602492115e-06,
      "loss": 0.4401,
      "mean_token_accuracy": 0.9491525292396545,
      "num_tokens": 68778.0,
      "step": 155
    },
    {
      "epoch": 0.011999076994077379,
      "grad_norm": 25.5,
      "learning_rate": 2.9820013845088838e-06,
      "loss": 0.1333,
      "mean_token_accuracy": 0.9901719689369202,
      "num_tokens": 69594.0,
      "step": 156
    },
    {
      "epoch": 0.012075994154295824,
      "grad_norm": 156.0,
      "learning_rate": 2.9818860087685565e-06,
      "loss": 0.4396,
      "mean_token_accuracy": 0.95652174949646,
      "num_tokens": 69872.0,
      "step": 157
    },
    {
      "epoch": 0.012152911314514268,
      "grad_norm": 27.75,
      "learning_rate": 2.9817706330282287e-06,
      "loss": 0.1898,
      "mean_token_accuracy": 0.9859402179718018,
      "num_tokens": 70442.0,
      "step": 158
    },
    {
      "epoch": 0.012229828474732713,
      "grad_norm": 10.875,
      "learning_rate": 2.981655257287901e-06,
      "loss": 0.1306,
      "mean_token_accuracy": 0.9902200698852539,
      "num_tokens": 71262.0,
      "step": 159
    },
    {
      "epoch": 0.012306745634951157,
      "grad_norm": 149.0,
      "learning_rate": 2.9815398815475732e-06,
      "loss": 0.4326,
      "mean_token_accuracy": 0.95652174949646,
      "num_tokens": 71540.0,
      "step": 160
    },
    {
      "epoch": 0.012383662795169602,
      "grad_norm": 14.8125,
      "learning_rate": 2.9814245058072455e-06,
      "loss": 0.1992,
      "mean_token_accuracy": 0.9820466637611389,
      "num_tokens": 72098.0,
      "step": 161
    },
    {
      "epoch": 0.012460579955388048,
      "grad_norm": 13.3125,
      "learning_rate": 2.981309130066918e-06,
      "loss": 0.1825,
      "mean_token_accuracy": 0.9871559739112854,
      "num_tokens": 72644.0,
      "step": 162
    },
    {
      "epoch": 0.012537497115606491,
      "grad_norm": 33.25,
      "learning_rate": 2.9811937543265904e-06,
      "loss": 0.2057,
      "mean_token_accuracy": 0.9833641648292542,
      "num_tokens": 73186.0,
      "step": 163
    },
    {
      "epoch": 0.012614414275824937,
      "grad_norm": 26.0,
      "learning_rate": 2.9810783785862627e-06,
      "loss": 0.202,
      "mean_token_accuracy": 0.9808027744293213,
      "num_tokens": 73760.0,
      "step": 164
    },
    {
      "epoch": 0.012691331436043382,
      "grad_norm": 82.5,
      "learning_rate": 2.980963002845935e-06,
      "loss": 0.4092,
      "mean_token_accuracy": 0.9604862928390503,
      "num_tokens": 74090.0,
      "step": 165
    },
    {
      "epoch": 0.012768248596261826,
      "grad_norm": 60.25,
      "learning_rate": 2.980847627105607e-06,
      "loss": 0.3414,
      "mean_token_accuracy": 0.9625360369682312,
      "num_tokens": 74440.0,
      "step": 166
    },
    {
      "epoch": 0.012845165756480271,
      "grad_norm": 69.5,
      "learning_rate": 2.9807322513652795e-06,
      "loss": 0.3904,
      "mean_token_accuracy": 0.9590443968772888,
      "num_tokens": 74734.0,
      "step": 167
    },
    {
      "epoch": 0.012922082916698715,
      "grad_norm": 63.0,
      "learning_rate": 2.980616875624952e-06,
      "loss": 0.3505,
      "mean_token_accuracy": 0.970059871673584,
      "num_tokens": 75070.0,
      "step": 168
    },
    {
      "epoch": 0.01299900007691716,
      "grad_norm": 226.0,
      "learning_rate": 2.9805014998846244e-06,
      "loss": 0.4743,
      "mean_token_accuracy": 0.9477351903915405,
      "num_tokens": 75358.0,
      "step": 169
    },
    {
      "epoch": 0.013075917237135606,
      "grad_norm": 46.5,
      "learning_rate": 2.9803861241442967e-06,
      "loss": 0.4247,
      "mean_token_accuracy": 0.9532710313796997,
      "num_tokens": 75680.0,
      "step": 170
    },
    {
      "epoch": 0.01315283439735405,
      "grad_norm": 183.0,
      "learning_rate": 2.980270748403969e-06,
      "loss": 0.3825,
      "mean_token_accuracy": 0.9641693830490112,
      "num_tokens": 75988.0,
      "step": 171
    },
    {
      "epoch": 0.013229751557572495,
      "grad_norm": 11.5625,
      "learning_rate": 2.980155372663641e-06,
      "loss": 0.1733,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 76574.0,
      "step": 172
    },
    {
      "epoch": 0.013306668717790938,
      "grad_norm": 354.0,
      "learning_rate": 2.9800399969233134e-06,
      "loss": 0.4765,
      "mean_token_accuracy": 0.9455252885818481,
      "num_tokens": 76832.0,
      "step": 173
    },
    {
      "epoch": 0.013383585878009384,
      "grad_norm": 19.75,
      "learning_rate": 2.979924621182986e-06,
      "loss": 0.1968,
      "mean_token_accuracy": 0.9845857620239258,
      "num_tokens": 77352.0,
      "step": 174
    },
    {
      "epoch": 0.01346050303822783,
      "grad_norm": 400.0,
      "learning_rate": 2.9798092454426584e-06,
      "loss": 0.3892,
      "mean_token_accuracy": 0.9640287756919861,
      "num_tokens": 77632.0,
      "step": 175
    },
    {
      "epoch": 0.013537420198446273,
      "grad_norm": 43.25,
      "learning_rate": 2.9796938697023306e-06,
      "loss": 0.2858,
      "mean_token_accuracy": 0.9750000238418579,
      "num_tokens": 77956.0,
      "step": 176
    },
    {
      "epoch": 0.013614337358664718,
      "grad_norm": 64.5,
      "learning_rate": 2.979578493962003e-06,
      "loss": 0.3767,
      "mean_token_accuracy": 0.9665551781654358,
      "num_tokens": 78256.0,
      "step": 177
    },
    {
      "epoch": 0.013691254518883162,
      "grad_norm": 30.125,
      "learning_rate": 2.979463118221675e-06,
      "loss": 0.1295,
      "mean_token_accuracy": 0.9890776872634888,
      "num_tokens": 79082.0,
      "step": 178
    },
    {
      "epoch": 0.013768171679101607,
      "grad_norm": 54.25,
      "learning_rate": 2.979347742481348e-06,
      "loss": 0.3491,
      "mean_token_accuracy": 0.978723406791687,
      "num_tokens": 79412.0,
      "step": 179
    },
    {
      "epoch": 0.013845088839320053,
      "grad_norm": 19.75,
      "learning_rate": 2.97923236674102e-06,
      "loss": 0.1338,
      "mean_token_accuracy": 0.9886934757232666,
      "num_tokens": 80210.0,
      "step": 180
    },
    {
      "epoch": 0.013922005999538497,
      "grad_norm": 116.5,
      "learning_rate": 2.9791169910006923e-06,
      "loss": 0.2947,
      "mean_token_accuracy": 0.9809523820877075,
      "num_tokens": 80528.0,
      "step": 181
    },
    {
      "epoch": 0.013998923159756942,
      "grad_norm": 54.25,
      "learning_rate": 2.9790016152603646e-06,
      "loss": 0.3326,
      "mean_token_accuracy": 0.9674556255340576,
      "num_tokens": 80868.0,
      "step": 182
    },
    {
      "epoch": 0.014075840319975386,
      "grad_norm": 426.0,
      "learning_rate": 2.978886239520037e-06,
      "loss": 0.4206,
      "mean_token_accuracy": 0.9540635943412781,
      "num_tokens": 81152.0,
      "step": 183
    },
    {
      "epoch": 0.014152757480193831,
      "grad_norm": 46.0,
      "learning_rate": 2.978770863779709e-06,
      "loss": 0.3965,
      "mean_token_accuracy": 0.9599999785423279,
      "num_tokens": 81478.0,
      "step": 184
    },
    {
      "epoch": 0.014229674640412276,
      "grad_norm": 190.0,
      "learning_rate": 2.978655488039382e-06,
      "loss": 0.4233,
      "mean_token_accuracy": 0.9415584206581116,
      "num_tokens": 81788.0,
      "step": 185
    },
    {
      "epoch": 0.01430659180063072,
      "grad_norm": 19.5,
      "learning_rate": 2.978540112299054e-06,
      "loss": 0.1751,
      "mean_token_accuracy": 0.9851577281951904,
      "num_tokens": 82328.0,
      "step": 186
    },
    {
      "epoch": 0.014383508960849166,
      "grad_norm": 342.0,
      "learning_rate": 2.9784247365587263e-06,
      "loss": 0.3624,
      "mean_token_accuracy": 0.9637681245803833,
      "num_tokens": 82606.0,
      "step": 187
    },
    {
      "epoch": 0.014460426121067611,
      "grad_norm": 528.0,
      "learning_rate": 2.9783093608183986e-06,
      "loss": 0.4217,
      "mean_token_accuracy": 0.9578543901443481,
      "num_tokens": 82868.0,
      "step": 188
    },
    {
      "epoch": 0.014537343281286055,
      "grad_norm": 15.8125,
      "learning_rate": 2.978193985078071e-06,
      "loss": 0.1824,
      "mean_token_accuracy": 0.9808027744293213,
      "num_tokens": 83442.0,
      "step": 189
    },
    {
      "epoch": 0.0146142604415045,
      "grad_norm": 14.625,
      "learning_rate": 2.9780786093377435e-06,
      "loss": 0.1701,
      "mean_token_accuracy": 0.9872029423713684,
      "num_tokens": 83990.0,
      "step": 190
    },
    {
      "epoch": 0.014691177601722944,
      "grad_norm": 10.8125,
      "learning_rate": 2.9779632335974158e-06,
      "loss": 0.1138,
      "mean_token_accuracy": 0.9913366436958313,
      "num_tokens": 84800.0,
      "step": 191
    },
    {
      "epoch": 0.01476809476194139,
      "grad_norm": 15.0,
      "learning_rate": 2.977847857857088e-06,
      "loss": 0.1581,
      "mean_token_accuracy": 0.9878261089324951,
      "num_tokens": 85376.0,
      "step": 192
    },
    {
      "epoch": 0.014845011922159835,
      "grad_norm": 161.0,
      "learning_rate": 2.9777324821167603e-06,
      "loss": 0.3603,
      "mean_token_accuracy": 0.9646643400192261,
      "num_tokens": 85660.0,
      "step": 193
    },
    {
      "epoch": 0.014921929082378278,
      "grad_norm": 27.0,
      "learning_rate": 2.9776171063764325e-06,
      "loss": 0.1834,
      "mean_token_accuracy": 0.9824868440628052,
      "num_tokens": 86232.0,
      "step": 194
    },
    {
      "epoch": 0.014998846242596724,
      "grad_norm": 26.375,
      "learning_rate": 2.977501730636105e-06,
      "loss": 0.186,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 86764.0,
      "step": 195
    },
    {
      "epoch": 0.015075763402815167,
      "grad_norm": 12.375,
      "learning_rate": 2.9773863548957775e-06,
      "loss": 0.0949,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 87812.0,
      "step": 196
    },
    {
      "epoch": 0.015152680563033613,
      "grad_norm": 18.5,
      "learning_rate": 2.9772709791554497e-06,
      "loss": 0.1736,
      "mean_token_accuracy": 0.9838420152664185,
      "num_tokens": 88370.0,
      "step": 197
    },
    {
      "epoch": 0.015229597723252058,
      "grad_norm": 16.375,
      "learning_rate": 2.977155603415122e-06,
      "loss": 0.1683,
      "mean_token_accuracy": 0.9871559739112854,
      "num_tokens": 88916.0,
      "step": 198
    },
    {
      "epoch": 0.015306514883470502,
      "grad_norm": 14.125,
      "learning_rate": 2.9770402276747943e-06,
      "loss": 0.1749,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 89460.0,
      "step": 199
    },
    {
      "epoch": 0.015383432043688947,
      "grad_norm": 59.25,
      "learning_rate": 2.9769248519344665e-06,
      "loss": 0.3683,
      "mean_token_accuracy": 0.9639639854431152,
      "num_tokens": 89794.0,
      "step": 200
    },
    {
      "epoch": 0.015460349203907391,
      "grad_norm": 37.0,
      "learning_rate": 2.976809476194139e-06,
      "loss": 0.365,
      "mean_token_accuracy": 0.9646302461624146,
      "num_tokens": 90106.0,
      "step": 201
    },
    {
      "epoch": 0.015537266364125836,
      "grad_norm": 19.5,
      "learning_rate": 2.9766941004538115e-06,
      "loss": 0.1189,
      "mean_token_accuracy": 0.9875930547714233,
      "num_tokens": 90914.0,
      "step": 202
    },
    {
      "epoch": 0.015614183524344282,
      "grad_norm": 19.125,
      "learning_rate": 2.9765787247134837e-06,
      "loss": 0.1609,
      "mean_token_accuracy": 0.9876543283462524,
      "num_tokens": 91482.0,
      "step": 203
    },
    {
      "epoch": 0.015691100684562725,
      "grad_norm": 48.75,
      "learning_rate": 2.976463348973156e-06,
      "loss": 0.1656,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 92068.0,
      "step": 204
    },
    {
      "epoch": 0.01576801784478117,
      "grad_norm": 584.0,
      "learning_rate": 2.9763479732328282e-06,
      "loss": 0.3633,
      "mean_token_accuracy": 0.9647058844566345,
      "num_tokens": 92324.0,
      "step": 205
    },
    {
      "epoch": 0.015844935004999616,
      "grad_norm": 11.9375,
      "learning_rate": 2.9762325974925005e-06,
      "loss": 0.1621,
      "mean_token_accuracy": 0.9871086478233337,
      "num_tokens": 92868.0,
      "step": 206
    },
    {
      "epoch": 0.01592185216521806,
      "grad_norm": 66.0,
      "learning_rate": 2.976117221752173e-06,
      "loss": 0.3015,
      "mean_token_accuracy": 0.9664804339408875,
      "num_tokens": 93228.0,
      "step": 207
    },
    {
      "epoch": 0.015998769325436504,
      "grad_norm": 19.625,
      "learning_rate": 2.9760018460118454e-06,
      "loss": 0.1745,
      "mean_token_accuracy": 0.984644889831543,
      "num_tokens": 93750.0,
      "step": 208
    },
    {
      "epoch": 0.01607568648565495,
      "grad_norm": 115.0,
      "learning_rate": 2.9758864702715177e-06,
      "loss": 0.3487,
      "mean_token_accuracy": 0.9681978821754456,
      "num_tokens": 94034.0,
      "step": 209
    },
    {
      "epoch": 0.016152603645873394,
      "grad_norm": 19.375,
      "learning_rate": 2.97577109453119e-06,
      "loss": 0.1637,
      "mean_token_accuracy": 0.9838420152664185,
      "num_tokens": 94592.0,
      "step": 210
    },
    {
      "epoch": 0.01622952080609184,
      "grad_norm": 23.75,
      "learning_rate": 2.975655718790862e-06,
      "loss": 0.1565,
      "mean_token_accuracy": 0.9876543283462524,
      "num_tokens": 95160.0,
      "step": 211
    },
    {
      "epoch": 0.016306437966310285,
      "grad_norm": 42.75,
      "learning_rate": 2.975540343050535e-06,
      "loss": 0.3533,
      "mean_token_accuracy": 0.9609120488166809,
      "num_tokens": 95468.0,
      "step": 212
    },
    {
      "epoch": 0.016383355126528727,
      "grad_norm": 13.1875,
      "learning_rate": 2.975424967310207e-06,
      "loss": 0.1584,
      "mean_token_accuracy": 0.9857397675514221,
      "num_tokens": 96030.0,
      "step": 213
    },
    {
      "epoch": 0.016460272286747173,
      "grad_norm": 11.5,
      "learning_rate": 2.9753095915698794e-06,
      "loss": 0.1576,
      "mean_token_accuracy": 0.985637366771698,
      "num_tokens": 96588.0,
      "step": 214
    },
    {
      "epoch": 0.016537189446965618,
      "grad_norm": 19.875,
      "learning_rate": 2.9751942158295517e-06,
      "loss": 0.1538,
      "mean_token_accuracy": 0.9840142130851746,
      "num_tokens": 97152.0,
      "step": 215
    },
    {
      "epoch": 0.016614106607184063,
      "grad_norm": 13.0625,
      "learning_rate": 2.975078840089224e-06,
      "loss": 0.151,
      "mean_token_accuracy": 0.9877408146858215,
      "num_tokens": 97724.0,
      "step": 216
    },
    {
      "epoch": 0.01669102376740251,
      "grad_norm": 14.5,
      "learning_rate": 2.974963464348896e-06,
      "loss": 0.1571,
      "mean_token_accuracy": 0.9852670431137085,
      "num_tokens": 98268.0,
      "step": 217
    },
    {
      "epoch": 0.01676794092762095,
      "grad_norm": 15.125,
      "learning_rate": 2.974848088608569e-06,
      "loss": 0.1481,
      "mean_token_accuracy": 0.9863713979721069,
      "num_tokens": 98856.0,
      "step": 218
    },
    {
      "epoch": 0.016844858087839396,
      "grad_norm": 32.75,
      "learning_rate": 2.974732712868241e-06,
      "loss": 0.3317,
      "mean_token_accuracy": 0.9646302461624146,
      "num_tokens": 99168.0,
      "step": 219
    },
    {
      "epoch": 0.01692177524805784,
      "grad_norm": 9.125,
      "learning_rate": 2.9746173371279134e-06,
      "loss": 0.1507,
      "mean_token_accuracy": 0.9874776601791382,
      "num_tokens": 99728.0,
      "step": 220
    },
    {
      "epoch": 0.016998692408276287,
      "grad_norm": 48.25,
      "learning_rate": 2.9745019613875856e-06,
      "loss": 0.3054,
      "mean_token_accuracy": 0.9716088175773621,
      "num_tokens": 100046.0,
      "step": 221
    },
    {
      "epoch": 0.017075609568494732,
      "grad_norm": 34.0,
      "learning_rate": 2.974386585647258e-06,
      "loss": 0.3189,
      "mean_token_accuracy": 0.9684210419654846,
      "num_tokens": 100332.0,
      "step": 222
    },
    {
      "epoch": 0.017152526728713174,
      "grad_norm": 29.375,
      "learning_rate": 2.97427120990693e-06,
      "loss": 0.2929,
      "mean_token_accuracy": 0.9740633964538574,
      "num_tokens": 100680.0,
      "step": 223
    },
    {
      "epoch": 0.01722944388893162,
      "grad_norm": 78.0,
      "learning_rate": 2.974155834166603e-06,
      "loss": 0.3187,
      "mean_token_accuracy": 0.9715302586555481,
      "num_tokens": 100962.0,
      "step": 224
    },
    {
      "epoch": 0.017306361049150065,
      "grad_norm": 33.0,
      "learning_rate": 2.9740404584262747e-06,
      "loss": 0.3214,
      "mean_token_accuracy": 0.9708737730979919,
      "num_tokens": 101272.0,
      "step": 225
    },
    {
      "epoch": 0.01738327820936851,
      "grad_norm": 49.25,
      "learning_rate": 2.9739250826859473e-06,
      "loss": 0.2918,
      "mean_token_accuracy": 0.9713466763496399,
      "num_tokens": 101622.0,
      "step": 226
    },
    {
      "epoch": 0.017460195369586956,
      "grad_norm": 39.5,
      "learning_rate": 2.9738097069456196e-06,
      "loss": 0.3151,
      "mean_token_accuracy": 0.969348669052124,
      "num_tokens": 101884.0,
      "step": 227
    },
    {
      "epoch": 0.017537112529805398,
      "grad_norm": 17.125,
      "learning_rate": 2.973694331205292e-06,
      "loss": 0.2772,
      "mean_token_accuracy": 0.9772727489471436,
      "num_tokens": 102194.0,
      "step": 228
    },
    {
      "epoch": 0.017614029690023843,
      "grad_norm": 15.4375,
      "learning_rate": 2.9735789554649645e-06,
      "loss": 0.1518,
      "mean_token_accuracy": 0.9874776601791382,
      "num_tokens": 102754.0,
      "step": 229
    },
    {
      "epoch": 0.01769094685024229,
      "grad_norm": 12.9375,
      "learning_rate": 2.973463579724637e-06,
      "loss": 0.1503,
      "mean_token_accuracy": 0.985637366771698,
      "num_tokens": 103312.0,
      "step": 230
    },
    {
      "epoch": 0.017767864010460734,
      "grad_norm": 15.625,
      "learning_rate": 2.9733482039843086e-06,
      "loss": 0.1392,
      "mean_token_accuracy": 0.9865546226501465,
      "num_tokens": 103908.0,
      "step": 231
    },
    {
      "epoch": 0.01784478117067918,
      "grad_norm": 30.75,
      "learning_rate": 2.9732328282439813e-06,
      "loss": 0.3121,
      "mean_token_accuracy": 0.9681978821754456,
      "num_tokens": 104192.0,
      "step": 232
    },
    {
      "epoch": 0.01792169833089762,
      "grad_norm": 16.125,
      "learning_rate": 2.9731174525036536e-06,
      "loss": 0.1494,
      "mean_token_accuracy": 0.9854280352592468,
      "num_tokens": 104742.0,
      "step": 233
    },
    {
      "epoch": 0.017998615491116067,
      "grad_norm": 9.8125,
      "learning_rate": 2.973002076763326e-06,
      "loss": 0.1009,
      "mean_token_accuracy": 0.9912060499191284,
      "num_tokens": 105540.0,
      "step": 234
    },
    {
      "epoch": 0.018075532651334512,
      "grad_norm": 10.75,
      "learning_rate": 2.9728867010229985e-06,
      "loss": 0.104,
      "mean_token_accuracy": 0.9894179701805115,
      "num_tokens": 106298.0,
      "step": 235
    },
    {
      "epoch": 0.018152449811552958,
      "grad_norm": 7.40625,
      "learning_rate": 2.9727713252826704e-06,
      "loss": 0.1389,
      "mean_token_accuracy": 0.9855855703353882,
      "num_tokens": 106854.0,
      "step": 236
    },
    {
      "epoch": 0.018229366971771403,
      "grad_norm": 52.5,
      "learning_rate": 2.972655949542343e-06,
      "loss": 0.2813,
      "mean_token_accuracy": 0.9698996543884277,
      "num_tokens": 107154.0,
      "step": 237
    },
    {
      "epoch": 0.01830628413198985,
      "grad_norm": 13.875,
      "learning_rate": 2.9725405738020153e-06,
      "loss": 0.1381,
      "mean_token_accuracy": 0.9856887459754944,
      "num_tokens": 107714.0,
      "step": 238
    },
    {
      "epoch": 0.01838320129220829,
      "grad_norm": 8.5625,
      "learning_rate": 2.9724251980616875e-06,
      "loss": 0.0985,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 108500.0,
      "step": 239
    },
    {
      "epoch": 0.018460118452426736,
      "grad_norm": 37.0,
      "learning_rate": 2.9723098223213602e-06,
      "loss": 0.2757,
      "mean_token_accuracy": 0.9722222089767456,
      "num_tokens": 108790.0,
      "step": 240
    },
    {
      "epoch": 0.01853703561264518,
      "grad_norm": 24.5,
      "learning_rate": 2.9721944465810325e-06,
      "loss": 0.1556,
      "mean_token_accuracy": 0.9865125417709351,
      "num_tokens": 109310.0,
      "step": 241
    },
    {
      "epoch": 0.018613952772863627,
      "grad_norm": 42.25,
      "learning_rate": 2.9720790708407043e-06,
      "loss": 0.3121,
      "mean_token_accuracy": 0.9639344215393066,
      "num_tokens": 109616.0,
      "step": 242
    },
    {
      "epoch": 0.018690869933082072,
      "grad_norm": 89.0,
      "learning_rate": 2.971963695100377e-06,
      "loss": 0.2748,
      "mean_token_accuracy": 0.9661971926689148,
      "num_tokens": 109972.0,
      "step": 243
    },
    {
      "epoch": 0.018767787093300514,
      "grad_norm": 10.1875,
      "learning_rate": 2.9718483193600493e-06,
      "loss": 0.1338,
      "mean_token_accuracy": 0.9860383868217468,
      "num_tokens": 110546.0,
      "step": 244
    },
    {
      "epoch": 0.01884470425351896,
      "grad_norm": 18.875,
      "learning_rate": 2.9717329436197215e-06,
      "loss": 0.2845,
      "mean_token_accuracy": 0.9651567935943604,
      "num_tokens": 110834.0,
      "step": 245
    },
    {
      "epoch": 0.018921621413737405,
      "grad_norm": 13.25,
      "learning_rate": 2.971617567879394e-06,
      "loss": 0.1318,
      "mean_token_accuracy": 0.9827288389205933,
      "num_tokens": 111414.0,
      "step": 246
    },
    {
      "epoch": 0.01899853857395585,
      "grad_norm": 11.375,
      "learning_rate": 2.971502192139066e-06,
      "loss": 0.1307,
      "mean_token_accuracy": 0.9860383868217468,
      "num_tokens": 111988.0,
      "step": 247
    },
    {
      "epoch": 0.019075455734174296,
      "grad_norm": 8.125,
      "learning_rate": 2.9713868163987387e-06,
      "loss": 0.0929,
      "mean_token_accuracy": 0.9913793206214905,
      "num_tokens": 112802.0,
      "step": 248
    },
    {
      "epoch": 0.019152372894392738,
      "grad_norm": 58.25,
      "learning_rate": 2.971271440658411e-06,
      "loss": 0.2441,
      "mean_token_accuracy": 0.9750000238418579,
      "num_tokens": 113164.0,
      "step": 249
    },
    {
      "epoch": 0.019229290054611183,
      "grad_norm": 24.5,
      "learning_rate": 2.9711560649180832e-06,
      "loss": 0.2803,
      "mean_token_accuracy": 0.9681978821754456,
      "num_tokens": 113448.0,
      "step": 250
    },
    {
      "epoch": 0.01930620721482963,
      "grad_norm": 94.5,
      "learning_rate": 2.971040689177756e-06,
      "loss": 0.262,
      "mean_token_accuracy": 0.9755351543426514,
      "num_tokens": 113776.0,
      "step": 251
    },
    {
      "epoch": 0.019383124375048074,
      "grad_norm": 19.375,
      "learning_rate": 2.970925313437428e-06,
      "loss": 0.1507,
      "mean_token_accuracy": 0.9845260977745056,
      "num_tokens": 114294.0,
      "step": 252
    },
    {
      "epoch": 0.01946004153526652,
      "grad_norm": 18.125,
      "learning_rate": 2.9708099376971e-06,
      "loss": 0.1396,
      "mean_token_accuracy": 0.9855855703353882,
      "num_tokens": 114850.0,
      "step": 253
    },
    {
      "epoch": 0.01953695869548496,
      "grad_norm": 8.1875,
      "learning_rate": 2.9706945619567727e-06,
      "loss": 0.0714,
      "mean_token_accuracy": 0.9923591017723083,
      "num_tokens": 115900.0,
      "step": 254
    },
    {
      "epoch": 0.019613875855703407,
      "grad_norm": 60.5,
      "learning_rate": 2.970579186216445e-06,
      "loss": 0.2613,
      "mean_token_accuracy": 0.9690402746200562,
      "num_tokens": 116224.0,
      "step": 255
    },
    {
      "epoch": 0.019690793015921852,
      "grad_norm": 7.8125,
      "learning_rate": 2.970463810476117e-06,
      "loss": 0.0922,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 117010.0,
      "step": 256
    },
    {
      "epoch": 0.019767710176140298,
      "grad_norm": 9.0,
      "learning_rate": 2.97034843473579e-06,
      "loss": 0.1297,
      "mean_token_accuracy": 0.9853211045265198,
      "num_tokens": 117556.0,
      "step": 257
    },
    {
      "epoch": 0.019844627336358743,
      "grad_norm": 9.0,
      "learning_rate": 2.9702330589954617e-06,
      "loss": 0.1268,
      "mean_token_accuracy": 0.987522304058075,
      "num_tokens": 118118.0,
      "step": 258
    },
    {
      "epoch": 0.019921544496577185,
      "grad_norm": 38.25,
      "learning_rate": 2.9701176832551344e-06,
      "loss": 0.2342,
      "mean_token_accuracy": 0.9689265489578247,
      "num_tokens": 118474.0,
      "step": 259
    },
    {
      "epoch": 0.01999846165679563,
      "grad_norm": 30.625,
      "learning_rate": 2.9700023075148067e-06,
      "loss": 0.2533,
      "mean_token_accuracy": 0.9673202633857727,
      "num_tokens": 118782.0,
      "step": 260
    },
    {
      "epoch": 0.020075378817014076,
      "grad_norm": 9.25,
      "learning_rate": 2.969886931774479e-06,
      "loss": 0.1206,
      "mean_token_accuracy": 0.988034188747406,
      "num_tokens": 119368.0,
      "step": 261
    },
    {
      "epoch": 0.02015229597723252,
      "grad_norm": 20.875,
      "learning_rate": 2.9697715560341516e-06,
      "loss": 0.2242,
      "mean_token_accuracy": 0.9740259647369385,
      "num_tokens": 119678.0,
      "step": 262
    },
    {
      "epoch": 0.020229213137450967,
      "grad_norm": 33.0,
      "learning_rate": 2.9696561802938234e-06,
      "loss": 0.2239,
      "mean_token_accuracy": 0.978787899017334,
      "num_tokens": 120010.0,
      "step": 263
    },
    {
      "epoch": 0.02030613029766941,
      "grad_norm": 38.0,
      "learning_rate": 2.9695408045534957e-06,
      "loss": 0.2484,
      "mean_token_accuracy": 0.9650145769119263,
      "num_tokens": 120354.0,
      "step": 264
    },
    {
      "epoch": 0.020383047457887854,
      "grad_norm": 12.125,
      "learning_rate": 2.9694254288131684e-06,
      "loss": 0.1238,
      "mean_token_accuracy": 0.9873873591423035,
      "num_tokens": 120910.0,
      "step": 265
    },
    {
      "epoch": 0.0204599646181063,
      "grad_norm": 38.75,
      "learning_rate": 2.9693100530728406e-06,
      "loss": 0.2747,
      "mean_token_accuracy": 0.9636963605880737,
      "num_tokens": 121214.0,
      "step": 266
    },
    {
      "epoch": 0.020536881778324745,
      "grad_norm": 37.0,
      "learning_rate": 2.969194677332513e-06,
      "loss": 0.2736,
      "mean_token_accuracy": 0.9672130942344666,
      "num_tokens": 121520.0,
      "step": 267
    },
    {
      "epoch": 0.02061379893854319,
      "grad_norm": 35.0,
      "learning_rate": 2.9690793015921856e-06,
      "loss": 0.2661,
      "mean_token_accuracy": 0.9641577005386353,
      "num_tokens": 121800.0,
      "step": 268
    },
    {
      "epoch": 0.020690716098761632,
      "grad_norm": 72.0,
      "learning_rate": 2.9689639258518574e-06,
      "loss": 0.2575,
      "mean_token_accuracy": 0.9721361994743347,
      "num_tokens": 122124.0,
      "step": 269
    },
    {
      "epoch": 0.020767633258980078,
      "grad_norm": 25.125,
      "learning_rate": 2.96884855011153e-06,
      "loss": 0.212,
      "mean_token_accuracy": 0.96875,
      "num_tokens": 122448.0,
      "step": 270
    },
    {
      "epoch": 0.020844550419198523,
      "grad_norm": 26.25,
      "learning_rate": 2.9687331743712024e-06,
      "loss": 0.2393,
      "mean_token_accuracy": 0.9673202633857727,
      "num_tokens": 122756.0,
      "step": 271
    },
    {
      "epoch": 0.02092146757941697,
      "grad_norm": 37.5,
      "learning_rate": 2.9686177986308746e-06,
      "loss": 0.2509,
      "mean_token_accuracy": 0.9675324559211731,
      "num_tokens": 123066.0,
      "step": 272
    },
    {
      "epoch": 0.020998384739635414,
      "grad_norm": 12.125,
      "learning_rate": 2.968502422890547e-06,
      "loss": 0.1252,
      "mean_token_accuracy": 0.9852670431137085,
      "num_tokens": 123610.0,
      "step": 273
    },
    {
      "epoch": 0.021075301899853856,
      "grad_norm": 13.1875,
      "learning_rate": 2.968387047150219e-06,
      "loss": 0.0989,
      "mean_token_accuracy": 0.9898219108581543,
      "num_tokens": 124398.0,
      "step": 274
    },
    {
      "epoch": 0.0211522190600723,
      "grad_norm": 39.5,
      "learning_rate": 2.9682716714098914e-06,
      "loss": 0.2444,
      "mean_token_accuracy": 0.9714285731315613,
      "num_tokens": 124714.0,
      "step": 275
    },
    {
      "epoch": 0.021229136220290747,
      "grad_norm": 17.625,
      "learning_rate": 2.968156295669564e-06,
      "loss": 0.1389,
      "mean_token_accuracy": 0.9845260977745056,
      "num_tokens": 125232.0,
      "step": 276
    },
    {
      "epoch": 0.021306053380509192,
      "grad_norm": 47.75,
      "learning_rate": 2.9680409199292363e-06,
      "loss": 0.2393,
      "mean_token_accuracy": 0.9675516486167908,
      "num_tokens": 125572.0,
      "step": 277
    },
    {
      "epoch": 0.021382970540727637,
      "grad_norm": 32.0,
      "learning_rate": 2.9679255441889086e-06,
      "loss": 0.2661,
      "mean_token_accuracy": 0.9723183512687683,
      "num_tokens": 125862.0,
      "step": 278
    },
    {
      "epoch": 0.02145988770094608,
      "grad_norm": 81.5,
      "learning_rate": 2.9678101684485813e-06,
      "loss": 0.2389,
      "mean_token_accuracy": 0.9688473343849182,
      "num_tokens": 126184.0,
      "step": 279
    },
    {
      "epoch": 0.021536804861164525,
      "grad_norm": 142.0,
      "learning_rate": 2.967694792708253e-06,
      "loss": 0.2348,
      "mean_token_accuracy": 0.9694189429283142,
      "num_tokens": 126512.0,
      "step": 280
    },
    {
      "epoch": 0.02161372202138297,
      "grad_norm": 83.5,
      "learning_rate": 2.9675794169679254e-06,
      "loss": 0.2435,
      "mean_token_accuracy": 0.970099687576294,
      "num_tokens": 126814.0,
      "step": 281
    },
    {
      "epoch": 0.021690639181601416,
      "grad_norm": 35.25,
      "learning_rate": 2.967464041227598e-06,
      "loss": 0.2541,
      "mean_token_accuracy": 0.9731800556182861,
      "num_tokens": 127076.0,
      "step": 282
    },
    {
      "epoch": 0.02176755634181986,
      "grad_norm": 7.625,
      "learning_rate": 2.9673486654872703e-06,
      "loss": 0.1214,
      "mean_token_accuracy": 0.9872029423713684,
      "num_tokens": 127624.0,
      "step": 283
    },
    {
      "epoch": 0.021844473502038306,
      "grad_norm": 22.625,
      "learning_rate": 2.9672332897469426e-06,
      "loss": 0.125,
      "mean_token_accuracy": 0.980461835861206,
      "num_tokens": 128188.0,
      "step": 284
    },
    {
      "epoch": 0.02192139066225675,
      "grad_norm": 10.25,
      "learning_rate": 2.967117914006615e-06,
      "loss": 0.0661,
      "mean_token_accuracy": 0.9921337366104126,
      "num_tokens": 129208.0,
      "step": 285
    },
    {
      "epoch": 0.021998307822475194,
      "grad_norm": 90.5,
      "learning_rate": 2.967002538266287e-06,
      "loss": 0.2196,
      "mean_token_accuracy": 0.9724770784378052,
      "num_tokens": 129536.0,
      "step": 286
    },
    {
      "epoch": 0.02207522498269364,
      "grad_norm": 57.0,
      "learning_rate": 2.9668871625259598e-06,
      "loss": 0.2278,
      "mean_token_accuracy": 0.9723076820373535,
      "num_tokens": 129862.0,
      "step": 287
    },
    {
      "epoch": 0.022152142142912085,
      "grad_norm": 80.5,
      "learning_rate": 2.966771786785632e-06,
      "loss": 0.225,
      "mean_token_accuracy": 0.9721361994743347,
      "num_tokens": 130186.0,
      "step": 288
    },
    {
      "epoch": 0.02222905930313053,
      "grad_norm": 15.3125,
      "learning_rate": 2.9666564110453043e-06,
      "loss": 0.1264,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 130730.0,
      "step": 289
    },
    {
      "epoch": 0.022305976463348972,
      "grad_norm": 13.125,
      "learning_rate": 2.966541035304977e-06,
      "loss": 0.1159,
      "mean_token_accuracy": 0.9860869646072388,
      "num_tokens": 131306.0,
      "step": 290
    },
    {
      "epoch": 0.022382893623567417,
      "grad_norm": 5.375,
      "learning_rate": 2.9664256595646488e-06,
      "loss": 0.0799,
      "mean_token_accuracy": 0.9925558567047119,
      "num_tokens": 132114.0,
      "step": 291
    },
    {
      "epoch": 0.022459810783785863,
      "grad_norm": 9.75,
      "learning_rate": 2.966310283824321e-06,
      "loss": 0.1215,
      "mean_token_accuracy": 0.9874776601791382,
      "num_tokens": 132674.0,
      "step": 292
    },
    {
      "epoch": 0.022536727944004308,
      "grad_norm": 40.25,
      "learning_rate": 2.9661949080839937e-06,
      "loss": 0.2674,
      "mean_token_accuracy": 0.9681274890899658,
      "num_tokens": 132926.0,
      "step": 293
    },
    {
      "epoch": 0.022613645104222754,
      "grad_norm": 27.125,
      "learning_rate": 2.966079532343666e-06,
      "loss": 0.2254,
      "mean_token_accuracy": 0.9665551781654358,
      "num_tokens": 133228.0,
      "step": 294
    },
    {
      "epoch": 0.022690562264441196,
      "grad_norm": 12.875,
      "learning_rate": 2.9659641566033382e-06,
      "loss": 0.1202,
      "mean_token_accuracy": 0.9857904314994812,
      "num_tokens": 133792.0,
      "step": 295
    },
    {
      "epoch": 0.02276747942465964,
      "grad_norm": 34.5,
      "learning_rate": 2.9658487808630105e-06,
      "loss": 0.2508,
      "mean_token_accuracy": 0.9688715934753418,
      "num_tokens": 134050.0,
      "step": 296
    },
    {
      "epoch": 0.022844396584878086,
      "grad_norm": 7.9375,
      "learning_rate": 2.9657334051226828e-06,
      "loss": 0.0617,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 135094.0,
      "step": 297
    },
    {
      "epoch": 0.022921313745096532,
      "grad_norm": 31.375,
      "learning_rate": 2.9656180293823554e-06,
      "loss": 0.2273,
      "mean_token_accuracy": 0.9696969985961914,
      "num_tokens": 135392.0,
      "step": 298
    },
    {
      "epoch": 0.022998230905314977,
      "grad_norm": 14.5,
      "learning_rate": 2.9655026536420277e-06,
      "loss": 0.1286,
      "mean_token_accuracy": 0.9850467443466187,
      "num_tokens": 135928.0,
      "step": 299
    },
    {
      "epoch": 0.02307514806553342,
      "grad_norm": 26.625,
      "learning_rate": 2.9653872779017e-06,
      "loss": 0.2378,
      "mean_token_accuracy": 0.971222996711731,
      "num_tokens": 136208.0,
      "step": 300
    },
    {
      "epoch": 0.023152065225751865,
      "grad_norm": 31.0,
      "learning_rate": 2.965271902161372e-06,
      "loss": 0.0795,
      "mean_token_accuracy": 0.9902676343917847,
      "num_tokens": 137032.0,
      "step": 301
    },
    {
      "epoch": 0.02322898238597031,
      "grad_norm": 8.5625,
      "learning_rate": 2.9651565264210445e-06,
      "loss": 0.1222,
      "mean_token_accuracy": 0.984644889831543,
      "num_tokens": 137554.0,
      "step": 302
    },
    {
      "epoch": 0.023305899546188755,
      "grad_norm": 26.125,
      "learning_rate": 2.9650411506807167e-06,
      "loss": 0.243,
      "mean_token_accuracy": 0.9678456783294678,
      "num_tokens": 137866.0,
      "step": 303
    },
    {
      "epoch": 0.0233828167064072,
      "grad_norm": 35.0,
      "learning_rate": 2.9649257749403894e-06,
      "loss": 0.2279,
      "mean_token_accuracy": 0.9682539701461792,
      "num_tokens": 138182.0,
      "step": 304
    },
    {
      "epoch": 0.023459733866625643,
      "grad_norm": 32.0,
      "learning_rate": 2.9648103992000617e-06,
      "loss": 0.2209,
      "mean_token_accuracy": 0.9742765426635742,
      "num_tokens": 138494.0,
      "step": 305
    },
    {
      "epoch": 0.023536651026844088,
      "grad_norm": 22.625,
      "learning_rate": 2.964695023459734e-06,
      "loss": 0.2361,
      "mean_token_accuracy": 0.974452555179596,
      "num_tokens": 138770.0,
      "step": 306
    },
    {
      "epoch": 0.023613568187062534,
      "grad_norm": 6.3125,
      "learning_rate": 2.964579647719406e-06,
      "loss": 0.1165,
      "mean_token_accuracy": 0.9885277152061462,
      "num_tokens": 139294.0,
      "step": 307
    },
    {
      "epoch": 0.02369048534728098,
      "grad_norm": 10.125,
      "learning_rate": 2.9644642719790784e-06,
      "loss": 0.0834,
      "mean_token_accuracy": 0.9897959232330322,
      "num_tokens": 140080.0,
      "step": 308
    },
    {
      "epoch": 0.023767402507499424,
      "grad_norm": 45.0,
      "learning_rate": 2.964348896238751e-06,
      "loss": 0.2065,
      "mean_token_accuracy": 0.9674556255340576,
      "num_tokens": 140420.0,
      "step": 309
    },
    {
      "epoch": 0.023844319667717866,
      "grad_norm": 14.3125,
      "learning_rate": 2.9642335204984234e-06,
      "loss": 0.117,
      "mean_token_accuracy": 0.9865546226501465,
      "num_tokens": 141016.0,
      "step": 310
    },
    {
      "epoch": 0.023921236827936312,
      "grad_norm": 44.25,
      "learning_rate": 2.9641181447580956e-06,
      "loss": 0.2023,
      "mean_token_accuracy": 0.9742120504379272,
      "num_tokens": 141366.0,
      "step": 311
    },
    {
      "epoch": 0.023998153988154757,
      "grad_norm": 35.0,
      "learning_rate": 2.964002769017768e-06,
      "loss": 0.2352,
      "mean_token_accuracy": 0.9635258316993713,
      "num_tokens": 141696.0,
      "step": 312
    },
    {
      "epoch": 0.024075071148373203,
      "grad_norm": 32.75,
      "learning_rate": 2.96388739327744e-06,
      "loss": 0.1942,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 142024.0,
      "step": 313
    },
    {
      "epoch": 0.024151988308591648,
      "grad_norm": 8.6875,
      "learning_rate": 2.9637720175371124e-06,
      "loss": 0.1155,
      "mean_token_accuracy": 0.9853211045265198,
      "num_tokens": 142570.0,
      "step": 314
    },
    {
      "epoch": 0.02422890546881009,
      "grad_norm": 27.75,
      "learning_rate": 2.963656641796785e-06,
      "loss": 0.2334,
      "mean_token_accuracy": 0.967391312122345,
      "num_tokens": 142848.0,
      "step": 315
    },
    {
      "epoch": 0.024305822629028535,
      "grad_norm": 23.75,
      "learning_rate": 2.9635412660564574e-06,
      "loss": 0.2084,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 143148.0,
      "step": 316
    },
    {
      "epoch": 0.02438273978924698,
      "grad_norm": 44.75,
      "learning_rate": 2.9634258903161296e-06,
      "loss": 0.1976,
      "mean_token_accuracy": 0.967930018901825,
      "num_tokens": 143492.0,
      "step": 317
    },
    {
      "epoch": 0.024459656949465426,
      "grad_norm": 9.3125,
      "learning_rate": 2.963310514575802e-06,
      "loss": 0.1076,
      "mean_token_accuracy": 0.9862778782844543,
      "num_tokens": 144076.0,
      "step": 318
    },
    {
      "epoch": 0.02453657410968387,
      "grad_norm": 7.28125,
      "learning_rate": 2.963195138835474e-06,
      "loss": 0.1105,
      "mean_token_accuracy": 0.9855855703353882,
      "num_tokens": 144632.0,
      "step": 319
    },
    {
      "epoch": 0.024613491269902314,
      "grad_norm": 20.625,
      "learning_rate": 2.963079763095147e-06,
      "loss": 0.194,
      "mean_token_accuracy": 0.9749216437339783,
      "num_tokens": 144952.0,
      "step": 320
    },
    {
      "epoch": 0.02469040843012076,
      "grad_norm": 16.0,
      "learning_rate": 2.962964387354819e-06,
      "loss": 0.1243,
      "mean_token_accuracy": 0.984644889831543,
      "num_tokens": 145474.0,
      "step": 321
    },
    {
      "epoch": 0.024767325590339204,
      "grad_norm": 70.5,
      "learning_rate": 2.9628490116144913e-06,
      "loss": 0.2082,
      "mean_token_accuracy": 0.9699699878692627,
      "num_tokens": 145808.0,
      "step": 322
    },
    {
      "epoch": 0.02484424275055765,
      "grad_norm": 40.0,
      "learning_rate": 2.9627336358741636e-06,
      "loss": 0.2356,
      "mean_token_accuracy": 0.9679715037345886,
      "num_tokens": 146090.0,
      "step": 323
    },
    {
      "epoch": 0.024921159910776095,
      "grad_norm": 13.125,
      "learning_rate": 2.962618260133836e-06,
      "loss": 0.1134,
      "mean_token_accuracy": 0.9857397675514221,
      "num_tokens": 146652.0,
      "step": 324
    },
    {
      "epoch": 0.02499807707099454,
      "grad_norm": 12.0625,
      "learning_rate": 2.962502884393508e-06,
      "loss": 0.1213,
      "mean_token_accuracy": 0.9845857620239258,
      "num_tokens": 147172.0,
      "step": 325
    },
    {
      "epoch": 0.025074994231212983,
      "grad_norm": 17.75,
      "learning_rate": 2.9623875086531808e-06,
      "loss": 0.1177,
      "mean_token_accuracy": 0.982425332069397,
      "num_tokens": 147742.0,
      "step": 326
    },
    {
      "epoch": 0.025151911391431428,
      "grad_norm": 21.875,
      "learning_rate": 2.962272132912853e-06,
      "loss": 0.1163,
      "mean_token_accuracy": 0.9865642786026001,
      "num_tokens": 148264.0,
      "step": 327
    },
    {
      "epoch": 0.025228828551649873,
      "grad_norm": 42.25,
      "learning_rate": 2.9621567571725253e-06,
      "loss": 0.1997,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 148568.0,
      "step": 328
    },
    {
      "epoch": 0.02530574571186832,
      "grad_norm": 25.625,
      "learning_rate": 2.9620413814321976e-06,
      "loss": 0.2265,
      "mean_token_accuracy": 0.9704797267913818,
      "num_tokens": 148840.0,
      "step": 329
    },
    {
      "epoch": 0.025382662872086764,
      "grad_norm": 6.3125,
      "learning_rate": 2.96192600569187e-06,
      "loss": 0.0748,
      "mean_token_accuracy": 0.989924430847168,
      "num_tokens": 149636.0,
      "step": 330
    },
    {
      "epoch": 0.025459580032305206,
      "grad_norm": 27.5,
      "learning_rate": 2.961810629951542e-06,
      "loss": 0.1917,
      "mean_token_accuracy": 0.9757575988769531,
      "num_tokens": 149968.0,
      "step": 331
    },
    {
      "epoch": 0.02553649719252365,
      "grad_norm": 23.625,
      "learning_rate": 2.9616952542112148e-06,
      "loss": 0.1952,
      "mean_token_accuracy": 0.9772079586982727,
      "num_tokens": 150320.0,
      "step": 332
    },
    {
      "epoch": 0.025613414352742097,
      "grad_norm": 26.125,
      "learning_rate": 2.961579878470887e-06,
      "loss": 0.2131,
      "mean_token_accuracy": 0.9639344215393066,
      "num_tokens": 150626.0,
      "step": 333
    },
    {
      "epoch": 0.025690331512960542,
      "grad_norm": 6.40625,
      "learning_rate": 2.9614645027305593e-06,
      "loss": 0.0746,
      "mean_token_accuracy": 0.9926650524139404,
      "num_tokens": 151446.0,
      "step": 334
    },
    {
      "epoch": 0.025767248673178988,
      "grad_norm": 28.125,
      "learning_rate": 2.9613491269902315e-06,
      "loss": 0.1849,
      "mean_token_accuracy": 0.976401150226593,
      "num_tokens": 151786.0,
      "step": 335
    },
    {
      "epoch": 0.02584416583339743,
      "grad_norm": 6.1875,
      "learning_rate": 2.9612337512499038e-06,
      "loss": 0.103,
      "mean_token_accuracy": 0.9861830472946167,
      "num_tokens": 152366.0,
      "step": 336
    },
    {
      "epoch": 0.025921082993615875,
      "grad_norm": 24.75,
      "learning_rate": 2.9611183755095765e-06,
      "loss": 0.2224,
      "mean_token_accuracy": 0.9686411023139954,
      "num_tokens": 152654.0,
      "step": 337
    },
    {
      "epoch": 0.02599800015383432,
      "grad_norm": 26.25,
      "learning_rate": 2.9610029997692487e-06,
      "loss": 0.2312,
      "mean_token_accuracy": 0.9686274528503418,
      "num_tokens": 152910.0,
      "step": 338
    },
    {
      "epoch": 0.026074917314052766,
      "grad_norm": 14.4375,
      "learning_rate": 2.9608876240289206e-06,
      "loss": 0.1882,
      "mean_token_accuracy": 0.9749216437339783,
      "num_tokens": 153230.0,
      "step": 339
    },
    {
      "epoch": 0.02615183447427121,
      "grad_norm": 5.84375,
      "learning_rate": 2.9607722482885932e-06,
      "loss": 0.1011,
      "mean_token_accuracy": 0.9861830472946167,
      "num_tokens": 153810.0,
      "step": 340
    },
    {
      "epoch": 0.026228751634489653,
      "grad_norm": 9.0,
      "learning_rate": 2.9606568725482655e-06,
      "loss": 0.1083,
      "mean_token_accuracy": 0.9850467443466187,
      "num_tokens": 154346.0,
      "step": 341
    },
    {
      "epoch": 0.0263056687947081,
      "grad_norm": 22.125,
      "learning_rate": 2.9605414968079378e-06,
      "loss": 0.2003,
      "mean_token_accuracy": 0.9775640964508057,
      "num_tokens": 154660.0,
      "step": 342
    },
    {
      "epoch": 0.026382585954926544,
      "grad_norm": 11.0,
      "learning_rate": 2.9604261210676104e-06,
      "loss": 0.1667,
      "mean_token_accuracy": 0.9793510437011719,
      "num_tokens": 155000.0,
      "step": 343
    },
    {
      "epoch": 0.02645950311514499,
      "grad_norm": 6.28125,
      "learning_rate": 2.9603107453272827e-06,
      "loss": 0.1032,
      "mean_token_accuracy": 0.9874326586723328,
      "num_tokens": 155558.0,
      "step": 344
    },
    {
      "epoch": 0.026536420275363435,
      "grad_norm": 8.5625,
      "learning_rate": 2.960195369586955e-06,
      "loss": 0.1076,
      "mean_token_accuracy": 0.9856887459754944,
      "num_tokens": 156118.0,
      "step": 345
    },
    {
      "epoch": 0.026613337435581877,
      "grad_norm": 23.375,
      "learning_rate": 2.9600799938466272e-06,
      "loss": 0.1872,
      "mean_token_accuracy": 0.9752321839332581,
      "num_tokens": 156442.0,
      "step": 346
    },
    {
      "epoch": 0.026690254595800322,
      "grad_norm": 19.5,
      "learning_rate": 2.9599646181062995e-06,
      "loss": 0.1941,
      "mean_token_accuracy": 0.9770491719245911,
      "num_tokens": 156748.0,
      "step": 347
    },
    {
      "epoch": 0.026767171756018768,
      "grad_norm": 5.625,
      "learning_rate": 2.959849242365972e-06,
      "loss": 0.0717,
      "mean_token_accuracy": 0.9912717938423157,
      "num_tokens": 157552.0,
      "step": 348
    },
    {
      "epoch": 0.026844088916237213,
      "grad_norm": 31.875,
      "learning_rate": 2.9597338666256444e-06,
      "loss": 0.1896,
      "mean_token_accuracy": 0.975683867931366,
      "num_tokens": 157882.0,
      "step": 349
    },
    {
      "epoch": 0.02692100607645566,
      "grad_norm": 6.5,
      "learning_rate": 2.9596184908853162e-06,
      "loss": 0.0693,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 158708.0,
      "step": 350
    },
    {
      "epoch": 0.0269979232366741,
      "grad_norm": 17.0,
      "learning_rate": 2.959503115144989e-06,
      "loss": 0.1677,
      "mean_token_accuracy": 0.976331353187561,
      "num_tokens": 159048.0,
      "step": 351
    },
    {
      "epoch": 0.027074840396892546,
      "grad_norm": 16.5,
      "learning_rate": 2.959387739404661e-06,
      "loss": 0.208,
      "mean_token_accuracy": 0.9752650260925293,
      "num_tokens": 159332.0,
      "step": 352
    },
    {
      "epoch": 0.02715175755711099,
      "grad_norm": 5.5625,
      "learning_rate": 2.9592723636643334e-06,
      "loss": 0.0996,
      "mean_token_accuracy": 0.9841827750205994,
      "num_tokens": 159902.0,
      "step": 353
    },
    {
      "epoch": 0.027228674717329437,
      "grad_norm": 22.875,
      "learning_rate": 2.959156987924006e-06,
      "loss": 0.19,
      "mean_token_accuracy": 0.9746031761169434,
      "num_tokens": 160218.0,
      "step": 354
    },
    {
      "epoch": 0.027305591877547882,
      "grad_norm": 23.25,
      "learning_rate": 2.9590416121836784e-06,
      "loss": 0.1833,
      "mean_token_accuracy": 0.9692307710647583,
      "num_tokens": 160544.0,
      "step": 355
    },
    {
      "epoch": 0.027382509037766324,
      "grad_norm": 7.3125,
      "learning_rate": 2.9589262364433506e-06,
      "loss": 0.0698,
      "mean_token_accuracy": 0.9902200698852539,
      "num_tokens": 161364.0,
      "step": 356
    },
    {
      "epoch": 0.02745942619798477,
      "grad_norm": 16.5,
      "learning_rate": 2.958810860703023e-06,
      "loss": 0.1816,
      "mean_token_accuracy": 0.9749216437339783,
      "num_tokens": 161684.0,
      "step": 357
    },
    {
      "epoch": 0.027536343358203215,
      "grad_norm": 25.875,
      "learning_rate": 2.958695484962695e-06,
      "loss": 0.1995,
      "mean_token_accuracy": 0.9698996543884277,
      "num_tokens": 161984.0,
      "step": 358
    },
    {
      "epoch": 0.02761326051842166,
      "grad_norm": 10.625,
      "learning_rate": 2.958580109222368e-06,
      "loss": 0.104,
      "mean_token_accuracy": 0.9817850589752197,
      "num_tokens": 162534.0,
      "step": 359
    },
    {
      "epoch": 0.027690177678640106,
      "grad_norm": 22.0,
      "learning_rate": 2.95846473348204e-06,
      "loss": 0.2212,
      "mean_token_accuracy": 0.9725490212440491,
      "num_tokens": 162790.0,
      "step": 360
    },
    {
      "epoch": 0.027767094838858548,
      "grad_norm": 20.5,
      "learning_rate": 2.958349357741712e-06,
      "loss": 0.2063,
      "mean_token_accuracy": 0.9704797267913818,
      "num_tokens": 163062.0,
      "step": 361
    },
    {
      "epoch": 0.027844011999076993,
      "grad_norm": 11.6875,
      "learning_rate": 2.9582339820013846e-06,
      "loss": 0.0999,
      "mean_token_accuracy": 0.9842382073402405,
      "num_tokens": 163634.0,
      "step": 362
    },
    {
      "epoch": 0.02792092915929544,
      "grad_norm": 9.4375,
      "learning_rate": 2.958118606261057e-06,
      "loss": 0.1115,
      "mean_token_accuracy": 0.9866156578063965,
      "num_tokens": 164158.0,
      "step": 363
    },
    {
      "epoch": 0.027997846319513884,
      "grad_norm": 15.875,
      "learning_rate": 2.958003230520729e-06,
      "loss": 0.177,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 164486.0,
      "step": 364
    },
    {
      "epoch": 0.02807476347973233,
      "grad_norm": 6.8125,
      "learning_rate": 2.957887854780402e-06,
      "loss": 0.0986,
      "mean_token_accuracy": 0.987522304058075,
      "num_tokens": 165048.0,
      "step": 365
    },
    {
      "epoch": 0.02815168063995077,
      "grad_norm": 17.75,
      "learning_rate": 2.9577724790400736e-06,
      "loss": 0.2046,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 165306.0,
      "step": 366
    },
    {
      "epoch": 0.028228597800169217,
      "grad_norm": 23.0,
      "learning_rate": 2.9576571032997463e-06,
      "loss": 0.1857,
      "mean_token_accuracy": 0.9755351543426514,
      "num_tokens": 165634.0,
      "step": 367
    },
    {
      "epoch": 0.028305514960387662,
      "grad_norm": 22.375,
      "learning_rate": 2.9575417275594186e-06,
      "loss": 0.1951,
      "mean_token_accuracy": 0.9716312289237976,
      "num_tokens": 165918.0,
      "step": 368
    },
    {
      "epoch": 0.028382432120606108,
      "grad_norm": 11.0625,
      "learning_rate": 2.957426351819091e-06,
      "loss": 0.1038,
      "mean_token_accuracy": 0.9860383868217468,
      "num_tokens": 166492.0,
      "step": 369
    },
    {
      "epoch": 0.028459349280824553,
      "grad_norm": 16.375,
      "learning_rate": 2.9573109760787635e-06,
      "loss": 0.1823,
      "mean_token_accuracy": 0.9719626307487488,
      "num_tokens": 166814.0,
      "step": 370
    },
    {
      "epoch": 0.028536266441043,
      "grad_norm": 17.25,
      "learning_rate": 2.9571956003384358e-06,
      "loss": 0.1632,
      "mean_token_accuracy": 0.9759759902954102,
      "num_tokens": 167150.0,
      "step": 371
    },
    {
      "epoch": 0.02861318360126144,
      "grad_norm": 18.125,
      "learning_rate": 2.9570802245981076e-06,
      "loss": 0.175,
      "mean_token_accuracy": 0.9789789915084839,
      "num_tokens": 167484.0,
      "step": 372
    },
    {
      "epoch": 0.028690100761479886,
      "grad_norm": 10.4375,
      "learning_rate": 2.9569648488577803e-06,
      "loss": 0.1546,
      "mean_token_accuracy": 0.9827089309692383,
      "num_tokens": 167832.0,
      "step": 373
    },
    {
      "epoch": 0.02876701792169833,
      "grad_norm": 22.125,
      "learning_rate": 2.9568494731174526e-06,
      "loss": 0.1903,
      "mean_token_accuracy": 0.9739413857460022,
      "num_tokens": 168140.0,
      "step": 374
    },
    {
      "epoch": 0.028843935081916777,
      "grad_norm": 10.75,
      "learning_rate": 2.956734097377125e-06,
      "loss": 0.11,
      "mean_token_accuracy": 0.9845260977745056,
      "num_tokens": 168658.0,
      "step": 375
    },
    {
      "epoch": 0.028920852242135222,
      "grad_norm": 10.375,
      "learning_rate": 2.9566187216367975e-06,
      "loss": 0.1643,
      "mean_token_accuracy": 0.9815384745597839,
      "num_tokens": 168984.0,
      "step": 376
    },
    {
      "epoch": 0.028997769402353664,
      "grad_norm": 8.0625,
      "learning_rate": 2.9565033458964693e-06,
      "loss": 0.0954,
      "mean_token_accuracy": 0.9863247871398926,
      "num_tokens": 169570.0,
      "step": 377
    },
    {
      "epoch": 0.02907468656257211,
      "grad_norm": 8.5,
      "learning_rate": 2.956387970156142e-06,
      "loss": 0.1536,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 169908.0,
      "step": 378
    },
    {
      "epoch": 0.029151603722790555,
      "grad_norm": 33.75,
      "learning_rate": 2.9562725944158143e-06,
      "loss": 0.1799,
      "mean_token_accuracy": 0.975683867931366,
      "num_tokens": 170238.0,
      "step": 379
    },
    {
      "epoch": 0.029228520883009,
      "grad_norm": 9.375,
      "learning_rate": 2.9561572186754865e-06,
      "loss": 0.0924,
      "mean_token_accuracy": 0.9865546226501465,
      "num_tokens": 170834.0,
      "step": 380
    },
    {
      "epoch": 0.029305438043227446,
      "grad_norm": 7.15625,
      "learning_rate": 2.956041842935159e-06,
      "loss": 0.0706,
      "mean_token_accuracy": 0.9900249242782593,
      "num_tokens": 171638.0,
      "step": 381
    },
    {
      "epoch": 0.029382355203445887,
      "grad_norm": 17.625,
      "learning_rate": 2.9559264671948315e-06,
      "loss": 0.198,
      "mean_token_accuracy": 0.9752650260925293,
      "num_tokens": 171922.0,
      "step": 382
    },
    {
      "epoch": 0.029459272363664333,
      "grad_norm": 6.09375,
      "learning_rate": 2.9558110914545033e-06,
      "loss": 0.0672,
      "mean_token_accuracy": 0.9913151264190674,
      "num_tokens": 172730.0,
      "step": 383
    },
    {
      "epoch": 0.02953618952388278,
      "grad_norm": 16.0,
      "learning_rate": 2.955695715714176e-06,
      "loss": 0.1659,
      "mean_token_accuracy": 0.9755351543426514,
      "num_tokens": 173058.0,
      "step": 384
    },
    {
      "epoch": 0.029613106684101224,
      "grad_norm": 19.0,
      "learning_rate": 2.9555803399738482e-06,
      "loss": 0.1571,
      "mean_token_accuracy": 0.9732937812805176,
      "num_tokens": 173396.0,
      "step": 385
    },
    {
      "epoch": 0.02969002384431967,
      "grad_norm": 16.0,
      "learning_rate": 2.9554649642335205e-06,
      "loss": 0.1646,
      "mean_token_accuracy": 0.9755351543426514,
      "num_tokens": 173724.0,
      "step": 386
    },
    {
      "epoch": 0.02976694100453811,
      "grad_norm": 19.125,
      "learning_rate": 2.955349588493193e-06,
      "loss": 0.1547,
      "mean_token_accuracy": 0.9766764044761658,
      "num_tokens": 174068.0,
      "step": 387
    },
    {
      "epoch": 0.029843858164756557,
      "grad_norm": 18.625,
      "learning_rate": 2.955234212752865e-06,
      "loss": 0.1774,
      "mean_token_accuracy": 0.9738562107086182,
      "num_tokens": 174376.0,
      "step": 388
    },
    {
      "epoch": 0.029920775324975002,
      "grad_norm": 11.6875,
      "learning_rate": 2.9551188370125377e-06,
      "loss": 0.1157,
      "mean_token_accuracy": 0.9837728142738342,
      "num_tokens": 174870.0,
      "step": 389
    },
    {
      "epoch": 0.029997692485193447,
      "grad_norm": 9.0625,
      "learning_rate": 2.95500346127221e-06,
      "loss": 0.1054,
      "mean_token_accuracy": 0.984644889831543,
      "num_tokens": 175392.0,
      "step": 390
    },
    {
      "epoch": 0.030074609645411893,
      "grad_norm": 5.1875,
      "learning_rate": 2.9548880855318822e-06,
      "loss": 0.0888,
      "mean_token_accuracy": 0.9863247871398926,
      "num_tokens": 175978.0,
      "step": 391
    },
    {
      "epoch": 0.030151526805630335,
      "grad_norm": 30.625,
      "learning_rate": 2.9547727097915545e-06,
      "loss": 0.218,
      "mean_token_accuracy": 0.9688715934753418,
      "num_tokens": 176236.0,
      "step": 392
    },
    {
      "epoch": 0.03022844396584878,
      "grad_norm": 5.5625,
      "learning_rate": 2.954657334051227e-06,
      "loss": 0.0932,
      "mean_token_accuracy": 0.9858906269073486,
      "num_tokens": 176804.0,
      "step": 393
    },
    {
      "epoch": 0.030305361126067226,
      "grad_norm": 13.1875,
      "learning_rate": 2.954541958310899e-06,
      "loss": 0.1524,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 177150.0,
      "step": 394
    },
    {
      "epoch": 0.03038227828628567,
      "grad_norm": 20.25,
      "learning_rate": 2.9544265825705717e-06,
      "loss": 0.1923,
      "mean_token_accuracy": 0.9715302586555481,
      "num_tokens": 177432.0,
      "step": 395
    },
    {
      "epoch": 0.030459195446504116,
      "grad_norm": 7.9375,
      "learning_rate": 2.954311206830244e-06,
      "loss": 0.0987,
      "mean_token_accuracy": 0.9853211045265198,
      "num_tokens": 177978.0,
      "step": 396
    },
    {
      "epoch": 0.03053611260672256,
      "grad_norm": 5.78125,
      "learning_rate": 2.954195831089916e-06,
      "loss": 0.0951,
      "mean_token_accuracy": 0.9870610237121582,
      "num_tokens": 178520.0,
      "step": 397
    },
    {
      "epoch": 0.030613029766941004,
      "grad_norm": 19.375,
      "learning_rate": 2.954080455349589e-06,
      "loss": 0.1593,
      "mean_token_accuracy": 0.9759036302566528,
      "num_tokens": 178854.0,
      "step": 398
    },
    {
      "epoch": 0.03068994692715945,
      "grad_norm": 8.5,
      "learning_rate": 2.9539650796092607e-06,
      "loss": 0.0927,
      "mean_token_accuracy": 0.9877836108207703,
      "num_tokens": 179428.0,
      "step": 399
    },
    {
      "epoch": 0.030766864087377895,
      "grad_norm": 14.9375,
      "learning_rate": 2.953849703868933e-06,
      "loss": 0.181,
      "mean_token_accuracy": 0.9754385948181152,
      "num_tokens": 179714.0,
      "step": 400
    },
    {
      "epoch": 0.03084378124759634,
      "grad_norm": 13.1875,
      "learning_rate": 2.9537343281286056e-06,
      "loss": 0.1525,
      "mean_token_accuracy": 0.9770773649215698,
      "num_tokens": 180064.0,
      "step": 401
    },
    {
      "epoch": 0.030920698407814782,
      "grad_norm": 6.90625,
      "learning_rate": 2.953618952388278e-06,
      "loss": 0.0964,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 180608.0,
      "step": 402
    },
    {
      "epoch": 0.030997615568033227,
      "grad_norm": 15.0,
      "learning_rate": 2.95350357664795e-06,
      "loss": 0.1815,
      "mean_token_accuracy": 0.9750889539718628,
      "num_tokens": 180890.0,
      "step": 403
    },
    {
      "epoch": 0.031074532728251673,
      "grad_norm": 15.875,
      "learning_rate": 2.9533882009076224e-06,
      "loss": 0.1545,
      "mean_token_accuracy": 0.9737609624862671,
      "num_tokens": 181234.0,
      "step": 404
    },
    {
      "epoch": 0.031151449888470118,
      "grad_norm": 13.8125,
      "learning_rate": 2.9532728251672947e-06,
      "loss": 0.1639,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 181538.0,
      "step": 405
    },
    {
      "epoch": 0.031228367048688564,
      "grad_norm": 8.125,
      "learning_rate": 2.9531574494269674e-06,
      "loss": 0.0956,
      "mean_token_accuracy": 0.9872495532035828,
      "num_tokens": 182088.0,
      "step": 406
    },
    {
      "epoch": 0.031305284208907005,
      "grad_norm": 14.8125,
      "learning_rate": 2.9530420736866396e-06,
      "loss": 0.0941,
      "mean_token_accuracy": 0.980461835861206,
      "num_tokens": 182652.0,
      "step": 407
    },
    {
      "epoch": 0.03138220136912545,
      "grad_norm": 6.53125,
      "learning_rate": 2.952926697946312e-06,
      "loss": 0.0659,
      "mean_token_accuracy": 0.9889975786209106,
      "num_tokens": 183472.0,
      "step": 408
    },
    {
      "epoch": 0.031459118529343896,
      "grad_norm": 10.9375,
      "learning_rate": 2.9528113222059846e-06,
      "loss": 0.1657,
      "mean_token_accuracy": 0.9739413857460022,
      "num_tokens": 183780.0,
      "step": 409
    },
    {
      "epoch": 0.03153603568956234,
      "grad_norm": 11.375,
      "learning_rate": 2.9526959464656564e-06,
      "loss": 0.1608,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 184088.0,
      "step": 410
    },
    {
      "epoch": 0.03161295284978079,
      "grad_norm": 12.8125,
      "learning_rate": 2.9525805707253286e-06,
      "loss": 0.1759,
      "mean_token_accuracy": 0.9726962447166443,
      "num_tokens": 184382.0,
      "step": 411
    },
    {
      "epoch": 0.03168987000999923,
      "grad_norm": 6.6875,
      "learning_rate": 2.9524651949850013e-06,
      "loss": 0.0864,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 184960.0,
      "step": 412
    },
    {
      "epoch": 0.03176678717021768,
      "grad_norm": 14.0625,
      "learning_rate": 2.9523498192446736e-06,
      "loss": 0.1643,
      "mean_token_accuracy": 0.9747633934020996,
      "num_tokens": 185278.0,
      "step": 413
    },
    {
      "epoch": 0.03184370433043612,
      "grad_norm": 6.34375,
      "learning_rate": 2.952234443504346e-06,
      "loss": 0.0639,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 186104.0,
      "step": 414
    },
    {
      "epoch": 0.03192062149065456,
      "grad_norm": 6.15625,
      "learning_rate": 2.952119067764018e-06,
      "loss": 0.0912,
      "mean_token_accuracy": 0.9872958064079285,
      "num_tokens": 186656.0,
      "step": 415
    },
    {
      "epoch": 0.03199753865087301,
      "grad_norm": 5.40625,
      "learning_rate": 2.9520036920236904e-06,
      "loss": 0.0645,
      "mean_token_accuracy": 0.9910485744476318,
      "num_tokens": 187440.0,
      "step": 416
    },
    {
      "epoch": 0.03207445581109145,
      "grad_norm": 12.0,
      "learning_rate": 2.951888316283363e-06,
      "loss": 0.1466,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 187778.0,
      "step": 417
    },
    {
      "epoch": 0.0321513729713099,
      "grad_norm": 9.125,
      "learning_rate": 2.9517729405430353e-06,
      "loss": 0.153,
      "mean_token_accuracy": 0.9784615635871887,
      "num_tokens": 188104.0,
      "step": 418
    },
    {
      "epoch": 0.032228290131528343,
      "grad_norm": 15.25,
      "learning_rate": 2.9516575648027076e-06,
      "loss": 0.1636,
      "mean_token_accuracy": 0.9746031761169434,
      "num_tokens": 188420.0,
      "step": 419
    },
    {
      "epoch": 0.03230520729174679,
      "grad_norm": 18.625,
      "learning_rate": 2.9515421890623802e-06,
      "loss": 0.1585,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 188748.0,
      "step": 420
    },
    {
      "epoch": 0.032382124451965234,
      "grad_norm": 7.71875,
      "learning_rate": 2.951426813322052e-06,
      "loss": 0.0813,
      "mean_token_accuracy": 0.9883138537406921,
      "num_tokens": 189348.0,
      "step": 421
    },
    {
      "epoch": 0.03245904161218368,
      "grad_norm": 11.0625,
      "learning_rate": 2.9513114375817243e-06,
      "loss": 0.1565,
      "mean_token_accuracy": 0.9781931638717651,
      "num_tokens": 189670.0,
      "step": 422
    },
    {
      "epoch": 0.032535958772402125,
      "grad_norm": 10.125,
      "learning_rate": 2.951196061841397e-06,
      "loss": 0.1667,
      "mean_token_accuracy": 0.9764309525489807,
      "num_tokens": 189968.0,
      "step": 423
    },
    {
      "epoch": 0.03261287593262057,
      "grad_norm": 10.375,
      "learning_rate": 2.9510806861010693e-06,
      "loss": 0.1535,
      "mean_token_accuracy": 0.9750778675079346,
      "num_tokens": 190290.0,
      "step": 424
    },
    {
      "epoch": 0.03268979309283901,
      "grad_norm": 22.625,
      "learning_rate": 2.9509653103607415e-06,
      "loss": 0.1948,
      "mean_token_accuracy": 0.9688715934753418,
      "num_tokens": 190548.0,
      "step": 425
    },
    {
      "epoch": 0.032766710253057454,
      "grad_norm": 19.5,
      "learning_rate": 2.950849934620414e-06,
      "loss": 0.1532,
      "mean_token_accuracy": 0.9694189429283142,
      "num_tokens": 190876.0,
      "step": 426
    },
    {
      "epoch": 0.0328436274132759,
      "grad_norm": 6.3125,
      "learning_rate": 2.950734558880086e-06,
      "loss": 0.0889,
      "mean_token_accuracy": 0.985637366771698,
      "num_tokens": 191434.0,
      "step": 427
    },
    {
      "epoch": 0.032920544573494345,
      "grad_norm": 7.5,
      "learning_rate": 2.9506191831397587e-06,
      "loss": 0.1589,
      "mean_token_accuracy": 0.9764309525489807,
      "num_tokens": 191732.0,
      "step": 428
    },
    {
      "epoch": 0.03299746173371279,
      "grad_norm": 15.3125,
      "learning_rate": 2.950503807399431e-06,
      "loss": 0.157,
      "mean_token_accuracy": 0.9768977165222168,
      "num_tokens": 192036.0,
      "step": 429
    },
    {
      "epoch": 0.033074378893931236,
      "grad_norm": 17.375,
      "learning_rate": 2.9503884316591032e-06,
      "loss": 0.1512,
      "mean_token_accuracy": 0.9741935729980469,
      "num_tokens": 192348.0,
      "step": 430
    },
    {
      "epoch": 0.03315129605414968,
      "grad_norm": 12.0,
      "learning_rate": 2.950273055918776e-06,
      "loss": 0.1576,
      "mean_token_accuracy": 0.9753845930099487,
      "num_tokens": 192674.0,
      "step": 431
    },
    {
      "epoch": 0.03322821321436813,
      "grad_norm": 12.5,
      "learning_rate": 2.9501576801784478e-06,
      "loss": 0.1581,
      "mean_token_accuracy": 0.9702970385551453,
      "num_tokens": 192978.0,
      "step": 432
    },
    {
      "epoch": 0.03330513037458657,
      "grad_norm": 18.875,
      "learning_rate": 2.95004230443812e-06,
      "loss": 0.1575,
      "mean_token_accuracy": 0.9743589758872986,
      "num_tokens": 193330.0,
      "step": 433
    },
    {
      "epoch": 0.03338204753480502,
      "grad_norm": 14.0625,
      "learning_rate": 2.9499269286977927e-06,
      "loss": 0.1498,
      "mean_token_accuracy": 0.9742120504379272,
      "num_tokens": 193680.0,
      "step": 434
    },
    {
      "epoch": 0.03345896469502346,
      "grad_norm": 10.8125,
      "learning_rate": 2.949811552957465e-06,
      "loss": 0.1511,
      "mean_token_accuracy": 0.9717868566513062,
      "num_tokens": 194000.0,
      "step": 435
    },
    {
      "epoch": 0.0335358818552419,
      "grad_norm": 8.25,
      "learning_rate": 2.9496961772171372e-06,
      "loss": 0.0838,
      "mean_token_accuracy": 0.9897785186767578,
      "num_tokens": 194588.0,
      "step": 436
    },
    {
      "epoch": 0.03361279901546035,
      "grad_norm": 6.96875,
      "learning_rate": 2.9495808014768095e-06,
      "loss": 0.0878,
      "mean_token_accuracy": 0.9881154298782349,
      "num_tokens": 195178.0,
      "step": 437
    },
    {
      "epoch": 0.03368971617567879,
      "grad_norm": 7.84375,
      "learning_rate": 2.9494654257364817e-06,
      "loss": 0.1559,
      "mean_token_accuracy": 0.9764309525489807,
      "num_tokens": 195476.0,
      "step": 438
    },
    {
      "epoch": 0.03376663333589724,
      "grad_norm": 13.9375,
      "learning_rate": 2.9493500499961544e-06,
      "loss": 0.1422,
      "mean_token_accuracy": 0.9759036302566528,
      "num_tokens": 195810.0,
      "step": 439
    },
    {
      "epoch": 0.03384355049611568,
      "grad_norm": 23.625,
      "learning_rate": 2.9492346742558267e-06,
      "loss": 0.1805,
      "mean_token_accuracy": 0.9725490212440491,
      "num_tokens": 196066.0,
      "step": 440
    },
    {
      "epoch": 0.03392046765633413,
      "grad_norm": 8.4375,
      "learning_rate": 2.949119298515499e-06,
      "loss": 0.0901,
      "mean_token_accuracy": 0.9852670431137085,
      "num_tokens": 196610.0,
      "step": 441
    },
    {
      "epoch": 0.033997384816552574,
      "grad_norm": 7.9375,
      "learning_rate": 2.949003922775171e-06,
      "loss": 0.0902,
      "mean_token_accuracy": 0.9868173003196716,
      "num_tokens": 197142.0,
      "step": 442
    },
    {
      "epoch": 0.03407430197677102,
      "grad_norm": 9.1875,
      "learning_rate": 2.9488885470348434e-06,
      "loss": 0.0673,
      "mean_token_accuracy": 0.9900249242782593,
      "num_tokens": 197946.0,
      "step": 443
    },
    {
      "epoch": 0.034151219136989465,
      "grad_norm": 6.78125,
      "learning_rate": 2.9487731712945157e-06,
      "loss": 0.1545,
      "mean_token_accuracy": 0.9764309525489807,
      "num_tokens": 198244.0,
      "step": 444
    },
    {
      "epoch": 0.03422813629720791,
      "grad_norm": 15.4375,
      "learning_rate": 2.9486577955541884e-06,
      "loss": 0.1424,
      "mean_token_accuracy": 0.9728915691375732,
      "num_tokens": 198578.0,
      "step": 445
    },
    {
      "epoch": 0.03430505345742635,
      "grad_norm": 8.6875,
      "learning_rate": 2.9485424198138606e-06,
      "loss": 0.088,
      "mean_token_accuracy": 0.9858906269073486,
      "num_tokens": 199146.0,
      "step": 446
    },
    {
      "epoch": 0.034381970617644794,
      "grad_norm": 19.875,
      "learning_rate": 2.948427044073533e-06,
      "loss": 0.1732,
      "mean_token_accuracy": 0.9684210419654846,
      "num_tokens": 199432.0,
      "step": 447
    },
    {
      "epoch": 0.03445888777786324,
      "grad_norm": 7.625,
      "learning_rate": 2.948311668333205e-06,
      "loss": 0.0839,
      "mean_token_accuracy": 0.9860383868217468,
      "num_tokens": 200006.0,
      "step": 448
    },
    {
      "epoch": 0.034535804938081685,
      "grad_norm": 25.25,
      "learning_rate": 2.9481962925928774e-06,
      "loss": 0.1892,
      "mean_token_accuracy": 0.9688715934753418,
      "num_tokens": 200264.0,
      "step": 449
    },
    {
      "epoch": 0.03461272209830013,
      "grad_norm": 13.375,
      "learning_rate": 2.9480809168525497e-06,
      "loss": 0.1757,
      "mean_token_accuracy": 0.9727626442909241,
      "num_tokens": 200522.0,
      "step": 450
    },
    {
      "epoch": 0.034689639258518576,
      "grad_norm": 13.625,
      "learning_rate": 2.9479655411122224e-06,
      "loss": 0.1503,
      "mean_token_accuracy": 0.9749216437339783,
      "num_tokens": 200842.0,
      "step": 451
    },
    {
      "epoch": 0.03476655641873702,
      "grad_norm": 11.8125,
      "learning_rate": 2.9478501653718946e-06,
      "loss": 0.1438,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 201170.0,
      "step": 452
    },
    {
      "epoch": 0.03484347357895547,
      "grad_norm": 8.5625,
      "learning_rate": 2.947734789631567e-06,
      "loss": 0.0826,
      "mean_token_accuracy": 0.9823633432388306,
      "num_tokens": 201738.0,
      "step": 453
    },
    {
      "epoch": 0.03492039073917391,
      "grad_norm": 18.25,
      "learning_rate": 2.947619413891239e-06,
      "loss": 0.1673,
      "mean_token_accuracy": 0.9713261723518372,
      "num_tokens": 202018.0,
      "step": 454
    },
    {
      "epoch": 0.03499730789939236,
      "grad_norm": 6.90625,
      "learning_rate": 2.9475040381509114e-06,
      "loss": 0.0804,
      "mean_token_accuracy": 0.9845626354217529,
      "num_tokens": 202602.0,
      "step": 455
    },
    {
      "epoch": 0.035074225059610796,
      "grad_norm": 4.5625,
      "learning_rate": 2.947388662410584e-06,
      "loss": 0.0591,
      "mean_token_accuracy": 0.9897959232330322,
      "num_tokens": 203388.0,
      "step": 456
    },
    {
      "epoch": 0.03515114221982924,
      "grad_norm": 9.8125,
      "learning_rate": 2.9472732866702563e-06,
      "loss": 0.0834,
      "mean_token_accuracy": 0.9858906269073486,
      "num_tokens": 203956.0,
      "step": 457
    },
    {
      "epoch": 0.03522805938004769,
      "grad_norm": 4.59375,
      "learning_rate": 2.9471579109299286e-06,
      "loss": 0.058,
      "mean_token_accuracy": 0.9897435903549194,
      "num_tokens": 204738.0,
      "step": 458
    },
    {
      "epoch": 0.03530497654026613,
      "grad_norm": 17.125,
      "learning_rate": 2.947042535189601e-06,
      "loss": 0.1642,
      "mean_token_accuracy": 0.9719298481941223,
      "num_tokens": 205024.0,
      "step": 459
    },
    {
      "epoch": 0.03538189370048458,
      "grad_norm": 4.84375,
      "learning_rate": 2.946927159449273e-06,
      "loss": 0.0812,
      "mean_token_accuracy": 0.9841827750205994,
      "num_tokens": 205594.0,
      "step": 460
    },
    {
      "epoch": 0.03545881086070302,
      "grad_norm": 12.9375,
      "learning_rate": 2.9468117837089454e-06,
      "loss": 0.1327,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 205940.0,
      "step": 461
    },
    {
      "epoch": 0.03553572802092147,
      "grad_norm": 19.125,
      "learning_rate": 2.946696407968618e-06,
      "loss": 0.1693,
      "mean_token_accuracy": 0.9679715037345886,
      "num_tokens": 206222.0,
      "step": 462
    },
    {
      "epoch": 0.035612645181139914,
      "grad_norm": 14.375,
      "learning_rate": 2.9465810322282903e-06,
      "loss": 0.1248,
      "mean_token_accuracy": 0.98525071144104,
      "num_tokens": 206562.0,
      "step": 463
    },
    {
      "epoch": 0.03568956234135836,
      "grad_norm": 10.0625,
      "learning_rate": 2.9464656564879626e-06,
      "loss": 0.1396,
      "mean_token_accuracy": 0.9810725450515747,
      "num_tokens": 206880.0,
      "step": 464
    },
    {
      "epoch": 0.035766479501576805,
      "grad_norm": 6.75,
      "learning_rate": 2.946350280747635e-06,
      "loss": 0.0592,
      "mean_token_accuracy": 0.9910025596618652,
      "num_tokens": 207660.0,
      "step": 465
    },
    {
      "epoch": 0.03584339666179524,
      "grad_norm": 19.75,
      "learning_rate": 2.946234905007307e-06,
      "loss": 0.169,
      "mean_token_accuracy": 0.9749103784561157,
      "num_tokens": 207940.0,
      "step": 466
    },
    {
      "epoch": 0.03592031382201369,
      "grad_norm": 21.0,
      "learning_rate": 2.9461195292669798e-06,
      "loss": 0.1726,
      "mean_token_accuracy": 0.970802903175354,
      "num_tokens": 208216.0,
      "step": 467
    },
    {
      "epoch": 0.035997230982232134,
      "grad_norm": 26.125,
      "learning_rate": 2.946004153526652e-06,
      "loss": 0.1564,
      "mean_token_accuracy": 0.971061110496521,
      "num_tokens": 208528.0,
      "step": 468
    },
    {
      "epoch": 0.03607414814245058,
      "grad_norm": 7.78125,
      "learning_rate": 2.945888777786324e-06,
      "loss": 0.0883,
      "mean_token_accuracy": 0.9864077568054199,
      "num_tokens": 209044.0,
      "step": 469
    },
    {
      "epoch": 0.036151065302669025,
      "grad_norm": 19.0,
      "learning_rate": 2.9457734020459965e-06,
      "loss": 0.1415,
      "mean_token_accuracy": 0.9781249761581421,
      "num_tokens": 209368.0,
      "step": 470
    },
    {
      "epoch": 0.03622798246288747,
      "grad_norm": 15.4375,
      "learning_rate": 2.945658026305669e-06,
      "loss": 0.147,
      "mean_token_accuracy": 0.970588207244873,
      "num_tokens": 209676.0,
      "step": 471
    },
    {
      "epoch": 0.036304899623105916,
      "grad_norm": 13.625,
      "learning_rate": 2.945542650565341e-06,
      "loss": 0.1586,
      "mean_token_accuracy": 0.9715302586555481,
      "num_tokens": 209958.0,
      "step": 472
    },
    {
      "epoch": 0.03638181678332436,
      "grad_norm": 14.3125,
      "learning_rate": 2.9454272748250137e-06,
      "loss": 0.1734,
      "mean_token_accuracy": 0.9688715934753418,
      "num_tokens": 210216.0,
      "step": 473
    },
    {
      "epoch": 0.036458733943542806,
      "grad_norm": 9.6875,
      "learning_rate": 2.945311899084686e-06,
      "loss": 0.082,
      "mean_token_accuracy": 0.9852670431137085,
      "num_tokens": 210760.0,
      "step": 474
    },
    {
      "epoch": 0.03653565110376125,
      "grad_norm": 11.125,
      "learning_rate": 2.9451965233443582e-06,
      "loss": 0.1569,
      "mean_token_accuracy": 0.9752650260925293,
      "num_tokens": 211044.0,
      "step": 475
    },
    {
      "epoch": 0.0366125682639797,
      "grad_norm": 9.75,
      "learning_rate": 2.9450811476040305e-06,
      "loss": 0.1685,
      "mean_token_accuracy": 0.9762845635414124,
      "num_tokens": 211298.0,
      "step": 476
    },
    {
      "epoch": 0.036689485424198136,
      "grad_norm": 6.9375,
      "learning_rate": 2.9449657718637028e-06,
      "loss": 0.0842,
      "mean_token_accuracy": 0.9873873591423035,
      "num_tokens": 211854.0,
      "step": 477
    },
    {
      "epoch": 0.03676640258441658,
      "grad_norm": 9.5625,
      "learning_rate": 2.9448503961233754e-06,
      "loss": 0.1557,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 212112.0,
      "step": 478
    },
    {
      "epoch": 0.03684331974463503,
      "grad_norm": 5.78125,
      "learning_rate": 2.9447350203830477e-06,
      "loss": 0.058,
      "mean_token_accuracy": 0.9910025596618652,
      "num_tokens": 212892.0,
      "step": 479
    },
    {
      "epoch": 0.03692023690485347,
      "grad_norm": 9.6875,
      "learning_rate": 2.9446196446427195e-06,
      "loss": 0.0859,
      "mean_token_accuracy": 0.9845626354217529,
      "num_tokens": 213476.0,
      "step": 480
    },
    {
      "epoch": 0.03699715406507192,
      "grad_norm": 5.78125,
      "learning_rate": 2.9445042689023922e-06,
      "loss": 0.0581,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 214262.0,
      "step": 481
    },
    {
      "epoch": 0.03707407122529036,
      "grad_norm": 16.875,
      "learning_rate": 2.9443888931620645e-06,
      "loss": 0.153,
      "mean_token_accuracy": 0.9732441306114197,
      "num_tokens": 214564.0,
      "step": 482
    },
    {
      "epoch": 0.03715098838550881,
      "grad_norm": 7.125,
      "learning_rate": 2.9442735174217367e-06,
      "loss": 0.0808,
      "mean_token_accuracy": 0.9854280352592468,
      "num_tokens": 215114.0,
      "step": 483
    },
    {
      "epoch": 0.037227905545727254,
      "grad_norm": 10.375,
      "learning_rate": 2.9441581416814094e-06,
      "loss": 0.0864,
      "mean_token_accuracy": 0.9850467443466187,
      "num_tokens": 215650.0,
      "step": 484
    },
    {
      "epoch": 0.0373048227059457,
      "grad_norm": 4.59375,
      "learning_rate": 2.9440427659410817e-06,
      "loss": 0.0549,
      "mean_token_accuracy": 0.9897435903549194,
      "num_tokens": 216432.0,
      "step": 485
    },
    {
      "epoch": 0.037381739866164145,
      "grad_norm": 6.21875,
      "learning_rate": 2.943927390200754e-06,
      "loss": 0.0787,
      "mean_token_accuracy": 0.9857904314994812,
      "num_tokens": 216996.0,
      "step": 486
    },
    {
      "epoch": 0.03745865702638258,
      "grad_norm": 16.625,
      "learning_rate": 2.943812014460426e-06,
      "loss": 0.1539,
      "mean_token_accuracy": 0.9796609878540039,
      "num_tokens": 217292.0,
      "step": 487
    },
    {
      "epoch": 0.03753557418660103,
      "grad_norm": 14.5625,
      "learning_rate": 2.9436966387200985e-06,
      "loss": 0.1409,
      "mean_token_accuracy": 0.975683867931366,
      "num_tokens": 217622.0,
      "step": 488
    },
    {
      "epoch": 0.037612491346819474,
      "grad_norm": 3.390625,
      "learning_rate": 2.943581262979771e-06,
      "loss": 0.0525,
      "mean_token_accuracy": 0.991094172000885,
      "num_tokens": 218410.0,
      "step": 489
    },
    {
      "epoch": 0.03768940850703792,
      "grad_norm": 18.625,
      "learning_rate": 2.9434658872394434e-06,
      "loss": 0.1567,
      "mean_token_accuracy": 0.9717868566513062,
      "num_tokens": 218730.0,
      "step": 490
    },
    {
      "epoch": 0.037766325667256365,
      "grad_norm": 13.1875,
      "learning_rate": 2.9433505114991152e-06,
      "loss": 0.1313,
      "mean_token_accuracy": 0.9742120504379272,
      "num_tokens": 219080.0,
      "step": 491
    },
    {
      "epoch": 0.03784324282747481,
      "grad_norm": 7.03125,
      "learning_rate": 2.943235135758788e-06,
      "loss": 0.0687,
      "mean_token_accuracy": 0.9883527159690857,
      "num_tokens": 219682.0,
      "step": 492
    },
    {
      "epoch": 0.037920159987693255,
      "grad_norm": 4.75,
      "learning_rate": 2.94311976001846e-06,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9950835704803467,
      "num_tokens": 220702.0,
      "step": 493
    },
    {
      "epoch": 0.0379970771479117,
      "grad_norm": 16.125,
      "learning_rate": 2.9430043842781324e-06,
      "loss": 0.1504,
      "mean_token_accuracy": 0.9749216437339783,
      "num_tokens": 221022.0,
      "step": 494
    },
    {
      "epoch": 0.038073994308130146,
      "grad_norm": 19.125,
      "learning_rate": 2.942889008537805e-06,
      "loss": 0.1543,
      "mean_token_accuracy": 0.9708737730979919,
      "num_tokens": 221332.0,
      "step": 495
    },
    {
      "epoch": 0.03815091146834859,
      "grad_norm": 22.625,
      "learning_rate": 2.9427736327974774e-06,
      "loss": 0.1707,
      "mean_token_accuracy": 0.9710144996643066,
      "num_tokens": 221610.0,
      "step": 496
    },
    {
      "epoch": 0.03822782862856703,
      "grad_norm": 8.75,
      "learning_rate": 2.9426582570571496e-06,
      "loss": 0.077,
      "mean_token_accuracy": 0.9879102110862732,
      "num_tokens": 222190.0,
      "step": 497
    },
    {
      "epoch": 0.038304745788785476,
      "grad_norm": 8.125,
      "learning_rate": 2.942542881316822e-06,
      "loss": 0.087,
      "mean_token_accuracy": 0.9845260977745056,
      "num_tokens": 222708.0,
      "step": 498
    },
    {
      "epoch": 0.03838166294900392,
      "grad_norm": 5.5625,
      "learning_rate": 2.942427505576494e-06,
      "loss": 0.0768,
      "mean_token_accuracy": 0.9874326586723328,
      "num_tokens": 223266.0,
      "step": 499
    },
    {
      "epoch": 0.038458580109222366,
      "grad_norm": 4.40625,
      "learning_rate": 2.9423121298361664e-06,
      "loss": 0.0754,
      "mean_token_accuracy": 0.9874326586723328,
      "num_tokens": 223824.0,
      "step": 500
    },
    {
      "epoch": 0.03853549726944081,
      "grad_norm": 4.90625,
      "learning_rate": 2.942196754095839e-06,
      "loss": 0.0517,
      "mean_token_accuracy": 0.988916277885437,
      "num_tokens": 224638.0,
      "step": 501
    },
    {
      "epoch": 0.03861241442965926,
      "grad_norm": 6.375,
      "learning_rate": 2.942081378355511e-06,
      "loss": 0.0829,
      "mean_token_accuracy": 0.9845260977745056,
      "num_tokens": 225156.0,
      "step": 502
    },
    {
      "epoch": 0.0386893315898777,
      "grad_norm": 3.578125,
      "learning_rate": 2.9419660026151836e-06,
      "loss": 0.0721,
      "mean_token_accuracy": 0.9858906269073486,
      "num_tokens": 225724.0,
      "step": 503
    },
    {
      "epoch": 0.03876624875009615,
      "grad_norm": 17.625,
      "learning_rate": 2.941850626874856e-06,
      "loss": 0.1712,
      "mean_token_accuracy": 0.9652509689331055,
      "num_tokens": 225984.0,
      "step": 504
    },
    {
      "epoch": 0.038843165910314593,
      "grad_norm": 10.75,
      "learning_rate": 2.941735251134528e-06,
      "loss": 0.1355,
      "mean_token_accuracy": 0.9777777791023254,
      "num_tokens": 226300.0,
      "step": 505
    },
    {
      "epoch": 0.03892008307053304,
      "grad_norm": 12.875,
      "learning_rate": 2.941619875394201e-06,
      "loss": 0.153,
      "mean_token_accuracy": 0.974452555179596,
      "num_tokens": 226576.0,
      "step": 506
    },
    {
      "epoch": 0.03899700023075148,
      "grad_norm": 21.0,
      "learning_rate": 2.9415044996538726e-06,
      "loss": 0.1601,
      "mean_token_accuracy": 0.9655172228813171,
      "num_tokens": 226838.0,
      "step": 507
    },
    {
      "epoch": 0.03907391739096992,
      "grad_norm": 7.46875,
      "learning_rate": 2.941389123913545e-06,
      "loss": 0.0832,
      "mean_token_accuracy": 0.9849340915679932,
      "num_tokens": 227370.0,
      "step": 508
    },
    {
      "epoch": 0.03915083455118837,
      "grad_norm": 12.0,
      "learning_rate": 2.9412737481732176e-06,
      "loss": 0.123,
      "mean_token_accuracy": 0.9795918464660645,
      "num_tokens": 227714.0,
      "step": 509
    },
    {
      "epoch": 0.039227751711406814,
      "grad_norm": 10.9375,
      "learning_rate": 2.94115837243289e-06,
      "loss": 0.1248,
      "mean_token_accuracy": 0.9755351543426514,
      "num_tokens": 228042.0,
      "step": 510
    },
    {
      "epoch": 0.03930466887162526,
      "grad_norm": 9.3125,
      "learning_rate": 2.941042996692562e-06,
      "loss": 0.1364,
      "mean_token_accuracy": 0.9773463010787964,
      "num_tokens": 228352.0,
      "step": 511
    },
    {
      "epoch": 0.039381586031843704,
      "grad_norm": 5.46875,
      "learning_rate": 2.9409276209522348e-06,
      "loss": 0.0786,
      "mean_token_accuracy": 0.9865642786026001,
      "num_tokens": 228874.0,
      "step": 512
    },
    {
      "epoch": 0.03945850319206215,
      "grad_norm": 8.875,
      "learning_rate": 2.9408122452119066e-06,
      "loss": 0.1277,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 229182.0,
      "step": 513
    },
    {
      "epoch": 0.039535420352280595,
      "grad_norm": 15.25,
      "learning_rate": 2.9406968694715793e-06,
      "loss": 0.132,
      "mean_token_accuracy": 0.9757575988769531,
      "num_tokens": 229514.0,
      "step": 514
    },
    {
      "epoch": 0.03961233751249904,
      "grad_norm": 11.125,
      "learning_rate": 2.9405814937312515e-06,
      "loss": 0.1299,
      "mean_token_accuracy": 0.9770491719245911,
      "num_tokens": 229820.0,
      "step": 515
    },
    {
      "epoch": 0.039689254672717486,
      "grad_norm": 13.5,
      "learning_rate": 2.940466117990924e-06,
      "loss": 0.1372,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 230120.0,
      "step": 516
    },
    {
      "epoch": 0.039766171832935925,
      "grad_norm": 7.90625,
      "learning_rate": 2.9403507422505965e-06,
      "loss": 0.0806,
      "mean_token_accuracy": 0.9845857620239258,
      "num_tokens": 230640.0,
      "step": 517
    },
    {
      "epoch": 0.03984308899315437,
      "grad_norm": 5.0625,
      "learning_rate": 2.9402353665102683e-06,
      "loss": 0.0718,
      "mean_token_accuracy": 0.9879931211471558,
      "num_tokens": 231224.0,
      "step": 518
    },
    {
      "epoch": 0.039920006153372815,
      "grad_norm": 4.25,
      "learning_rate": 2.9401199907699406e-06,
      "loss": 0.074,
      "mean_token_accuracy": 0.9871086478233337,
      "num_tokens": 231768.0,
      "step": 519
    },
    {
      "epoch": 0.03999692331359126,
      "grad_norm": 13.3125,
      "learning_rate": 2.9400046150296133e-06,
      "loss": 0.1404,
      "mean_token_accuracy": 0.9773463010787964,
      "num_tokens": 232078.0,
      "step": 520
    },
    {
      "epoch": 0.040073840473809706,
      "grad_norm": 12.875,
      "learning_rate": 2.9398892392892855e-06,
      "loss": 0.1189,
      "mean_token_accuracy": 0.9774011373519897,
      "num_tokens": 232434.0,
      "step": 521
    },
    {
      "epoch": 0.04015075763402815,
      "grad_norm": 14.75,
      "learning_rate": 2.9397738635489578e-06,
      "loss": 0.1459,
      "mean_token_accuracy": 0.9752650260925293,
      "num_tokens": 232718.0,
      "step": 522
    },
    {
      "epoch": 0.0402276747942466,
      "grad_norm": 5.71875,
      "learning_rate": 2.9396584878086304e-06,
      "loss": 0.0537,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 233544.0,
      "step": 523
    },
    {
      "epoch": 0.04030459195446504,
      "grad_norm": 14.6875,
      "learning_rate": 2.9395431120683023e-06,
      "loss": 0.1436,
      "mean_token_accuracy": 0.9710144996643066,
      "num_tokens": 233822.0,
      "step": 524
    },
    {
      "epoch": 0.04038150911468349,
      "grad_norm": 19.625,
      "learning_rate": 2.939427736327975e-06,
      "loss": 0.1501,
      "mean_token_accuracy": 0.9722222089767456,
      "num_tokens": 234112.0,
      "step": 525
    },
    {
      "epoch": 0.04045842627490193,
      "grad_norm": 7.625,
      "learning_rate": 2.9393123605876472e-06,
      "loss": 0.0799,
      "mean_token_accuracy": 0.9838383793830872,
      "num_tokens": 234608.0,
      "step": 526
    },
    {
      "epoch": 0.04053534343512038,
      "grad_norm": 12.125,
      "learning_rate": 2.9391969848473195e-06,
      "loss": 0.1165,
      "mean_token_accuracy": 0.9750000238418579,
      "num_tokens": 234970.0,
      "step": 527
    },
    {
      "epoch": 0.04061226059533882,
      "grad_norm": 11.8125,
      "learning_rate": 2.939081609106992e-06,
      "loss": 0.1143,
      "mean_token_accuracy": 0.9745762944221497,
      "num_tokens": 235326.0,
      "step": 528
    },
    {
      "epoch": 0.04068917775555726,
      "grad_norm": 10.5,
      "learning_rate": 2.938966233366664e-06,
      "loss": 0.1275,
      "mean_token_accuracy": 0.9717868566513062,
      "num_tokens": 235646.0,
      "step": 529
    },
    {
      "epoch": 0.04076609491577571,
      "grad_norm": 14.3125,
      "learning_rate": 2.9388508576263363e-06,
      "loss": 0.1279,
      "mean_token_accuracy": 0.9743589758872986,
      "num_tokens": 235960.0,
      "step": 530
    },
    {
      "epoch": 0.04084301207599415,
      "grad_norm": 17.0,
      "learning_rate": 2.938735481886009e-06,
      "loss": 0.1483,
      "mean_token_accuracy": 0.970802903175354,
      "num_tokens": 236236.0,
      "step": 531
    },
    {
      "epoch": 0.0409199292362126,
      "grad_norm": 14.5625,
      "learning_rate": 2.938620106145681e-06,
      "loss": 0.1552,
      "mean_token_accuracy": 0.9683794379234314,
      "num_tokens": 236490.0,
      "step": 532
    },
    {
      "epoch": 0.040996846396431044,
      "grad_norm": 28.375,
      "learning_rate": 2.9385047304053535e-06,
      "loss": 0.0716,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 237070.0,
      "step": 533
    },
    {
      "epoch": 0.04107376355664949,
      "grad_norm": 6.71875,
      "learning_rate": 2.938389354665026e-06,
      "loss": 0.0735,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 237610.0,
      "step": 534
    },
    {
      "epoch": 0.041150680716867935,
      "grad_norm": 24.25,
      "learning_rate": 2.938273978924698e-06,
      "loss": 0.152,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 237890.0,
      "step": 535
    },
    {
      "epoch": 0.04122759787708638,
      "grad_norm": 7.15625,
      "learning_rate": 2.9381586031843707e-06,
      "loss": 0.1122,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 238236.0,
      "step": 536
    },
    {
      "epoch": 0.041304515037304826,
      "grad_norm": 7.34375,
      "learning_rate": 2.938043227444043e-06,
      "loss": 0.0718,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 238826.0,
      "step": 537
    },
    {
      "epoch": 0.041381432197523264,
      "grad_norm": 13.3125,
      "learning_rate": 2.937927851703715e-06,
      "loss": 0.1238,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 239140.0,
      "step": 538
    },
    {
      "epoch": 0.04145834935774171,
      "grad_norm": 22.125,
      "learning_rate": 2.937812475963388e-06,
      "loss": 0.1298,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 239478.0,
      "step": 539
    },
    {
      "epoch": 0.041535266517960155,
      "grad_norm": 11.1875,
      "learning_rate": 2.9376971002230597e-06,
      "loss": 0.1399,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 239756.0,
      "step": 540
    },
    {
      "epoch": 0.0416121836781786,
      "grad_norm": 15.875,
      "learning_rate": 2.937581724482732e-06,
      "loss": 0.1482,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 240026.0,
      "step": 541
    },
    {
      "epoch": 0.041689100838397046,
      "grad_norm": 14.3125,
      "learning_rate": 2.9374663487424046e-06,
      "loss": 0.1143,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 240386.0,
      "step": 542
    },
    {
      "epoch": 0.04176601799861549,
      "grad_norm": 5.34375,
      "learning_rate": 2.937350973002077e-06,
      "loss": 0.0674,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 240950.0,
      "step": 543
    },
    {
      "epoch": 0.04184293515883394,
      "grad_norm": 6.0,
      "learning_rate": 2.937235597261749e-06,
      "loss": 0.0757,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 241474.0,
      "step": 544
    },
    {
      "epoch": 0.04191985231905238,
      "grad_norm": 11.8125,
      "learning_rate": 2.9371202215214214e-06,
      "loss": 0.1371,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 241758.0,
      "step": 545
    },
    {
      "epoch": 0.04199676947927083,
      "grad_norm": 9.3125,
      "learning_rate": 2.9370048457810937e-06,
      "loss": 0.1139,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 242102.0,
      "step": 546
    },
    {
      "epoch": 0.04207368663948927,
      "grad_norm": 11.8125,
      "learning_rate": 2.9368894700407663e-06,
      "loss": 0.1299,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 242400.0,
      "step": 547
    },
    {
      "epoch": 0.04215060379970771,
      "grad_norm": 17.25,
      "learning_rate": 2.9367740943004386e-06,
      "loss": 0.1035,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 242772.0,
      "step": 548
    },
    {
      "epoch": 0.04222752095992616,
      "grad_norm": 9.375,
      "learning_rate": 2.936658718560111e-06,
      "loss": 0.1292,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 243070.0,
      "step": 549
    },
    {
      "epoch": 0.0423044381201446,
      "grad_norm": 15.3125,
      "learning_rate": 2.936543342819783e-06,
      "loss": 0.1077,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 243410.0,
      "step": 550
    },
    {
      "epoch": 0.04238135528036305,
      "grad_norm": 16.0,
      "learning_rate": 2.9364279670794554e-06,
      "loss": 0.1377,
      "mean_token_accuracy": 0.9824561476707458,
      "num_tokens": 243696.0,
      "step": 551
    },
    {
      "epoch": 0.04245827244058149,
      "grad_norm": 12.1875,
      "learning_rate": 2.9363125913391276e-06,
      "loss": 0.1136,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 244046.0,
      "step": 552
    },
    {
      "epoch": 0.04253518960079994,
      "grad_norm": 20.375,
      "learning_rate": 2.9361972155988003e-06,
      "loss": 0.1269,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 244360.0,
      "step": 553
    },
    {
      "epoch": 0.042612106761018384,
      "grad_norm": 13.125,
      "learning_rate": 2.9360818398584726e-06,
      "loss": 0.1284,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 244656.0,
      "step": 554
    },
    {
      "epoch": 0.04268902392123683,
      "grad_norm": 14.4375,
      "learning_rate": 2.935966464118145e-06,
      "loss": 0.119,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 244988.0,
      "step": 555
    },
    {
      "epoch": 0.042765941081455275,
      "grad_norm": 10.0625,
      "learning_rate": 2.935851088377817e-06,
      "loss": 0.0755,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 245514.0,
      "step": 556
    },
    {
      "epoch": 0.04284285824167372,
      "grad_norm": 20.25,
      "learning_rate": 2.9357357126374893e-06,
      "loss": 0.1205,
      "mean_token_accuracy": 0.982300877571106,
      "num_tokens": 245854.0,
      "step": 557
    },
    {
      "epoch": 0.04291977540189216,
      "grad_norm": 10.3125,
      "learning_rate": 2.9356203368971616e-06,
      "loss": 0.0756,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 246394.0,
      "step": 558
    },
    {
      "epoch": 0.042996692562110604,
      "grad_norm": 13.75,
      "learning_rate": 2.9355049611568343e-06,
      "loss": 0.1179,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 246712.0,
      "step": 559
    },
    {
      "epoch": 0.04307360972232905,
      "grad_norm": 6.21875,
      "learning_rate": 2.9353895854165065e-06,
      "loss": 0.0682,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 247286.0,
      "step": 560
    },
    {
      "epoch": 0.043150526882547495,
      "grad_norm": 19.0,
      "learning_rate": 2.935274209676179e-06,
      "loss": 0.1297,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 247594.0,
      "step": 561
    },
    {
      "epoch": 0.04322744404276594,
      "grad_norm": 5.625,
      "learning_rate": 2.935158833935851e-06,
      "loss": 0.051,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 248404.0,
      "step": 562
    },
    {
      "epoch": 0.043304361202984386,
      "grad_norm": 6.59375,
      "learning_rate": 2.9350434581955233e-06,
      "loss": 0.0713,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 248966.0,
      "step": 563
    },
    {
      "epoch": 0.04338127836320283,
      "grad_norm": 6.625,
      "learning_rate": 2.934928082455196e-06,
      "loss": 0.07,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 249526.0,
      "step": 564
    },
    {
      "epoch": 0.04345819552342128,
      "grad_norm": 13.375,
      "learning_rate": 2.9348127067148683e-06,
      "loss": 0.1361,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 249808.0,
      "step": 565
    },
    {
      "epoch": 0.04353511268363972,
      "grad_norm": 13.5,
      "learning_rate": 2.9346973309745405e-06,
      "loss": 0.1143,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 250144.0,
      "step": 566
    },
    {
      "epoch": 0.04361202984385817,
      "grad_norm": 3.78125,
      "learning_rate": 2.9345819552342128e-06,
      "loss": 0.0636,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 250718.0,
      "step": 567
    },
    {
      "epoch": 0.04368894700407661,
      "grad_norm": 6.3125,
      "learning_rate": 2.934466579493885e-06,
      "loss": 0.0647,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 251288.0,
      "step": 568
    },
    {
      "epoch": 0.04376586416429505,
      "grad_norm": 25.75,
      "learning_rate": 2.9343512037535573e-06,
      "loss": 0.1425,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 251590.0,
      "step": 569
    },
    {
      "epoch": 0.0438427813245135,
      "grad_norm": 16.25,
      "learning_rate": 2.93423582801323e-06,
      "loss": 0.1276,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 251878.0,
      "step": 570
    },
    {
      "epoch": 0.04391969848473194,
      "grad_norm": 12.0,
      "learning_rate": 2.9341204522729022e-06,
      "loss": 0.1247,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 252200.0,
      "step": 571
    },
    {
      "epoch": 0.04399661564495039,
      "grad_norm": 5.4375,
      "learning_rate": 2.9340050765325745e-06,
      "loss": 0.0731,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 252722.0,
      "step": 572
    },
    {
      "epoch": 0.04407353280516883,
      "grad_norm": 21.5,
      "learning_rate": 2.9338897007922467e-06,
      "loss": 0.1257,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 253050.0,
      "step": 573
    },
    {
      "epoch": 0.04415044996538728,
      "grad_norm": 9.5,
      "learning_rate": 2.933774325051919e-06,
      "loss": 0.1321,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 253338.0,
      "step": 574
    },
    {
      "epoch": 0.044227367125605724,
      "grad_norm": 10.3125,
      "learning_rate": 2.9336589493115917e-06,
      "loss": 0.1257,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 253638.0,
      "step": 575
    },
    {
      "epoch": 0.04430428428582417,
      "grad_norm": 9.0,
      "learning_rate": 2.933543573571264e-06,
      "loss": 0.1139,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 253960.0,
      "step": 576
    },
    {
      "epoch": 0.044381201446042615,
      "grad_norm": 12.875,
      "learning_rate": 2.933428197830936e-06,
      "loss": 0.1121,
      "mean_token_accuracy": 0.9827089309692383,
      "num_tokens": 254308.0,
      "step": 577
    },
    {
      "epoch": 0.04445811860626106,
      "grad_norm": 6.6875,
      "learning_rate": 2.9333128220906085e-06,
      "loss": 0.0653,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 254892.0,
      "step": 578
    },
    {
      "epoch": 0.0445350357664795,
      "grad_norm": 5.84375,
      "learning_rate": 2.9331974463502807e-06,
      "loss": 0.0632,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 255482.0,
      "step": 579
    },
    {
      "epoch": 0.044611952926697944,
      "grad_norm": 11.25,
      "learning_rate": 2.933082070609953e-06,
      "loss": 0.1188,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 255790.0,
      "step": 580
    },
    {
      "epoch": 0.04468887008691639,
      "grad_norm": 4.59375,
      "learning_rate": 2.9329666948696257e-06,
      "loss": 0.066,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 256348.0,
      "step": 581
    },
    {
      "epoch": 0.044765787247134835,
      "grad_norm": 9.0,
      "learning_rate": 2.932851319129298e-06,
      "loss": 0.0704,
      "mean_token_accuracy": 0.9894551634788513,
      "num_tokens": 256918.0,
      "step": 582
    },
    {
      "epoch": 0.04484270440735328,
      "grad_norm": 12.5625,
      "learning_rate": 2.93273594338897e-06,
      "loss": 0.122,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 257228.0,
      "step": 583
    },
    {
      "epoch": 0.044919621567571726,
      "grad_norm": 6.75,
      "learning_rate": 2.9326205676486424e-06,
      "loss": 0.0701,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 257774.0,
      "step": 584
    },
    {
      "epoch": 0.04499653872779017,
      "grad_norm": 4.25,
      "learning_rate": 2.9325051919083147e-06,
      "loss": 0.0459,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 258582.0,
      "step": 585
    },
    {
      "epoch": 0.045073455888008616,
      "grad_norm": 8.0,
      "learning_rate": 2.9323898161679874e-06,
      "loss": 0.0672,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 259150.0,
      "step": 586
    },
    {
      "epoch": 0.04515037304822706,
      "grad_norm": 4.0625,
      "learning_rate": 2.9322744404276596e-06,
      "loss": 0.0467,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 259936.0,
      "step": 587
    },
    {
      "epoch": 0.04522729020844551,
      "grad_norm": 6.15625,
      "learning_rate": 2.932159064687332e-06,
      "loss": 0.0713,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 260452.0,
      "step": 588
    },
    {
      "epoch": 0.045304207368663946,
      "grad_norm": 6.1875,
      "learning_rate": 2.932043688947004e-06,
      "loss": 0.0635,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 261038.0,
      "step": 589
    },
    {
      "epoch": 0.04538112452888239,
      "grad_norm": 7.875,
      "learning_rate": 2.9319283132066764e-06,
      "loss": 0.0745,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 261534.0,
      "step": 590
    },
    {
      "epoch": 0.04545804168910084,
      "grad_norm": 7.28125,
      "learning_rate": 2.9318129374663487e-06,
      "loss": 0.0711,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 262052.0,
      "step": 591
    },
    {
      "epoch": 0.04553495884931928,
      "grad_norm": 11.1875,
      "learning_rate": 2.9316975617260213e-06,
      "loss": 0.1278,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 262350.0,
      "step": 592
    },
    {
      "epoch": 0.04561187600953773,
      "grad_norm": 12.875,
      "learning_rate": 2.9315821859856936e-06,
      "loss": 0.1195,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 262672.0,
      "step": 593
    },
    {
      "epoch": 0.04568879316975617,
      "grad_norm": 5.21875,
      "learning_rate": 2.931466810245366e-06,
      "loss": 0.0645,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 263234.0,
      "step": 594
    },
    {
      "epoch": 0.04576571032997462,
      "grad_norm": 15.3125,
      "learning_rate": 2.931351434505038e-06,
      "loss": 0.1245,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 263524.0,
      "step": 595
    },
    {
      "epoch": 0.045842627490193064,
      "grad_norm": 14.5625,
      "learning_rate": 2.9312360587647104e-06,
      "loss": 0.1289,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 263800.0,
      "step": 596
    },
    {
      "epoch": 0.04591954465041151,
      "grad_norm": 3.359375,
      "learning_rate": 2.931120683024383e-06,
      "loss": 0.0622,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 264346.0,
      "step": 597
    },
    {
      "epoch": 0.045996461810629954,
      "grad_norm": 6.0625,
      "learning_rate": 2.9310053072840553e-06,
      "loss": 0.067,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 264892.0,
      "step": 598
    },
    {
      "epoch": 0.04607337897084839,
      "grad_norm": 11.0,
      "learning_rate": 2.9308899315437276e-06,
      "loss": 0.0637,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 265504.0,
      "step": 599
    },
    {
      "epoch": 0.04615029613106684,
      "grad_norm": 17.25,
      "learning_rate": 2.9307745558034e-06,
      "loss": 0.1204,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 265816.0,
      "step": 600
    },
    {
      "epoch": 0.046227213291285284,
      "grad_norm": 4.875,
      "learning_rate": 2.930659180063072e-06,
      "loss": 0.0656,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 266356.0,
      "step": 601
    },
    {
      "epoch": 0.04630413045150373,
      "grad_norm": 7.1875,
      "learning_rate": 2.9305438043227443e-06,
      "loss": 0.0684,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 266898.0,
      "step": 602
    },
    {
      "epoch": 0.046381047611722175,
      "grad_norm": 13.75,
      "learning_rate": 2.930428428582417e-06,
      "loss": 0.1134,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 267222.0,
      "step": 603
    },
    {
      "epoch": 0.04645796477194062,
      "grad_norm": 12.6875,
      "learning_rate": 2.9303130528420893e-06,
      "loss": 0.1127,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 267524.0,
      "step": 604
    },
    {
      "epoch": 0.046534881932159065,
      "grad_norm": 6.15625,
      "learning_rate": 2.9301976771017615e-06,
      "loss": 0.0662,
      "mean_token_accuracy": 0.9961464405059814,
      "num_tokens": 268044.0,
      "step": 605
    },
    {
      "epoch": 0.04661179909237751,
      "grad_norm": 13.8125,
      "learning_rate": 2.930082301361434e-06,
      "loss": 0.1153,
      "mean_token_accuracy": 0.9840764403343201,
      "num_tokens": 268360.0,
      "step": 606
    },
    {
      "epoch": 0.046688716252595956,
      "grad_norm": 33.25,
      "learning_rate": 2.929966925621106e-06,
      "loss": 0.1228,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 268664.0,
      "step": 607
    },
    {
      "epoch": 0.0467656334128144,
      "grad_norm": 9.375,
      "learning_rate": 2.9298515498807783e-06,
      "loss": 0.0726,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 269156.0,
      "step": 608
    },
    {
      "epoch": 0.04684255057303285,
      "grad_norm": 5.4375,
      "learning_rate": 2.929736174140451e-06,
      "loss": 0.065,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 269708.0,
      "step": 609
    },
    {
      "epoch": 0.046919467733251286,
      "grad_norm": 12.9375,
      "learning_rate": 2.929620798400123e-06,
      "loss": 0.1166,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 270008.0,
      "step": 610
    },
    {
      "epoch": 0.04699638489346973,
      "grad_norm": 10.6875,
      "learning_rate": 2.9295054226597955e-06,
      "loss": 0.1099,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 270324.0,
      "step": 611
    },
    {
      "epoch": 0.047073302053688176,
      "grad_norm": 11.875,
      "learning_rate": 2.9293900469194678e-06,
      "loss": 0.1117,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 270650.0,
      "step": 612
    },
    {
      "epoch": 0.04715021921390662,
      "grad_norm": 6.21875,
      "learning_rate": 2.92927467117914e-06,
      "loss": 0.0663,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 271166.0,
      "step": 613
    },
    {
      "epoch": 0.04722713637412507,
      "grad_norm": 7.59375,
      "learning_rate": 2.9291592954388127e-06,
      "loss": 0.1048,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 271490.0,
      "step": 614
    },
    {
      "epoch": 0.04730405353434351,
      "grad_norm": 6.5,
      "learning_rate": 2.929043919698485e-06,
      "loss": 0.0619,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 272068.0,
      "step": 615
    },
    {
      "epoch": 0.04738097069456196,
      "grad_norm": 4.0625,
      "learning_rate": 2.928928543958157e-06,
      "loss": 0.0621,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 272614.0,
      "step": 616
    },
    {
      "epoch": 0.0474578878547804,
      "grad_norm": 5.90625,
      "learning_rate": 2.9288131682178295e-06,
      "loss": 0.0643,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 273160.0,
      "step": 617
    },
    {
      "epoch": 0.04753480501499885,
      "grad_norm": 204.0,
      "learning_rate": 2.9286977924775017e-06,
      "loss": 0.1137,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 273472.0,
      "step": 618
    },
    {
      "epoch": 0.047611722175217294,
      "grad_norm": 18.25,
      "learning_rate": 2.928582416737174e-06,
      "loss": 0.13,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 273750.0,
      "step": 619
    },
    {
      "epoch": 0.04768863933543573,
      "grad_norm": 19.625,
      "learning_rate": 2.9284670409968467e-06,
      "loss": 0.1507,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 273982.0,
      "step": 620
    },
    {
      "epoch": 0.04776555649565418,
      "grad_norm": 13.4375,
      "learning_rate": 2.9283516652565185e-06,
      "loss": 0.1183,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 274272.0,
      "step": 621
    },
    {
      "epoch": 0.047842473655872624,
      "grad_norm": 24.625,
      "learning_rate": 2.928236289516191e-06,
      "loss": 0.0979,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 274632.0,
      "step": 622
    },
    {
      "epoch": 0.04791939081609107,
      "grad_norm": 4.125,
      "learning_rate": 2.9281209137758635e-06,
      "loss": 0.0606,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 275178.0,
      "step": 623
    },
    {
      "epoch": 0.047996307976309514,
      "grad_norm": 12.25,
      "learning_rate": 2.9280055380355357e-06,
      "loss": 0.0972,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 275522.0,
      "step": 624
    },
    {
      "epoch": 0.04807322513652796,
      "grad_norm": 6.25,
      "learning_rate": 2.9278901622952084e-06,
      "loss": 0.0612,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 276106.0,
      "step": 625
    },
    {
      "epoch": 0.048150142296746405,
      "grad_norm": 5.34375,
      "learning_rate": 2.9277747865548807e-06,
      "loss": 0.061,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 276662.0,
      "step": 626
    },
    {
      "epoch": 0.04822705945696485,
      "grad_norm": 14.375,
      "learning_rate": 2.9276594108145525e-06,
      "loss": 0.1289,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 276934.0,
      "step": 627
    },
    {
      "epoch": 0.048303976617183296,
      "grad_norm": 4.40625,
      "learning_rate": 2.927544035074225e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 277978.0,
      "step": 628
    },
    {
      "epoch": 0.04838089377740174,
      "grad_norm": 4.625,
      "learning_rate": 2.9274286593338974e-06,
      "loss": 0.0634,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 278496.0,
      "step": 629
    },
    {
      "epoch": 0.04845781093762018,
      "grad_norm": 7.375,
      "learning_rate": 2.9273132835935697e-06,
      "loss": 0.0629,
      "mean_token_accuracy": 0.9894179701805115,
      "num_tokens": 279064.0,
      "step": 630
    },
    {
      "epoch": 0.048534728097838625,
      "grad_norm": 8.4375,
      "learning_rate": 2.9271979078532424e-06,
      "loss": 0.0547,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 279628.0,
      "step": 631
    },
    {
      "epoch": 0.04861164525805707,
      "grad_norm": 12.875,
      "learning_rate": 2.927082532112914e-06,
      "loss": 0.0968,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 279980.0,
      "step": 632
    },
    {
      "epoch": 0.048688562418275516,
      "grad_norm": 13.6875,
      "learning_rate": 2.926967156372587e-06,
      "loss": 0.1211,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 280266.0,
      "step": 633
    },
    {
      "epoch": 0.04876547957849396,
      "grad_norm": 11.3125,
      "learning_rate": 2.926851780632259e-06,
      "loss": 0.1189,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 280560.0,
      "step": 634
    },
    {
      "epoch": 0.04884239673871241,
      "grad_norm": 5.09375,
      "learning_rate": 2.9267364048919314e-06,
      "loss": 0.0605,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 281076.0,
      "step": 635
    },
    {
      "epoch": 0.04891931389893085,
      "grad_norm": 9.25,
      "learning_rate": 2.926621029151604e-06,
      "loss": 0.0976,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 281426.0,
      "step": 636
    },
    {
      "epoch": 0.0489962310591493,
      "grad_norm": 3.828125,
      "learning_rate": 2.9265056534112763e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 282222.0,
      "step": 637
    },
    {
      "epoch": 0.04907314821936774,
      "grad_norm": 4.59375,
      "learning_rate": 2.926390277670948e-06,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 283004.0,
      "step": 638
    },
    {
      "epoch": 0.04915006537958619,
      "grad_norm": 10.9375,
      "learning_rate": 2.926274901930621e-06,
      "loss": 0.112,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 283304.0,
      "step": 639
    },
    {
      "epoch": 0.04922698253980463,
      "grad_norm": 5.625,
      "learning_rate": 2.926159526190293e-06,
      "loss": 0.0567,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 283874.0,
      "step": 640
    },
    {
      "epoch": 0.04930389970002307,
      "grad_norm": 12.8125,
      "learning_rate": 2.9260441504499654e-06,
      "loss": 0.122,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 284158.0,
      "step": 641
    },
    {
      "epoch": 0.04938081686024152,
      "grad_norm": 5.21875,
      "learning_rate": 2.925928774709638e-06,
      "loss": 0.0601,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 284704.0,
      "step": 642
    },
    {
      "epoch": 0.04945773402045996,
      "grad_norm": 7.96875,
      "learning_rate": 2.92581339896931e-06,
      "loss": 0.1025,
      "mean_token_accuracy": 0.9842271208763123,
      "num_tokens": 285022.0,
      "step": 643
    },
    {
      "epoch": 0.04953465118067841,
      "grad_norm": 23.5,
      "learning_rate": 2.9256980232289826e-06,
      "loss": 0.1025,
      "mean_token_accuracy": 0.9789789915084839,
      "num_tokens": 285358.0,
      "step": 644
    },
    {
      "epoch": 0.049611568340896854,
      "grad_norm": 10.625,
      "learning_rate": 2.925582647488655e-06,
      "loss": 0.1022,
      "mean_token_accuracy": 0.9966777563095093,
      "num_tokens": 285660.0,
      "step": 645
    },
    {
      "epoch": 0.0496884855011153,
      "grad_norm": 6.03125,
      "learning_rate": 2.925467271748327e-06,
      "loss": 0.0585,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 286228.0,
      "step": 646
    },
    {
      "epoch": 0.049765402661333745,
      "grad_norm": 6.0,
      "learning_rate": 2.9253518960079998e-06,
      "loss": 0.0583,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 286784.0,
      "step": 647
    },
    {
      "epoch": 0.04984231982155219,
      "grad_norm": 3.546875,
      "learning_rate": 2.9252365202676716e-06,
      "loss": 0.0593,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 287308.0,
      "step": 648
    },
    {
      "epoch": 0.049919236981770636,
      "grad_norm": 14.6875,
      "learning_rate": 2.925121144527344e-06,
      "loss": 0.1175,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 287588.0,
      "step": 649
    },
    {
      "epoch": 0.04999615414198908,
      "grad_norm": 4.0,
      "learning_rate": 2.9250057687870165e-06,
      "loss": 0.0548,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 288152.0,
      "step": 650
    },
    {
      "epoch": 0.05007307130220752,
      "grad_norm": 11.5,
      "learning_rate": 2.924890393046689e-06,
      "loss": 0.1079,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 288448.0,
      "step": 651
    },
    {
      "epoch": 0.050149988462425965,
      "grad_norm": 8.75,
      "learning_rate": 2.924775017306361e-06,
      "loss": 0.0989,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 288770.0,
      "step": 652
    },
    {
      "epoch": 0.05022690562264441,
      "grad_norm": 9.0,
      "learning_rate": 2.9246596415660337e-06,
      "loss": 0.1112,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 289048.0,
      "step": 653
    },
    {
      "epoch": 0.050303822782862856,
      "grad_norm": 4.6875,
      "learning_rate": 2.9245442658257056e-06,
      "loss": 0.0603,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 289570.0,
      "step": 654
    },
    {
      "epoch": 0.0503807399430813,
      "grad_norm": 8.0625,
      "learning_rate": 2.9244288900853783e-06,
      "loss": 0.0976,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 289890.0,
      "step": 655
    },
    {
      "epoch": 0.05045765710329975,
      "grad_norm": 10.5,
      "learning_rate": 2.9243135143450505e-06,
      "loss": 0.0565,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 290472.0,
      "step": 656
    },
    {
      "epoch": 0.05053457426351819,
      "grad_norm": 4.59375,
      "learning_rate": 2.9241981386047228e-06,
      "loss": 0.0533,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 291058.0,
      "step": 657
    },
    {
      "epoch": 0.05061149142373664,
      "grad_norm": 10.625,
      "learning_rate": 2.924082762864395e-06,
      "loss": 0.0949,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 291386.0,
      "step": 658
    },
    {
      "epoch": 0.05068840858395508,
      "grad_norm": 9.125,
      "learning_rate": 2.9239673871240673e-06,
      "loss": 0.1001,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 291696.0,
      "step": 659
    },
    {
      "epoch": 0.05076532574417353,
      "grad_norm": 7.71875,
      "learning_rate": 2.9238520113837396e-06,
      "loss": 0.089,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 292042.0,
      "step": 660
    },
    {
      "epoch": 0.05084224290439197,
      "grad_norm": 3.90625,
      "learning_rate": 2.9237366356434122e-06,
      "loss": 0.0547,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 292586.0,
      "step": 661
    },
    {
      "epoch": 0.05091916006461041,
      "grad_norm": 5.03125,
      "learning_rate": 2.9236212599030845e-06,
      "loss": 0.0555,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 293144.0,
      "step": 662
    },
    {
      "epoch": 0.05099607722482886,
      "grad_norm": 9.0,
      "learning_rate": 2.9235058841627567e-06,
      "loss": 0.0862,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 293494.0,
      "step": 663
    },
    {
      "epoch": 0.0510729943850473,
      "grad_norm": 8.4375,
      "learning_rate": 2.9233905084224294e-06,
      "loss": 0.096,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 293804.0,
      "step": 664
    },
    {
      "epoch": 0.05114991154526575,
      "grad_norm": 5.96875,
      "learning_rate": 2.9232751326821013e-06,
      "loss": 0.0574,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 294352.0,
      "step": 665
    },
    {
      "epoch": 0.051226828705484194,
      "grad_norm": 3.5,
      "learning_rate": 2.9231597569417735e-06,
      "loss": 0.0511,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 294926.0,
      "step": 666
    },
    {
      "epoch": 0.05130374586570264,
      "grad_norm": 7.03125,
      "learning_rate": 2.923044381201446e-06,
      "loss": 0.0846,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 295270.0,
      "step": 667
    },
    {
      "epoch": 0.051380663025921085,
      "grad_norm": 12.0,
      "learning_rate": 2.9229290054611185e-06,
      "loss": 0.1089,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 295552.0,
      "step": 668
    },
    {
      "epoch": 0.05145758018613953,
      "grad_norm": 13.125,
      "learning_rate": 2.9228136297207907e-06,
      "loss": 0.1064,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 295842.0,
      "step": 669
    },
    {
      "epoch": 0.051534497346357976,
      "grad_norm": 7.09375,
      "learning_rate": 2.922698253980463e-06,
      "loss": 0.0941,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 296154.0,
      "step": 670
    },
    {
      "epoch": 0.051611414506576414,
      "grad_norm": 5.5625,
      "learning_rate": 2.9225828782401352e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 296952.0,
      "step": 671
    },
    {
      "epoch": 0.05168833166679486,
      "grad_norm": 7.84375,
      "learning_rate": 2.922467502499808e-06,
      "loss": 0.0825,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 297302.0,
      "step": 672
    },
    {
      "epoch": 0.051765248827013305,
      "grad_norm": 7.65625,
      "learning_rate": 2.92235212675948e-06,
      "loss": 0.0823,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 297652.0,
      "step": 673
    },
    {
      "epoch": 0.05184216598723175,
      "grad_norm": 6.75,
      "learning_rate": 2.9222367510191524e-06,
      "loss": 0.0543,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 298210.0,
      "step": 674
    },
    {
      "epoch": 0.051919083147450196,
      "grad_norm": 5.75,
      "learning_rate": 2.9221213752788247e-06,
      "loss": 0.0523,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 298756.0,
      "step": 675
    },
    {
      "epoch": 0.05199600030766864,
      "grad_norm": 11.75,
      "learning_rate": 2.922005999538497e-06,
      "loss": 0.0879,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 299084.0,
      "step": 676
    },
    {
      "epoch": 0.052072917467887087,
      "grad_norm": 4.78125,
      "learning_rate": 2.921890623798169e-06,
      "loss": 0.0478,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 299658.0,
      "step": 677
    },
    {
      "epoch": 0.05214983462810553,
      "grad_norm": 9.5,
      "learning_rate": 2.921775248057842e-06,
      "loss": 0.0508,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 300222.0,
      "step": 678
    },
    {
      "epoch": 0.05222675178832398,
      "grad_norm": 6.46875,
      "learning_rate": 2.921659872317514e-06,
      "loss": 0.0485,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 300792.0,
      "step": 679
    },
    {
      "epoch": 0.05230366894854242,
      "grad_norm": 4.5,
      "learning_rate": 2.9215444965771864e-06,
      "loss": 0.0509,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 301352.0,
      "step": 680
    },
    {
      "epoch": 0.05238058610876086,
      "grad_norm": 9.0625,
      "learning_rate": 2.9214291208368587e-06,
      "loss": 0.0897,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 301672.0,
      "step": 681
    },
    {
      "epoch": 0.05245750326897931,
      "grad_norm": 6.1875,
      "learning_rate": 2.921313745096531e-06,
      "loss": 0.0556,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 302204.0,
      "step": 682
    },
    {
      "epoch": 0.05253442042919775,
      "grad_norm": 14.4375,
      "learning_rate": 2.9211983693562036e-06,
      "loss": 0.0909,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 302518.0,
      "step": 683
    },
    {
      "epoch": 0.0526113375894162,
      "grad_norm": 14.0,
      "learning_rate": 2.921082993615876e-06,
      "loss": 0.0867,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 302830.0,
      "step": 684
    },
    {
      "epoch": 0.05268825474963464,
      "grad_norm": 11.5,
      "learning_rate": 2.920967617875548e-06,
      "loss": 0.0802,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 303162.0,
      "step": 685
    },
    {
      "epoch": 0.05276517190985309,
      "grad_norm": 12.375,
      "learning_rate": 2.9208522421352204e-06,
      "loss": 0.0904,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 303466.0,
      "step": 686
    },
    {
      "epoch": 0.052842089070071534,
      "grad_norm": 7.1875,
      "learning_rate": 2.9207368663948926e-06,
      "loss": 0.0532,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 303990.0,
      "step": 687
    },
    {
      "epoch": 0.05291900623028998,
      "grad_norm": 8.25,
      "learning_rate": 2.920621490654565e-06,
      "loss": 0.0823,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 304312.0,
      "step": 688
    },
    {
      "epoch": 0.052995923390508425,
      "grad_norm": 6.8125,
      "learning_rate": 2.9205061149142376e-06,
      "loss": 0.0504,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 304836.0,
      "step": 689
    },
    {
      "epoch": 0.05307284055072687,
      "grad_norm": 6.71875,
      "learning_rate": 2.92039073917391e-06,
      "loss": 0.0832,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 305130.0,
      "step": 690
    },
    {
      "epoch": 0.05314975771094531,
      "grad_norm": 9.875,
      "learning_rate": 2.920275363433582e-06,
      "loss": 0.0649,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 305486.0,
      "step": 691
    },
    {
      "epoch": 0.053226674871163754,
      "grad_norm": 9.0,
      "learning_rate": 2.9201599876932544e-06,
      "loss": 0.0494,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 306058.0,
      "step": 692
    },
    {
      "epoch": 0.0533035920313822,
      "grad_norm": 4.96875,
      "learning_rate": 2.9200446119529266e-06,
      "loss": 0.0437,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 306604.0,
      "step": 693
    },
    {
      "epoch": 0.053380509191600645,
      "grad_norm": 11.5625,
      "learning_rate": 2.9199292362125993e-06,
      "loss": 0.0796,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 306916.0,
      "step": 694
    },
    {
      "epoch": 0.05345742635181909,
      "grad_norm": 11.3125,
      "learning_rate": 2.9198138604722715e-06,
      "loss": 0.0499,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 307480.0,
      "step": 695
    },
    {
      "epoch": 0.053534343512037535,
      "grad_norm": 12.8125,
      "learning_rate": 2.919698484731944e-06,
      "loss": 0.0703,
      "mean_token_accuracy": 0.9830028414726257,
      "num_tokens": 307834.0,
      "step": 696
    },
    {
      "epoch": 0.05361126067225598,
      "grad_norm": 10.625,
      "learning_rate": 2.919583108991616e-06,
      "loss": 0.0831,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 308112.0,
      "step": 697
    },
    {
      "epoch": 0.053688177832474426,
      "grad_norm": 10.5,
      "learning_rate": 2.9194677332512883e-06,
      "loss": 0.0725,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 308430.0,
      "step": 698
    },
    {
      "epoch": 0.05376509499269287,
      "grad_norm": 5.5,
      "learning_rate": 2.9193523575109606e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 309250.0,
      "step": 699
    },
    {
      "epoch": 0.05384201215291132,
      "grad_norm": 10.5625,
      "learning_rate": 2.9192369817706333e-06,
      "loss": 0.0712,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 309574.0,
      "step": 700
    },
    {
      "epoch": 0.05391892931312976,
      "grad_norm": 9.125,
      "learning_rate": 2.9191216060303055e-06,
      "loss": 0.0645,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 309898.0,
      "step": 701
    },
    {
      "epoch": 0.0539958464733482,
      "grad_norm": 8.1875,
      "learning_rate": 2.9190062302899778e-06,
      "loss": 0.0683,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 310182.0,
      "step": 702
    },
    {
      "epoch": 0.054072763633566646,
      "grad_norm": 16.375,
      "learning_rate": 2.91889085454965e-06,
      "loss": 0.0571,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 310530.0,
      "step": 703
    },
    {
      "epoch": 0.05414968079378509,
      "grad_norm": 10.5,
      "learning_rate": 2.9187754788093223e-06,
      "loss": 0.066,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 310836.0,
      "step": 704
    },
    {
      "epoch": 0.05422659795400354,
      "grad_norm": 6.46875,
      "learning_rate": 2.918660103068995e-06,
      "loss": 0.038,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 311376.0,
      "step": 705
    },
    {
      "epoch": 0.05430351511422198,
      "grad_norm": 5.46875,
      "learning_rate": 2.9185447273286672e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 311960.0,
      "step": 706
    },
    {
      "epoch": 0.05438043227444043,
      "grad_norm": 6.21875,
      "learning_rate": 2.9184293515883395e-06,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 312504.0,
      "step": 707
    },
    {
      "epoch": 0.054457349434658874,
      "grad_norm": 9.8125,
      "learning_rate": 2.9183139758480118e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 313086.0,
      "step": 708
    },
    {
      "epoch": 0.05453426659487732,
      "grad_norm": 9.4375,
      "learning_rate": 2.918198600107684e-06,
      "loss": 0.0528,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 313416.0,
      "step": 709
    },
    {
      "epoch": 0.054611183755095764,
      "grad_norm": 10.375,
      "learning_rate": 2.9180832243673563e-06,
      "loss": 0.0611,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 313708.0,
      "step": 710
    },
    {
      "epoch": 0.05468810091531421,
      "grad_norm": 10.8125,
      "learning_rate": 2.917967848627029e-06,
      "loss": 0.0494,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 314060.0,
      "step": 711
    },
    {
      "epoch": 0.05476501807553265,
      "grad_norm": 5.84375,
      "learning_rate": 2.917852472886701e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 314582.0,
      "step": 712
    },
    {
      "epoch": 0.054841935235751094,
      "grad_norm": 4.34375,
      "learning_rate": 2.9177370971463735e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 315378.0,
      "step": 713
    },
    {
      "epoch": 0.05491885239596954,
      "grad_norm": 7.0,
      "learning_rate": 2.9176217214060457e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 315914.0,
      "step": 714
    },
    {
      "epoch": 0.054995769556187984,
      "grad_norm": 9.6875,
      "learning_rate": 2.917506345665718e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 316254.0,
      "step": 715
    },
    {
      "epoch": 0.05507268671640643,
      "grad_norm": 4.46875,
      "learning_rate": 2.9173909699253902e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 316818.0,
      "step": 716
    },
    {
      "epoch": 0.055149603876624875,
      "grad_norm": 9.625,
      "learning_rate": 2.917275594185063e-06,
      "loss": 0.0518,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 317110.0,
      "step": 717
    },
    {
      "epoch": 0.05522652103684332,
      "grad_norm": 14.125,
      "learning_rate": 2.917160218444735e-06,
      "loss": 0.0458,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 317474.0,
      "step": 718
    },
    {
      "epoch": 0.055303438197061766,
      "grad_norm": 9.25,
      "learning_rate": 2.9170448427044074e-06,
      "loss": 0.0465,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 317766.0,
      "step": 719
    },
    {
      "epoch": 0.05538035535728021,
      "grad_norm": 11.875,
      "learning_rate": 2.9169294669640797e-06,
      "loss": 0.0544,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 318044.0,
      "step": 720
    },
    {
      "epoch": 0.05545727251749866,
      "grad_norm": 4.90625,
      "learning_rate": 2.916814091223752e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 318612.0,
      "step": 721
    },
    {
      "epoch": 0.055534189677717095,
      "grad_norm": 14.375,
      "learning_rate": 2.9166987154834246e-06,
      "loss": 0.052,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 318946.0,
      "step": 722
    },
    {
      "epoch": 0.05561110683793554,
      "grad_norm": 3.765625,
      "learning_rate": 2.916583339743097e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 319990.0,
      "step": 723
    },
    {
      "epoch": 0.055688023998153986,
      "grad_norm": 9.6875,
      "learning_rate": 2.9164679640027687e-06,
      "loss": 0.0458,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 320296.0,
      "step": 724
    },
    {
      "epoch": 0.05576494115837243,
      "grad_norm": 4.5625,
      "learning_rate": 2.9163525882624414e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 320838.0,
      "step": 725
    },
    {
      "epoch": 0.05584185831859088,
      "grad_norm": 7.59375,
      "learning_rate": 2.9162372125221137e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 321430.0,
      "step": 726
    },
    {
      "epoch": 0.05591877547880932,
      "grad_norm": 5.53125,
      "learning_rate": 2.916121836781786e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 322246.0,
      "step": 727
    },
    {
      "epoch": 0.05599569263902777,
      "grad_norm": 15.375,
      "learning_rate": 2.9160064610414586e-06,
      "loss": 0.0487,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 322532.0,
      "step": 728
    },
    {
      "epoch": 0.05607260979924621,
      "grad_norm": 10.125,
      "learning_rate": 2.915891085301131e-06,
      "loss": 0.0433,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 322856.0,
      "step": 729
    },
    {
      "epoch": 0.05614952695946466,
      "grad_norm": 7.5,
      "learning_rate": 2.915775709560803e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 323160.0,
      "step": 730
    },
    {
      "epoch": 0.056226444119683104,
      "grad_norm": 5.5625,
      "learning_rate": 2.9156603338204754e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 323708.0,
      "step": 731
    },
    {
      "epoch": 0.05630336127990154,
      "grad_norm": 9.6875,
      "learning_rate": 2.9155449580801476e-06,
      "loss": 0.0428,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 323996.0,
      "step": 732
    },
    {
      "epoch": 0.05638027844011999,
      "grad_norm": 11.0625,
      "learning_rate": 2.9154295823398203e-06,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 324306.0,
      "step": 733
    },
    {
      "epoch": 0.05645719560033843,
      "grad_norm": 11.3125,
      "learning_rate": 2.9153142065994926e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 324612.0,
      "step": 734
    },
    {
      "epoch": 0.05653411276055688,
      "grad_norm": 5.0,
      "learning_rate": 2.9151988308591644e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 325180.0,
      "step": 735
    },
    {
      "epoch": 0.056611029920775324,
      "grad_norm": 19.375,
      "learning_rate": 2.915083455118837e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 325480.0,
      "step": 736
    },
    {
      "epoch": 0.05668794708099377,
      "grad_norm": 8.75,
      "learning_rate": 2.9149680793785094e-06,
      "loss": 0.0431,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 325780.0,
      "step": 737
    },
    {
      "epoch": 0.056764864241212215,
      "grad_norm": 3.90625,
      "learning_rate": 2.9148527036381816e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 326326.0,
      "step": 738
    },
    {
      "epoch": 0.05684178140143066,
      "grad_norm": 6.125,
      "learning_rate": 2.9147373278978543e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 326868.0,
      "step": 739
    },
    {
      "epoch": 0.056918698561649106,
      "grad_norm": 4.96875,
      "learning_rate": 2.9146219521575266e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 327676.0,
      "step": 740
    },
    {
      "epoch": 0.05699561572186755,
      "grad_norm": 6.75,
      "learning_rate": 2.914506576417199e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.994575023651123,
      "num_tokens": 328230.0,
      "step": 741
    },
    {
      "epoch": 0.057072532882086,
      "grad_norm": 6.0,
      "learning_rate": 2.914391200676871e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 328514.0,
      "step": 742
    },
    {
      "epoch": 0.057149450042304435,
      "grad_norm": 17.375,
      "learning_rate": 2.9142758249365433e-06,
      "loss": 0.0433,
      "mean_token_accuracy": 0.989276111125946,
      "num_tokens": 328890.0,
      "step": 743
    },
    {
      "epoch": 0.05722636720252288,
      "grad_norm": 12.25,
      "learning_rate": 2.914160449196216e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 329250.0,
      "step": 744
    },
    {
      "epoch": 0.057303284362741326,
      "grad_norm": 6.90625,
      "learning_rate": 2.9140450734558883e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 329824.0,
      "step": 745
    },
    {
      "epoch": 0.05738020152295977,
      "grad_norm": 4.53125,
      "learning_rate": 2.91392969771556e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 330408.0,
      "step": 746
    },
    {
      "epoch": 0.05745711868317822,
      "grad_norm": 11.25,
      "learning_rate": 2.9138143219752328e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 330770.0,
      "step": 747
    },
    {
      "epoch": 0.05753403584339666,
      "grad_norm": 6.34375,
      "learning_rate": 2.913698946234905e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 331372.0,
      "step": 748
    },
    {
      "epoch": 0.05761095300361511,
      "grad_norm": 4.53125,
      "learning_rate": 2.9135835704945773e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 331906.0,
      "step": 749
    },
    {
      "epoch": 0.05768787016383355,
      "grad_norm": 11.0625,
      "learning_rate": 2.91346819475425e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 332226.0,
      "step": 750
    },
    {
      "epoch": 0.057764787324052,
      "grad_norm": 9.0,
      "learning_rate": 2.913352819013922e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 332778.0,
      "step": 751
    },
    {
      "epoch": 0.057841704484270444,
      "grad_norm": 12.375,
      "learning_rate": 2.9132374432735945e-06,
      "loss": 0.0412,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 333094.0,
      "step": 752
    },
    {
      "epoch": 0.05791862164448888,
      "grad_norm": 10.5625,
      "learning_rate": 2.9131220675332668e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 333620.0,
      "step": 753
    },
    {
      "epoch": 0.05799553880470733,
      "grad_norm": 2.171875,
      "learning_rate": 2.913006691792939e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 334220.0,
      "step": 754
    },
    {
      "epoch": 0.05807245596492577,
      "grad_norm": 14.6875,
      "learning_rate": 2.9128913160526117e-06,
      "loss": 0.0445,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 334526.0,
      "step": 755
    },
    {
      "epoch": 0.05814937312514422,
      "grad_norm": 22.125,
      "learning_rate": 2.912775940312284e-06,
      "loss": 0.0474,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 334790.0,
      "step": 756
    },
    {
      "epoch": 0.058226290285362664,
      "grad_norm": 5.40625,
      "learning_rate": 2.9126605645719558e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 335364.0,
      "step": 757
    },
    {
      "epoch": 0.05830320744558111,
      "grad_norm": 8.0625,
      "learning_rate": 2.9125451888316285e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 335688.0,
      "step": 758
    },
    {
      "epoch": 0.058380124605799555,
      "grad_norm": 15.125,
      "learning_rate": 2.9124298130913007e-06,
      "loss": 0.0455,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 335964.0,
      "step": 759
    },
    {
      "epoch": 0.058457041766018,
      "grad_norm": 12.5625,
      "learning_rate": 2.912314437350973e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 336248.0,
      "step": 760
    },
    {
      "epoch": 0.058533958926236446,
      "grad_norm": 4.34375,
      "learning_rate": 2.9121990616106457e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 337064.0,
      "step": 761
    },
    {
      "epoch": 0.05861087608645489,
      "grad_norm": 15.875,
      "learning_rate": 2.9120836858703175e-06,
      "loss": 0.0454,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 337372.0,
      "step": 762
    },
    {
      "epoch": 0.05868779324667333,
      "grad_norm": 5.46875,
      "learning_rate": 2.91196831012999e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 338208.0,
      "step": 763
    },
    {
      "epoch": 0.058764710406891775,
      "grad_norm": 8.5625,
      "learning_rate": 2.9118529343896624e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 338502.0,
      "step": 764
    },
    {
      "epoch": 0.05884162756711022,
      "grad_norm": 11.3125,
      "learning_rate": 2.9117375586493347e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 338834.0,
      "step": 765
    },
    {
      "epoch": 0.058918544727328666,
      "grad_norm": 4.15625,
      "learning_rate": 2.9116221829090074e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 339642.0,
      "step": 766
    },
    {
      "epoch": 0.05899546188754711,
      "grad_norm": 8.0,
      "learning_rate": 2.9115068071686796e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 339992.0,
      "step": 767
    },
    {
      "epoch": 0.05907237904776556,
      "grad_norm": 5.5625,
      "learning_rate": 2.9113914314283515e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 340550.0,
      "step": 768
    },
    {
      "epoch": 0.059149296207984,
      "grad_norm": 3.859375,
      "learning_rate": 2.911276055688024e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 341346.0,
      "step": 769
    },
    {
      "epoch": 0.05922621336820245,
      "grad_norm": 11.75,
      "learning_rate": 2.9111606799476964e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 341626.0,
      "step": 770
    },
    {
      "epoch": 0.05930313052842089,
      "grad_norm": 10.6875,
      "learning_rate": 2.9110453042073687e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 341958.0,
      "step": 771
    },
    {
      "epoch": 0.05938004768863934,
      "grad_norm": 13.625,
      "learning_rate": 2.9109299284670414e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 342220.0,
      "step": 772
    },
    {
      "epoch": 0.05945696484885778,
      "grad_norm": 10.3125,
      "learning_rate": 2.910814552726713e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 342500.0,
      "step": 773
    },
    {
      "epoch": 0.05953388200907622,
      "grad_norm": 11.9375,
      "learning_rate": 2.910699176986386e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 342804.0,
      "step": 774
    },
    {
      "epoch": 0.05961079916929467,
      "grad_norm": 4.3125,
      "learning_rate": 2.910583801246058e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9973474740982056,
      "num_tokens": 343560.0,
      "step": 775
    },
    {
      "epoch": 0.05968771632951311,
      "grad_norm": 6.0,
      "learning_rate": 2.9104684255057304e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 343882.0,
      "step": 776
    },
    {
      "epoch": 0.05976463348973156,
      "grad_norm": 5.625,
      "learning_rate": 2.9103530497654026e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 344440.0,
      "step": 777
    },
    {
      "epoch": 0.059841550649950004,
      "grad_norm": 11.75,
      "learning_rate": 2.910237674025075e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 344786.0,
      "step": 778
    },
    {
      "epoch": 0.05991846781016845,
      "grad_norm": 7.375,
      "learning_rate": 2.910122298284747e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 345096.0,
      "step": 779
    },
    {
      "epoch": 0.059995384970386895,
      "grad_norm": 10.5625,
      "learning_rate": 2.91000692254442e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 345406.0,
      "step": 780
    },
    {
      "epoch": 0.06007230213060534,
      "grad_norm": 8.875,
      "learning_rate": 2.909891546804092e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 345706.0,
      "step": 781
    },
    {
      "epoch": 0.060149219290823785,
      "grad_norm": 10.25,
      "learning_rate": 2.9097761710637644e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 346006.0,
      "step": 782
    },
    {
      "epoch": 0.06022613645104223,
      "grad_norm": 4.75,
      "learning_rate": 2.909660795323437e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 346552.0,
      "step": 783
    },
    {
      "epoch": 0.06030305361126067,
      "grad_norm": 11.625,
      "learning_rate": 2.909545419583109e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 346830.0,
      "step": 784
    },
    {
      "epoch": 0.060379970771479115,
      "grad_norm": 4.46875,
      "learning_rate": 2.909430043842781e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 347416.0,
      "step": 785
    },
    {
      "epoch": 0.06045688793169756,
      "grad_norm": 6.09375,
      "learning_rate": 2.909314668102454e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 348018.0,
      "step": 786
    },
    {
      "epoch": 0.060533805091916006,
      "grad_norm": 11.25,
      "learning_rate": 2.909199292362126e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 348304.0,
      "step": 787
    },
    {
      "epoch": 0.06061072225213445,
      "grad_norm": 4.84375,
      "learning_rate": 2.9090839166217983e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 348848.0,
      "step": 788
    },
    {
      "epoch": 0.060687639412352896,
      "grad_norm": 7.5,
      "learning_rate": 2.9089685408814706e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 349132.0,
      "step": 789
    },
    {
      "epoch": 0.06076455657257134,
      "grad_norm": 6.5,
      "learning_rate": 2.908853165141143e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 349420.0,
      "step": 790
    },
    {
      "epoch": 0.06084147373278979,
      "grad_norm": 7.0,
      "learning_rate": 2.9087377894008155e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 349732.0,
      "step": 791
    },
    {
      "epoch": 0.06091839089300823,
      "grad_norm": 3.5625,
      "learning_rate": 2.9086224136604878e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9971346855163574,
      "num_tokens": 350782.0,
      "step": 792
    },
    {
      "epoch": 0.06099530805322668,
      "grad_norm": 4.53125,
      "learning_rate": 2.90850703792016e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 351562.0,
      "step": 793
    },
    {
      "epoch": 0.06107222521344512,
      "grad_norm": 11.625,
      "learning_rate": 2.9083916621798327e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 351858.0,
      "step": 794
    },
    {
      "epoch": 0.06114914237366356,
      "grad_norm": 5.46875,
      "learning_rate": 2.9082762864395046e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 352414.0,
      "step": 795
    },
    {
      "epoch": 0.06122605953388201,
      "grad_norm": 3.78125,
      "learning_rate": 2.908160910699177e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 353194.0,
      "step": 796
    },
    {
      "epoch": 0.06130297669410045,
      "grad_norm": 5.5,
      "learning_rate": 2.9080455349588495e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 353714.0,
      "step": 797
    },
    {
      "epoch": 0.0613798938543189,
      "grad_norm": 4.21875,
      "learning_rate": 2.9079301592185218e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 354282.0,
      "step": 798
    },
    {
      "epoch": 0.061456811014537344,
      "grad_norm": 4.28125,
      "learning_rate": 2.907814783478194e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 354830.0,
      "step": 799
    },
    {
      "epoch": 0.06153372817475579,
      "grad_norm": 10.625,
      "learning_rate": 2.9076994077378663e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 355162.0,
      "step": 800
    },
    {
      "epoch": 0.061610645334974234,
      "grad_norm": 2.703125,
      "learning_rate": 2.9075840319975385e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 355708.0,
      "step": 801
    },
    {
      "epoch": 0.06168756249519268,
      "grad_norm": 5.78125,
      "learning_rate": 2.9074686562572112e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 356284.0,
      "step": 802
    },
    {
      "epoch": 0.061764479655411125,
      "grad_norm": 12.0,
      "learning_rate": 2.9073532805168835e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 356538.0,
      "step": 803
    },
    {
      "epoch": 0.061841396815629564,
      "grad_norm": 12.5625,
      "learning_rate": 2.9072379047765557e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 356800.0,
      "step": 804
    },
    {
      "epoch": 0.06191831397584801,
      "grad_norm": 2.53125,
      "learning_rate": 2.9071225290362284e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 357342.0,
      "step": 805
    },
    {
      "epoch": 0.061995231136066455,
      "grad_norm": 11.875,
      "learning_rate": 2.9070071532959002e-06,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 357614.0,
      "step": 806
    },
    {
      "epoch": 0.0620721482962849,
      "grad_norm": 12.8125,
      "learning_rate": 2.9068917775555725e-06,
      "loss": 0.0405,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 357942.0,
      "step": 807
    },
    {
      "epoch": 0.062149065456503345,
      "grad_norm": 9.4375,
      "learning_rate": 2.906776401815245e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 358282.0,
      "step": 808
    },
    {
      "epoch": 0.06222598261672179,
      "grad_norm": 3.890625,
      "learning_rate": 2.9066610260749174e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 358838.0,
      "step": 809
    },
    {
      "epoch": 0.062302899776940236,
      "grad_norm": 12.0625,
      "learning_rate": 2.9065456503345897e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 359178.0,
      "step": 810
    },
    {
      "epoch": 0.06237981693715868,
      "grad_norm": 6.25,
      "learning_rate": 2.906430274594262e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9888682961463928,
      "num_tokens": 359718.0,
      "step": 811
    },
    {
      "epoch": 0.06245673409737713,
      "grad_norm": 10.0,
      "learning_rate": 2.9063148988539342e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 359976.0,
      "step": 812
    },
    {
      "epoch": 0.06253365125759557,
      "grad_norm": 11.0,
      "learning_rate": 2.906199523113607e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 360288.0,
      "step": 813
    },
    {
      "epoch": 0.06261056841781401,
      "grad_norm": 9.3125,
      "learning_rate": 2.906084147373279e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 360614.0,
      "step": 814
    },
    {
      "epoch": 0.06268748557803246,
      "grad_norm": 9.0,
      "learning_rate": 2.9059687716329514e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 360954.0,
      "step": 815
    },
    {
      "epoch": 0.0627644027382509,
      "grad_norm": 6.25,
      "learning_rate": 2.9058533958926237e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 361300.0,
      "step": 816
    },
    {
      "epoch": 0.06284131989846935,
      "grad_norm": 4.0625,
      "learning_rate": 2.905738020152296e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 362086.0,
      "step": 817
    },
    {
      "epoch": 0.06291823705868779,
      "grad_norm": 10.0,
      "learning_rate": 2.905622644411968e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 362388.0,
      "step": 818
    },
    {
      "epoch": 0.06299515421890624,
      "grad_norm": 8.8125,
      "learning_rate": 2.905507268671641e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 362692.0,
      "step": 819
    },
    {
      "epoch": 0.06307207137912468,
      "grad_norm": 11.3125,
      "learning_rate": 2.905391892931313e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 362952.0,
      "step": 820
    },
    {
      "epoch": 0.06314898853934313,
      "grad_norm": 9.5625,
      "learning_rate": 2.9052765171909854e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 363222.0,
      "step": 821
    },
    {
      "epoch": 0.06322590569956157,
      "grad_norm": 14.6875,
      "learning_rate": 2.9051611414506576e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9880478382110596,
      "num_tokens": 363474.0,
      "step": 822
    },
    {
      "epoch": 0.06330282285978002,
      "grad_norm": 8.0625,
      "learning_rate": 2.90504576571033e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 363756.0,
      "step": 823
    },
    {
      "epoch": 0.06337974001999847,
      "grad_norm": 6.5625,
      "learning_rate": 2.9049303899700026e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 364066.0,
      "step": 824
    },
    {
      "epoch": 0.06345665718021691,
      "grad_norm": 4.625,
      "learning_rate": 2.904815014229675e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 364588.0,
      "step": 825
    },
    {
      "epoch": 0.06353357434043536,
      "grad_norm": 5.3125,
      "learning_rate": 2.904699638489347e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 365134.0,
      "step": 826
    },
    {
      "epoch": 0.0636104915006538,
      "grad_norm": 11.3125,
      "learning_rate": 2.9045842627490194e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 365474.0,
      "step": 827
    },
    {
      "epoch": 0.06368740866087225,
      "grad_norm": 9.5,
      "learning_rate": 2.9044688870086916e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 365792.0,
      "step": 828
    },
    {
      "epoch": 0.06376432582109069,
      "grad_norm": 11.75,
      "learning_rate": 2.904353511268364e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 366148.0,
      "step": 829
    },
    {
      "epoch": 0.06384124298130912,
      "grad_norm": 6.09375,
      "learning_rate": 2.9042381355280366e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 366712.0,
      "step": 830
    },
    {
      "epoch": 0.06391816014152757,
      "grad_norm": 13.125,
      "learning_rate": 2.904122759787709e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 366944.0,
      "step": 831
    },
    {
      "epoch": 0.06399507730174601,
      "grad_norm": 10.0,
      "learning_rate": 2.904007384047381e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 367230.0,
      "step": 832
    },
    {
      "epoch": 0.06407199446196446,
      "grad_norm": 5.5,
      "learning_rate": 2.9038920083070533e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 367542.0,
      "step": 833
    },
    {
      "epoch": 0.0641489116221829,
      "grad_norm": 8.375,
      "learning_rate": 2.9037766325667256e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9894366264343262,
      "num_tokens": 367828.0,
      "step": 834
    },
    {
      "epoch": 0.06422582878240135,
      "grad_norm": 2.9375,
      "learning_rate": 2.903661256826398e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 368352.0,
      "step": 835
    },
    {
      "epoch": 0.0643027459426198,
      "grad_norm": 5.28125,
      "learning_rate": 2.9035458810860705e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.993914783000946,
      "num_tokens": 368846.0,
      "step": 836
    },
    {
      "epoch": 0.06437966310283824,
      "grad_norm": 4.96875,
      "learning_rate": 2.9034305053457428e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 369396.0,
      "step": 837
    },
    {
      "epoch": 0.06445658026305669,
      "grad_norm": 8.25,
      "learning_rate": 2.903315129605415e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 369960.0,
      "step": 838
    },
    {
      "epoch": 0.06453349742327513,
      "grad_norm": 8.5,
      "learning_rate": 2.9031997538650873e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 370286.0,
      "step": 839
    },
    {
      "epoch": 0.06461041458349358,
      "grad_norm": 9.1875,
      "learning_rate": 2.9030843781247596e-06,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 370580.0,
      "step": 840
    },
    {
      "epoch": 0.06468733174371202,
      "grad_norm": 4.59375,
      "learning_rate": 2.9029690023844322e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 371388.0,
      "step": 841
    },
    {
      "epoch": 0.06476424890393047,
      "grad_norm": 9.75,
      "learning_rate": 2.9028536266441045e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 371726.0,
      "step": 842
    },
    {
      "epoch": 0.06484116606414891,
      "grad_norm": 6.8125,
      "learning_rate": 2.9027382509037768e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 372294.0,
      "step": 843
    },
    {
      "epoch": 0.06491808322436736,
      "grad_norm": 8.6875,
      "learning_rate": 2.902622875163449e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 372600.0,
      "step": 844
    },
    {
      "epoch": 0.0649950003845858,
      "grad_norm": 14.9375,
      "learning_rate": 2.9025074994231213e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 372920.0,
      "step": 845
    },
    {
      "epoch": 0.06507191754480425,
      "grad_norm": 5.625,
      "learning_rate": 2.9023921236827935e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 373440.0,
      "step": 846
    },
    {
      "epoch": 0.0651488347050227,
      "grad_norm": 3.5625,
      "learning_rate": 2.9022767479424662e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 374028.0,
      "step": 847
    },
    {
      "epoch": 0.06522575186524114,
      "grad_norm": 3.21875,
      "learning_rate": 2.9021613722021385e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 374576.0,
      "step": 848
    },
    {
      "epoch": 0.06530266902545959,
      "grad_norm": 7.5,
      "learning_rate": 2.9020459964618107e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 375116.0,
      "step": 849
    },
    {
      "epoch": 0.06537958618567802,
      "grad_norm": 7.5625,
      "learning_rate": 2.901930620721483e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 375438.0,
      "step": 850
    },
    {
      "epoch": 0.06545650334589646,
      "grad_norm": 6.96875,
      "learning_rate": 2.9018152449811552e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 375974.0,
      "step": 851
    },
    {
      "epoch": 0.06553342050611491,
      "grad_norm": 12.625,
      "learning_rate": 2.901699869240828e-06,
      "loss": 0.0394,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 376276.0,
      "step": 852
    },
    {
      "epoch": 0.06561033766633335,
      "grad_norm": 13.4375,
      "learning_rate": 2.9015844935005e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 376622.0,
      "step": 853
    },
    {
      "epoch": 0.0656872548265518,
      "grad_norm": 4.65625,
      "learning_rate": 2.901469117760172e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 377170.0,
      "step": 854
    },
    {
      "epoch": 0.06576417198677025,
      "grad_norm": 4.09375,
      "learning_rate": 2.9013537420198447e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 377742.0,
      "step": 855
    },
    {
      "epoch": 0.06584108914698869,
      "grad_norm": 3.046875,
      "learning_rate": 2.901238366279517e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 378328.0,
      "step": 856
    },
    {
      "epoch": 0.06591800630720714,
      "grad_norm": 3.28125,
      "learning_rate": 2.9011229905391892e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 379086.0,
      "step": 857
    },
    {
      "epoch": 0.06599492346742558,
      "grad_norm": 11.125,
      "learning_rate": 2.901007614798862e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 379356.0,
      "step": 858
    },
    {
      "epoch": 0.06607184062764403,
      "grad_norm": 10.375,
      "learning_rate": 2.900892239058534e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 379634.0,
      "step": 859
    },
    {
      "epoch": 0.06614875778786247,
      "grad_norm": 9.5,
      "learning_rate": 2.9007768633182064e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 379940.0,
      "step": 860
    },
    {
      "epoch": 0.06622567494808092,
      "grad_norm": 9.0,
      "learning_rate": 2.9006614875778787e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9827089309692383,
      "num_tokens": 380288.0,
      "step": 861
    },
    {
      "epoch": 0.06630259210829936,
      "grad_norm": 3.21875,
      "learning_rate": 2.900546111837551e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 381076.0,
      "step": 862
    },
    {
      "epoch": 0.06637950926851781,
      "grad_norm": 4.125,
      "learning_rate": 2.9004307360972236e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 381620.0,
      "step": 863
    },
    {
      "epoch": 0.06645642642873625,
      "grad_norm": 6.1875,
      "learning_rate": 2.900315360356896e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 382160.0,
      "step": 864
    },
    {
      "epoch": 0.0665333435889547,
      "grad_norm": 8.25,
      "learning_rate": 2.9001999846165677e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 382478.0,
      "step": 865
    },
    {
      "epoch": 0.06661026074917314,
      "grad_norm": 3.890625,
      "learning_rate": 2.9000846088762404e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 383042.0,
      "step": 866
    },
    {
      "epoch": 0.06668717790939159,
      "grad_norm": 7.09375,
      "learning_rate": 2.8999692331359126e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 383618.0,
      "step": 867
    },
    {
      "epoch": 0.06676409506961004,
      "grad_norm": 10.25,
      "learning_rate": 2.899853857395585e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 383974.0,
      "step": 868
    },
    {
      "epoch": 0.06684101222982848,
      "grad_norm": 3.265625,
      "learning_rate": 2.8997384816552576e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 384520.0,
      "step": 869
    },
    {
      "epoch": 0.06691792939004693,
      "grad_norm": 3.5625,
      "learning_rate": 2.89962310591493e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 385106.0,
      "step": 870
    },
    {
      "epoch": 0.06699484655026536,
      "grad_norm": 12.5,
      "learning_rate": 2.899507730174602e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 385420.0,
      "step": 871
    },
    {
      "epoch": 0.0670717637104838,
      "grad_norm": 3.53125,
      "learning_rate": 2.8993923544342744e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 385978.0,
      "step": 872
    },
    {
      "epoch": 0.06714868087070225,
      "grad_norm": 3.078125,
      "learning_rate": 2.8992769786939466e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 386798.0,
      "step": 873
    },
    {
      "epoch": 0.0672255980309207,
      "grad_norm": 7.5,
      "learning_rate": 2.8991616029536193e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 387410.0,
      "step": 874
    },
    {
      "epoch": 0.06730251519113914,
      "grad_norm": 9.25,
      "learning_rate": 2.8990462272132916e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 387754.0,
      "step": 875
    },
    {
      "epoch": 0.06737943235135758,
      "grad_norm": 7.3125,
      "learning_rate": 2.8989308514729634e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 388088.0,
      "step": 876
    },
    {
      "epoch": 0.06745634951157603,
      "grad_norm": 7.34375,
      "learning_rate": 2.898815475732636e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 388452.0,
      "step": 877
    },
    {
      "epoch": 0.06753326667179448,
      "grad_norm": 14.75,
      "learning_rate": 2.8987000999923083e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 388758.0,
      "step": 878
    },
    {
      "epoch": 0.06761018383201292,
      "grad_norm": 6.90625,
      "learning_rate": 2.8985847242519806e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 389304.0,
      "step": 879
    },
    {
      "epoch": 0.06768710099223137,
      "grad_norm": 12.6875,
      "learning_rate": 2.8984693485116533e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 389600.0,
      "step": 880
    },
    {
      "epoch": 0.06776401815244981,
      "grad_norm": 16.875,
      "learning_rate": 2.8983539727713255e-06,
      "loss": 0.048,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 389888.0,
      "step": 881
    },
    {
      "epoch": 0.06784093531266826,
      "grad_norm": 17.375,
      "learning_rate": 2.898238597030998e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 390168.0,
      "step": 882
    },
    {
      "epoch": 0.0679178524728867,
      "grad_norm": 7.28125,
      "learning_rate": 2.89812322129067e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 390718.0,
      "step": 883
    },
    {
      "epoch": 0.06799476963310515,
      "grad_norm": 9.8125,
      "learning_rate": 2.8980078455503423e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 391042.0,
      "step": 884
    },
    {
      "epoch": 0.0680716867933236,
      "grad_norm": 13.875,
      "learning_rate": 2.8978924698100146e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 391348.0,
      "step": 885
    },
    {
      "epoch": 0.06814860395354204,
      "grad_norm": 8.1875,
      "learning_rate": 2.8977770940696872e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 391676.0,
      "step": 886
    },
    {
      "epoch": 0.06822552111376048,
      "grad_norm": 11.125,
      "learning_rate": 2.897661718329359e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 391988.0,
      "step": 887
    },
    {
      "epoch": 0.06830243827397893,
      "grad_norm": 9.5625,
      "learning_rate": 2.8975463425890318e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 392288.0,
      "step": 888
    },
    {
      "epoch": 0.06837935543419738,
      "grad_norm": 4.3125,
      "learning_rate": 2.897430966848704e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 392832.0,
      "step": 889
    },
    {
      "epoch": 0.06845627259441582,
      "grad_norm": 11.9375,
      "learning_rate": 2.8973155911083763e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 393088.0,
      "step": 890
    },
    {
      "epoch": 0.06853318975463425,
      "grad_norm": 4.09375,
      "learning_rate": 2.897200215368049e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 393876.0,
      "step": 891
    },
    {
      "epoch": 0.0686101069148527,
      "grad_norm": 5.9375,
      "learning_rate": 2.897084839627721e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 394440.0,
      "step": 892
    },
    {
      "epoch": 0.06868702407507114,
      "grad_norm": 7.0,
      "learning_rate": 2.896969463887393e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 395042.0,
      "step": 893
    },
    {
      "epoch": 0.06876394123528959,
      "grad_norm": 7.90625,
      "learning_rate": 2.8968540881470657e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 395386.0,
      "step": 894
    },
    {
      "epoch": 0.06884085839550803,
      "grad_norm": 9.6875,
      "learning_rate": 2.896738712406738e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 395716.0,
      "step": 895
    },
    {
      "epoch": 0.06891777555572648,
      "grad_norm": 12.5,
      "learning_rate": 2.8966233366664102e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 396062.0,
      "step": 896
    },
    {
      "epoch": 0.06899469271594492,
      "grad_norm": 11.125,
      "learning_rate": 2.896507960926083e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 396402.0,
      "step": 897
    },
    {
      "epoch": 0.06907160987616337,
      "grad_norm": 3.53125,
      "learning_rate": 2.8963925851857548e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 397182.0,
      "step": 898
    },
    {
      "epoch": 0.06914852703638182,
      "grad_norm": 10.5,
      "learning_rate": 2.8962772094454274e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 397458.0,
      "step": 899
    },
    {
      "epoch": 0.06922544419660026,
      "grad_norm": 4.09375,
      "learning_rate": 2.8961618337050997e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 398042.0,
      "step": 900
    },
    {
      "epoch": 0.0693023613568187,
      "grad_norm": 3.203125,
      "learning_rate": 2.896046457964772e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 398850.0,
      "step": 901
    },
    {
      "epoch": 0.06937927851703715,
      "grad_norm": 3.828125,
      "learning_rate": 2.8959310822244446e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 399414.0,
      "step": 902
    },
    {
      "epoch": 0.0694561956772556,
      "grad_norm": 11.0625,
      "learning_rate": 2.8958157064841165e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 399748.0,
      "step": 903
    },
    {
      "epoch": 0.06953311283747404,
      "grad_norm": 11.75,
      "learning_rate": 2.8957003307437887e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 400092.0,
      "step": 904
    },
    {
      "epoch": 0.06961002999769249,
      "grad_norm": 7.46875,
      "learning_rate": 2.8955849550034614e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 400414.0,
      "step": 905
    },
    {
      "epoch": 0.06968694715791093,
      "grad_norm": 13.3125,
      "learning_rate": 2.8954695792631337e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 400778.0,
      "step": 906
    },
    {
      "epoch": 0.06976386431812938,
      "grad_norm": 7.5625,
      "learning_rate": 2.895354203522806e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 401102.0,
      "step": 907
    },
    {
      "epoch": 0.06984078147834782,
      "grad_norm": 4.09375,
      "learning_rate": 2.8952388277824786e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 401688.0,
      "step": 908
    },
    {
      "epoch": 0.06991769863856627,
      "grad_norm": 14.9375,
      "learning_rate": 2.8951234520421505e-06,
      "loss": 0.0508,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 401950.0,
      "step": 909
    },
    {
      "epoch": 0.06999461579878472,
      "grad_norm": 3.96875,
      "learning_rate": 2.895008076301823e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 402758.0,
      "step": 910
    },
    {
      "epoch": 0.07007153295900316,
      "grad_norm": 8.3125,
      "learning_rate": 2.8948927005614954e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 403080.0,
      "step": 911
    },
    {
      "epoch": 0.07014845011922159,
      "grad_norm": 3.125,
      "learning_rate": 2.8947773248211677e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 403660.0,
      "step": 912
    },
    {
      "epoch": 0.07022536727944004,
      "grad_norm": 4.15625,
      "learning_rate": 2.8946619490808403e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 404218.0,
      "step": 913
    },
    {
      "epoch": 0.07030228443965848,
      "grad_norm": 7.5,
      "learning_rate": 2.894546573340512e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 404540.0,
      "step": 914
    },
    {
      "epoch": 0.07037920159987693,
      "grad_norm": 3.46875,
      "learning_rate": 2.8944311976001844e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 405350.0,
      "step": 915
    },
    {
      "epoch": 0.07045611876009537,
      "grad_norm": 2.5625,
      "learning_rate": 2.894315821859857e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 405934.0,
      "step": 916
    },
    {
      "epoch": 0.07053303592031382,
      "grad_norm": 3.65625,
      "learning_rate": 2.8942004461195294e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 406502.0,
      "step": 917
    },
    {
      "epoch": 0.07060995308053226,
      "grad_norm": 7.15625,
      "learning_rate": 2.8940850703792016e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 406830.0,
      "step": 918
    },
    {
      "epoch": 0.07068687024075071,
      "grad_norm": 7.78125,
      "learning_rate": 2.893969694638874e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.994413435459137,
      "num_tokens": 407190.0,
      "step": 919
    },
    {
      "epoch": 0.07076378740096916,
      "grad_norm": 8.375,
      "learning_rate": 2.893854318898546e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 407510.0,
      "step": 920
    },
    {
      "epoch": 0.0708407045611876,
      "grad_norm": 5.125,
      "learning_rate": 2.893738943158219e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 408034.0,
      "step": 921
    },
    {
      "epoch": 0.07091762172140605,
      "grad_norm": 4.8125,
      "learning_rate": 2.893623567417891e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 408580.0,
      "step": 922
    },
    {
      "epoch": 0.07099453888162449,
      "grad_norm": 6.9375,
      "learning_rate": 2.8935081916775633e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 409148.0,
      "step": 923
    },
    {
      "epoch": 0.07107145604184294,
      "grad_norm": 4.53125,
      "learning_rate": 2.893392815937236e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 409722.0,
      "step": 924
    },
    {
      "epoch": 0.07114837320206138,
      "grad_norm": 7.9375,
      "learning_rate": 2.893277440196908e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 410046.0,
      "step": 925
    },
    {
      "epoch": 0.07122529036227983,
      "grad_norm": 9.125,
      "learning_rate": 2.89316206445658e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 410374.0,
      "step": 926
    },
    {
      "epoch": 0.07130220752249827,
      "grad_norm": 9.625,
      "learning_rate": 2.893046688716253e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 410692.0,
      "step": 927
    },
    {
      "epoch": 0.07137912468271672,
      "grad_norm": 9.5625,
      "learning_rate": 2.892931312975925e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 411008.0,
      "step": 928
    },
    {
      "epoch": 0.07145604184293516,
      "grad_norm": 6.5625,
      "learning_rate": 2.8928159372355973e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 411550.0,
      "step": 929
    },
    {
      "epoch": 0.07153295900315361,
      "grad_norm": 7.21875,
      "learning_rate": 2.8927005614952696e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 411896.0,
      "step": 930
    },
    {
      "epoch": 0.07160987616337205,
      "grad_norm": 15.375,
      "learning_rate": 2.892585185754942e-06,
      "loss": 0.0454,
      "mean_token_accuracy": 0.9880478382110596,
      "num_tokens": 412148.0,
      "step": 931
    },
    {
      "epoch": 0.07168679332359049,
      "grad_norm": 4.25,
      "learning_rate": 2.8924698100146145e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 412696.0,
      "step": 932
    },
    {
      "epoch": 0.07176371048380893,
      "grad_norm": 5.96875,
      "learning_rate": 2.8923544342742868e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 413252.0,
      "step": 933
    },
    {
      "epoch": 0.07184062764402738,
      "grad_norm": 4.28125,
      "learning_rate": 2.892239058533959e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 413564.0,
      "step": 934
    },
    {
      "epoch": 0.07191754480424582,
      "grad_norm": 8.5625,
      "learning_rate": 2.8921236827936313e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 413848.0,
      "step": 935
    },
    {
      "epoch": 0.07199446196446427,
      "grad_norm": 9.375,
      "learning_rate": 2.8920083070533035e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 414170.0,
      "step": 936
    },
    {
      "epoch": 0.07207137912468271,
      "grad_norm": 7.46875,
      "learning_rate": 2.891892931312976e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 414522.0,
      "step": 937
    },
    {
      "epoch": 0.07214829628490116,
      "grad_norm": 8.125,
      "learning_rate": 2.8917775555726485e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 414844.0,
      "step": 938
    },
    {
      "epoch": 0.0722252134451196,
      "grad_norm": 3.78125,
      "learning_rate": 2.8916621798323207e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 415390.0,
      "step": 939
    },
    {
      "epoch": 0.07230213060533805,
      "grad_norm": 5.59375,
      "learning_rate": 2.891546804091993e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 415962.0,
      "step": 940
    },
    {
      "epoch": 0.0723790477655565,
      "grad_norm": 5.65625,
      "learning_rate": 2.8914314283516653e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9931740760803223,
      "num_tokens": 416256.0,
      "step": 941
    },
    {
      "epoch": 0.07245596492577494,
      "grad_norm": 7.4375,
      "learning_rate": 2.8913160526113375e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9894179701805115,
      "num_tokens": 416824.0,
      "step": 942
    },
    {
      "epoch": 0.07253288208599339,
      "grad_norm": 9.25,
      "learning_rate": 2.8912006768710098e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 417138.0,
      "step": 943
    },
    {
      "epoch": 0.07260979924621183,
      "grad_norm": 10.375,
      "learning_rate": 2.8910853011306825e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 417488.0,
      "step": 944
    },
    {
      "epoch": 0.07268671640643028,
      "grad_norm": 7.0,
      "learning_rate": 2.8909699253903547e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 417796.0,
      "step": 945
    },
    {
      "epoch": 0.07276363356664872,
      "grad_norm": 5.90625,
      "learning_rate": 2.890854549650027e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 418346.0,
      "step": 946
    },
    {
      "epoch": 0.07284055072686717,
      "grad_norm": 6.8125,
      "learning_rate": 2.8907391739096992e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 418656.0,
      "step": 947
    },
    {
      "epoch": 0.07291746788708561,
      "grad_norm": 4.21875,
      "learning_rate": 2.8906237981693715e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 419216.0,
      "step": 948
    },
    {
      "epoch": 0.07299438504730406,
      "grad_norm": 3.640625,
      "learning_rate": 2.890508422429044e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 420040.0,
      "step": 949
    },
    {
      "epoch": 0.0730713022075225,
      "grad_norm": 3.390625,
      "learning_rate": 2.8903930466887164e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 420598.0,
      "step": 950
    },
    {
      "epoch": 0.07314821936774095,
      "grad_norm": 4.1875,
      "learning_rate": 2.8902776709483887e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 421148.0,
      "step": 951
    },
    {
      "epoch": 0.0732251365279594,
      "grad_norm": 7.40625,
      "learning_rate": 2.890162295208061e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 421690.0,
      "step": 952
    },
    {
      "epoch": 0.07330205368817783,
      "grad_norm": 7.34375,
      "learning_rate": 2.890046919467733e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 422230.0,
      "step": 953
    },
    {
      "epoch": 0.07337897084839627,
      "grad_norm": 9.0625,
      "learning_rate": 2.8899315437274055e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 422528.0,
      "step": 954
    },
    {
      "epoch": 0.07345588800861472,
      "grad_norm": 6.1875,
      "learning_rate": 2.889816167987078e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 423108.0,
      "step": 955
    },
    {
      "epoch": 0.07353280516883316,
      "grad_norm": 6.375,
      "learning_rate": 2.8897007922467504e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 423430.0,
      "step": 956
    },
    {
      "epoch": 0.07360972232905161,
      "grad_norm": 5.28125,
      "learning_rate": 2.8895854165064227e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 423998.0,
      "step": 957
    },
    {
      "epoch": 0.07368663948927005,
      "grad_norm": 6.15625,
      "learning_rate": 2.889470040766095e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 424320.0,
      "step": 958
    },
    {
      "epoch": 0.0737635566494885,
      "grad_norm": 8.6875,
      "learning_rate": 2.889354665025767e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 424630.0,
      "step": 959
    },
    {
      "epoch": 0.07384047380970694,
      "grad_norm": 11.6875,
      "learning_rate": 2.88923928928544e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 424942.0,
      "step": 960
    },
    {
      "epoch": 0.07391739096992539,
      "grad_norm": 4.53125,
      "learning_rate": 2.889123913545112e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 425506.0,
      "step": 961
    },
    {
      "epoch": 0.07399430813014383,
      "grad_norm": 3.375,
      "learning_rate": 2.8890085378047844e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 426332.0,
      "step": 962
    },
    {
      "epoch": 0.07407122529036228,
      "grad_norm": 8.0625,
      "learning_rate": 2.8888931620644566e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 426642.0,
      "step": 963
    },
    {
      "epoch": 0.07414814245058073,
      "grad_norm": 5.375,
      "learning_rate": 2.888777786324129e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 426992.0,
      "step": 964
    },
    {
      "epoch": 0.07422505961079917,
      "grad_norm": 7.53125,
      "learning_rate": 2.888662410583801e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 427292.0,
      "step": 965
    },
    {
      "epoch": 0.07430197677101762,
      "grad_norm": 11.75,
      "learning_rate": 2.888547034843474e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 427546.0,
      "step": 966
    },
    {
      "epoch": 0.07437889393123606,
      "grad_norm": 6.1875,
      "learning_rate": 2.888431659103146e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 428126.0,
      "step": 967
    },
    {
      "epoch": 0.07445581109145451,
      "grad_norm": 6.75,
      "learning_rate": 2.8883162833628183e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 428438.0,
      "step": 968
    },
    {
      "epoch": 0.07453272825167295,
      "grad_norm": 4.46875,
      "learning_rate": 2.8882009076224906e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 429194.0,
      "step": 969
    },
    {
      "epoch": 0.0746096454118914,
      "grad_norm": 6.53125,
      "learning_rate": 2.888085531882163e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 429546.0,
      "step": 970
    },
    {
      "epoch": 0.07468656257210984,
      "grad_norm": 5.4375,
      "learning_rate": 2.8879701561418355e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 430082.0,
      "step": 971
    },
    {
      "epoch": 0.07476347973232829,
      "grad_norm": 3.421875,
      "learning_rate": 2.887854780401508e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 430878.0,
      "step": 972
    },
    {
      "epoch": 0.07484039689254672,
      "grad_norm": 4.21875,
      "learning_rate": 2.88773940466118e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 431452.0,
      "step": 973
    },
    {
      "epoch": 0.07491731405276517,
      "grad_norm": 9.75,
      "learning_rate": 2.8876240289208523e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 431794.0,
      "step": 974
    },
    {
      "epoch": 0.07499423121298361,
      "grad_norm": 7.1875,
      "learning_rate": 2.8875086531805246e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9930555820465088,
      "num_tokens": 432084.0,
      "step": 975
    },
    {
      "epoch": 0.07507114837320206,
      "grad_norm": 9.75,
      "learning_rate": 2.887393277440197e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 432422.0,
      "step": 976
    },
    {
      "epoch": 0.0751480655334205,
      "grad_norm": 3.671875,
      "learning_rate": 2.8872779016998695e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 432972.0,
      "step": 977
    },
    {
      "epoch": 0.07522498269363895,
      "grad_norm": 12.3125,
      "learning_rate": 2.8871625259595418e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 433288.0,
      "step": 978
    },
    {
      "epoch": 0.07530189985385739,
      "grad_norm": 13.4375,
      "learning_rate": 2.887047150219214e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 433636.0,
      "step": 979
    },
    {
      "epoch": 0.07537881701407584,
      "grad_norm": 12.125,
      "learning_rate": 2.8869317744788863e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9768977165222168,
      "num_tokens": 433940.0,
      "step": 980
    },
    {
      "epoch": 0.07545573417429428,
      "grad_norm": 12.375,
      "learning_rate": 2.8868163987385585e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 434196.0,
      "step": 981
    },
    {
      "epoch": 0.07553265133451273,
      "grad_norm": 9.75,
      "learning_rate": 2.8867010229982312e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 434476.0,
      "step": 982
    },
    {
      "epoch": 0.07560956849473117,
      "grad_norm": 7.71875,
      "learning_rate": 2.8865856472579035e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 434828.0,
      "step": 983
    },
    {
      "epoch": 0.07568648565494962,
      "grad_norm": 3.765625,
      "learning_rate": 2.8864702715175757e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 435372.0,
      "step": 984
    },
    {
      "epoch": 0.07576340281516807,
      "grad_norm": 7.4375,
      "learning_rate": 2.886354895777248e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 435686.0,
      "step": 985
    },
    {
      "epoch": 0.07584031997538651,
      "grad_norm": 3.859375,
      "learning_rate": 2.8862395200369203e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 436260.0,
      "step": 986
    },
    {
      "epoch": 0.07591723713560496,
      "grad_norm": 9.6875,
      "learning_rate": 2.8861241442965925e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 436544.0,
      "step": 987
    },
    {
      "epoch": 0.0759941542958234,
      "grad_norm": 5.375,
      "learning_rate": 2.886008768556265e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 436832.0,
      "step": 988
    },
    {
      "epoch": 0.07607107145604185,
      "grad_norm": 7.71875,
      "learning_rate": 2.8858933928159375e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 437186.0,
      "step": 989
    },
    {
      "epoch": 0.07614798861626029,
      "grad_norm": 9.1875,
      "learning_rate": 2.8857780170756097e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 437466.0,
      "step": 990
    },
    {
      "epoch": 0.07622490577647874,
      "grad_norm": 11.75,
      "learning_rate": 2.885662641335282e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 437796.0,
      "step": 991
    },
    {
      "epoch": 0.07630182293669718,
      "grad_norm": 7.1875,
      "learning_rate": 2.8855472655949542e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 438104.0,
      "step": 992
    },
    {
      "epoch": 0.07637874009691561,
      "grad_norm": 6.96875,
      "learning_rate": 2.8854318898546265e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 438390.0,
      "step": 993
    },
    {
      "epoch": 0.07645565725713406,
      "grad_norm": 5.5,
      "learning_rate": 2.885316514114299e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 438710.0,
      "step": 994
    },
    {
      "epoch": 0.0765325744173525,
      "grad_norm": 5.71875,
      "learning_rate": 2.885201138373971e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 439014.0,
      "step": 995
    },
    {
      "epoch": 0.07660949157757095,
      "grad_norm": 5.4375,
      "learning_rate": 2.8850857626336437e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 439570.0,
      "step": 996
    },
    {
      "epoch": 0.0766864087377894,
      "grad_norm": 5.84375,
      "learning_rate": 2.884970386893316e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 440134.0,
      "step": 997
    },
    {
      "epoch": 0.07676332589800784,
      "grad_norm": 10.1875,
      "learning_rate": 2.884855011152988e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 440416.0,
      "step": 998
    },
    {
      "epoch": 0.07684024305822629,
      "grad_norm": 4.25,
      "learning_rate": 2.884739635412661e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 440938.0,
      "step": 999
    },
    {
      "epoch": 0.07691716021844473,
      "grad_norm": 5.3125,
      "learning_rate": 2.884624259672333e-06,
      "loss": 0.017,
      "step": 1000
    },
    {
      "epoch": 0.07691716021844473,
      "eval_loss": 0.02490871027112007,
      "eval_mean_token_accuracy": 0.9912069249153137,
      "eval_num_tokens": 441774.0,
      "eval_runtime": 21.3502,
      "eval_samples_per_second": 9.368,
      "eval_steps_per_second": 1.171,
      "step": 1000
    },
    {
      "epoch": 0.07699407737866318,
      "grad_norm": 4.25,
      "learning_rate": 2.884508883932005e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9958983361721039,
      "num_tokens": 442362.0,
      "step": 1001
    },
    {
      "epoch": 0.07707099453888162,
      "grad_norm": 10.6875,
      "learning_rate": 2.8843935081916777e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 442662.0,
      "step": 1002
    },
    {
      "epoch": 0.07714791169910007,
      "grad_norm": 5.3125,
      "learning_rate": 2.88427813245135e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 442944.0,
      "step": 1003
    },
    {
      "epoch": 0.07722482885931851,
      "grad_norm": 9.75,
      "learning_rate": 2.884162756711022e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 443250.0,
      "step": 1004
    },
    {
      "epoch": 0.07730174601953696,
      "grad_norm": 8.25,
      "learning_rate": 2.884047380970695e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 443610.0,
      "step": 1005
    },
    {
      "epoch": 0.0773786631797554,
      "grad_norm": 3.234375,
      "learning_rate": 2.8839320052303667e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 444418.0,
      "step": 1006
    },
    {
      "epoch": 0.07745558033997385,
      "grad_norm": 3.71875,
      "learning_rate": 2.8838166294900394e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 444964.0,
      "step": 1007
    },
    {
      "epoch": 0.0775324975001923,
      "grad_norm": 4.90625,
      "learning_rate": 2.8837012537497116e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 445488.0,
      "step": 1008
    },
    {
      "epoch": 0.07760941466041074,
      "grad_norm": 3.859375,
      "learning_rate": 2.883585878009384e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 446030.0,
      "step": 1009
    },
    {
      "epoch": 0.07768633182062919,
      "grad_norm": 6.9375,
      "learning_rate": 2.8834705022690566e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 446324.0,
      "step": 1010
    },
    {
      "epoch": 0.07776324898084763,
      "grad_norm": 7.25,
      "learning_rate": 2.883355126528729e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 446646.0,
      "step": 1011
    },
    {
      "epoch": 0.07784016614106608,
      "grad_norm": 2.671875,
      "learning_rate": 2.8832397507884007e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 447180.0,
      "step": 1012
    },
    {
      "epoch": 0.07791708330128452,
      "grad_norm": 8.75,
      "learning_rate": 2.8831243750480733e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.982300877571106,
      "num_tokens": 447520.0,
      "step": 1013
    },
    {
      "epoch": 0.07799400046150295,
      "grad_norm": 10.5625,
      "learning_rate": 2.8830089993077456e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 447834.0,
      "step": 1014
    },
    {
      "epoch": 0.0780709176217214,
      "grad_norm": 5.21875,
      "learning_rate": 2.882893623567418e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 448402.0,
      "step": 1015
    },
    {
      "epoch": 0.07814783478193985,
      "grad_norm": 3.359375,
      "learning_rate": 2.8827782478270905e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 448920.0,
      "step": 1016
    },
    {
      "epoch": 0.07822475194215829,
      "grad_norm": 10.4375,
      "learning_rate": 2.8826628720867624e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 449202.0,
      "step": 1017
    },
    {
      "epoch": 0.07830166910237674,
      "grad_norm": 15.9375,
      "learning_rate": 2.882547496346435e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 449454.0,
      "step": 1018
    },
    {
      "epoch": 0.07837858626259518,
      "grad_norm": 2.78125,
      "learning_rate": 2.8824321206061073e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 450016.0,
      "step": 1019
    },
    {
      "epoch": 0.07845550342281363,
      "grad_norm": 2.15625,
      "learning_rate": 2.8823167448657796e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9986772537231445,
      "num_tokens": 450774.0,
      "step": 1020
    },
    {
      "epoch": 0.07853242058303207,
      "grad_norm": 10.125,
      "learning_rate": 2.8822013691254523e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 451066.0,
      "step": 1021
    },
    {
      "epoch": 0.07860933774325052,
      "grad_norm": 11.75,
      "learning_rate": 2.882085993385124e-06,
      "loss": 0.0399,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 451394.0,
      "step": 1022
    },
    {
      "epoch": 0.07868625490346896,
      "grad_norm": 4.34375,
      "learning_rate": 2.8819706176447963e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 451956.0,
      "step": 1023
    },
    {
      "epoch": 0.07876317206368741,
      "grad_norm": 17.125,
      "learning_rate": 2.881855241904469e-06,
      "loss": 0.052,
      "mean_token_accuracy": 0.9789789915084839,
      "num_tokens": 452292.0,
      "step": 1024
    },
    {
      "epoch": 0.07884008922390585,
      "grad_norm": 4.46875,
      "learning_rate": 2.8817398661641413e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 452836.0,
      "step": 1025
    },
    {
      "epoch": 0.0789170063841243,
      "grad_norm": 9.875,
      "learning_rate": 2.8816244904238135e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 453176.0,
      "step": 1026
    },
    {
      "epoch": 0.07899392354434275,
      "grad_norm": 3.203125,
      "learning_rate": 2.8815091146834862e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 453746.0,
      "step": 1027
    },
    {
      "epoch": 0.07907084070456119,
      "grad_norm": 11.875,
      "learning_rate": 2.881393738943158e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 454110.0,
      "step": 1028
    },
    {
      "epoch": 0.07914775786477964,
      "grad_norm": 13.25,
      "learning_rate": 2.8812783632028307e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9789789915084839,
      "num_tokens": 454446.0,
      "step": 1029
    },
    {
      "epoch": 0.07922467502499808,
      "grad_norm": 3.546875,
      "learning_rate": 2.881162987462503e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 455008.0,
      "step": 1030
    },
    {
      "epoch": 0.07930159218521653,
      "grad_norm": 5.40625,
      "learning_rate": 2.8810476117221753e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 455292.0,
      "step": 1031
    },
    {
      "epoch": 0.07937850934543497,
      "grad_norm": 3.34375,
      "learning_rate": 2.880932235981848e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 455856.0,
      "step": 1032
    },
    {
      "epoch": 0.07945542650565342,
      "grad_norm": 5.875,
      "learning_rate": 2.8808168602415198e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 456140.0,
      "step": 1033
    },
    {
      "epoch": 0.07953234366587185,
      "grad_norm": 8.6875,
      "learning_rate": 2.880701484501192e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 456500.0,
      "step": 1034
    },
    {
      "epoch": 0.0796092608260903,
      "grad_norm": 8.4375,
      "learning_rate": 2.8805861087608647e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 456778.0,
      "step": 1035
    },
    {
      "epoch": 0.07968617798630874,
      "grad_norm": 7.0,
      "learning_rate": 2.880470733020537e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 457138.0,
      "step": 1036
    },
    {
      "epoch": 0.07976309514652719,
      "grad_norm": 2.765625,
      "learning_rate": 2.8803553572802092e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 457700.0,
      "step": 1037
    },
    {
      "epoch": 0.07984001230674563,
      "grad_norm": 7.53125,
      "learning_rate": 2.880239981539882e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 458034.0,
      "step": 1038
    },
    {
      "epoch": 0.07991692946696408,
      "grad_norm": 8.75,
      "learning_rate": 2.8801246057995537e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 458318.0,
      "step": 1039
    },
    {
      "epoch": 0.07999384662718252,
      "grad_norm": 5.0625,
      "learning_rate": 2.8800092300592264e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 458664.0,
      "step": 1040
    },
    {
      "epoch": 0.08007076378740097,
      "grad_norm": 6.125,
      "learning_rate": 2.8798938543188987e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 458990.0,
      "step": 1041
    },
    {
      "epoch": 0.08014768094761941,
      "grad_norm": 11.5,
      "learning_rate": 2.879778478578571e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 459272.0,
      "step": 1042
    },
    {
      "epoch": 0.08022459810783786,
      "grad_norm": 6.3125,
      "learning_rate": 2.879663102838243e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9842271208763123,
      "num_tokens": 459590.0,
      "step": 1043
    },
    {
      "epoch": 0.0803015152680563,
      "grad_norm": 6.78125,
      "learning_rate": 2.8795477270979155e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 459938.0,
      "step": 1044
    },
    {
      "epoch": 0.08037843242827475,
      "grad_norm": 4.03125,
      "learning_rate": 2.8794323513575877e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 460512.0,
      "step": 1045
    },
    {
      "epoch": 0.0804553495884932,
      "grad_norm": 7.5,
      "learning_rate": 2.8793169756172604e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 460818.0,
      "step": 1046
    },
    {
      "epoch": 0.08053226674871164,
      "grad_norm": 5.375,
      "learning_rate": 2.8792015998769327e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 461146.0,
      "step": 1047
    },
    {
      "epoch": 0.08060918390893008,
      "grad_norm": 5.78125,
      "learning_rate": 2.879086224136605e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 461470.0,
      "step": 1048
    },
    {
      "epoch": 0.08068610106914853,
      "grad_norm": 10.625,
      "learning_rate": 2.8789708483962776e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 461778.0,
      "step": 1049
    },
    {
      "epoch": 0.08076301822936698,
      "grad_norm": 4.46875,
      "learning_rate": 2.8788554726559494e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 462346.0,
      "step": 1050
    },
    {
      "epoch": 0.08083993538958542,
      "grad_norm": 2.984375,
      "learning_rate": 2.8787400969156217e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 462920.0,
      "step": 1051
    },
    {
      "epoch": 0.08091685254980387,
      "grad_norm": 5.53125,
      "learning_rate": 2.8786247211752944e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 463258.0,
      "step": 1052
    },
    {
      "epoch": 0.08099376971002231,
      "grad_norm": 5.1875,
      "learning_rate": 2.8785093454349666e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 463568.0,
      "step": 1053
    },
    {
      "epoch": 0.08107068687024076,
      "grad_norm": 5.96875,
      "learning_rate": 2.878393969694639e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 464094.0,
      "step": 1054
    },
    {
      "epoch": 0.08114760403045919,
      "grad_norm": 10.125,
      "learning_rate": 2.878278593954311e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 464462.0,
      "step": 1055
    },
    {
      "epoch": 0.08122452119067763,
      "grad_norm": 4.4375,
      "learning_rate": 2.8781632182139834e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 465018.0,
      "step": 1056
    },
    {
      "epoch": 0.08130143835089608,
      "grad_norm": 4.71875,
      "learning_rate": 2.878047842473656e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 465344.0,
      "step": 1057
    },
    {
      "epoch": 0.08137835551111453,
      "grad_norm": 6.25,
      "learning_rate": 2.8779324667333283e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 465650.0,
      "step": 1058
    },
    {
      "epoch": 0.08145527267133297,
      "grad_norm": 5.5625,
      "learning_rate": 2.8778170909930006e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 465990.0,
      "step": 1059
    },
    {
      "epoch": 0.08153218983155142,
      "grad_norm": 10.9375,
      "learning_rate": 2.877701715252673e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 466318.0,
      "step": 1060
    },
    {
      "epoch": 0.08160910699176986,
      "grad_norm": 7.96875,
      "learning_rate": 2.877586339512345e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 466638.0,
      "step": 1061
    },
    {
      "epoch": 0.0816860241519883,
      "grad_norm": 4.21875,
      "learning_rate": 2.8774709637720174e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 467434.0,
      "step": 1062
    },
    {
      "epoch": 0.08176294131220675,
      "grad_norm": 3.796875,
      "learning_rate": 2.87735558803169e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 468010.0,
      "step": 1063
    },
    {
      "epoch": 0.0818398584724252,
      "grad_norm": 7.3125,
      "learning_rate": 2.8772402122913623e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 468530.0,
      "step": 1064
    },
    {
      "epoch": 0.08191677563264364,
      "grad_norm": 13.4375,
      "learning_rate": 2.8771248365510346e-06,
      "loss": 0.0401,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 468802.0,
      "step": 1065
    },
    {
      "epoch": 0.08199369279286209,
      "grad_norm": 9.4375,
      "learning_rate": 2.877009460810707e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 469062.0,
      "step": 1066
    },
    {
      "epoch": 0.08207060995308053,
      "grad_norm": 3.203125,
      "learning_rate": 2.876894085070379e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 469636.0,
      "step": 1067
    },
    {
      "epoch": 0.08214752711329898,
      "grad_norm": 3.6875,
      "learning_rate": 2.8767787093300518e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 470192.0,
      "step": 1068
    },
    {
      "epoch": 0.08222444427351742,
      "grad_norm": 3.375,
      "learning_rate": 2.876663333589724e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 470760.0,
      "step": 1069
    },
    {
      "epoch": 0.08230136143373587,
      "grad_norm": 7.75,
      "learning_rate": 2.8765479578493963e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 471066.0,
      "step": 1070
    },
    {
      "epoch": 0.08237827859395432,
      "grad_norm": 6.46875,
      "learning_rate": 2.8764325821090685e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 471372.0,
      "step": 1071
    },
    {
      "epoch": 0.08245519575417276,
      "grad_norm": 3.078125,
      "learning_rate": 2.876317206368741e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 472186.0,
      "step": 1072
    },
    {
      "epoch": 0.0825321129143912,
      "grad_norm": 3.609375,
      "learning_rate": 2.876201830628413e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 472734.0,
      "step": 1073
    },
    {
      "epoch": 0.08260903007460965,
      "grad_norm": 8.375,
      "learning_rate": 2.8760864548880857e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 473040.0,
      "step": 1074
    },
    {
      "epoch": 0.08268594723482808,
      "grad_norm": 6.9375,
      "learning_rate": 2.875971079147758e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 473324.0,
      "step": 1075
    },
    {
      "epoch": 0.08276286439504653,
      "grad_norm": 5.5625,
      "learning_rate": 2.8758557034074303e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 473916.0,
      "step": 1076
    },
    {
      "epoch": 0.08283978155526497,
      "grad_norm": 9.75,
      "learning_rate": 2.8757403276671025e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 474254.0,
      "step": 1077
    },
    {
      "epoch": 0.08291669871548342,
      "grad_norm": 4.53125,
      "learning_rate": 2.8756249519267748e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 474824.0,
      "step": 1078
    },
    {
      "epoch": 0.08299361587570186,
      "grad_norm": 8.5625,
      "learning_rate": 2.8755095761864475e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 475106.0,
      "step": 1079
    },
    {
      "epoch": 0.08307053303592031,
      "grad_norm": 2.609375,
      "learning_rate": 2.8753942004461197e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 475890.0,
      "step": 1080
    },
    {
      "epoch": 0.08314745019613876,
      "grad_norm": 3.234375,
      "learning_rate": 2.875278824705792e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 476430.0,
      "step": 1081
    },
    {
      "epoch": 0.0832243673563572,
      "grad_norm": 6.59375,
      "learning_rate": 2.8751634489654642e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 476752.0,
      "step": 1082
    },
    {
      "epoch": 0.08330128451657565,
      "grad_norm": 9.0,
      "learning_rate": 2.8750480732251365e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 477086.0,
      "step": 1083
    },
    {
      "epoch": 0.08337820167679409,
      "grad_norm": 3.125,
      "learning_rate": 2.8749326974848087e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 477642.0,
      "step": 1084
    },
    {
      "epoch": 0.08345511883701254,
      "grad_norm": 3.21875,
      "learning_rate": 2.8748173217444814e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 478228.0,
      "step": 1085
    },
    {
      "epoch": 0.08353203599723098,
      "grad_norm": 8.4375,
      "learning_rate": 2.8747019460041537e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 478532.0,
      "step": 1086
    },
    {
      "epoch": 0.08360895315744943,
      "grad_norm": 4.1875,
      "learning_rate": 2.874586570263826e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 479120.0,
      "step": 1087
    },
    {
      "epoch": 0.08368587031766787,
      "grad_norm": 5.15625,
      "learning_rate": 2.874471194523498e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 479636.0,
      "step": 1088
    },
    {
      "epoch": 0.08376278747788632,
      "grad_norm": 8.125,
      "learning_rate": 2.8743558187831705e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 479980.0,
      "step": 1089
    },
    {
      "epoch": 0.08383970463810476,
      "grad_norm": 3.53125,
      "learning_rate": 2.874240443042843e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 480520.0,
      "step": 1090
    },
    {
      "epoch": 0.08391662179832321,
      "grad_norm": 8.75,
      "learning_rate": 2.8741250673025154e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 480816.0,
      "step": 1091
    },
    {
      "epoch": 0.08399353895854166,
      "grad_norm": 10.125,
      "learning_rate": 2.8740096915621877e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 481166.0,
      "step": 1092
    },
    {
      "epoch": 0.0840704561187601,
      "grad_norm": 4.3125,
      "learning_rate": 2.87389431582186e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 481740.0,
      "step": 1093
    },
    {
      "epoch": 0.08414737327897855,
      "grad_norm": 6.625,
      "learning_rate": 2.873778940081532e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 482074.0,
      "step": 1094
    },
    {
      "epoch": 0.08422429043919699,
      "grad_norm": 8.875,
      "learning_rate": 2.8736635643412044e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 482362.0,
      "step": 1095
    },
    {
      "epoch": 0.08430120759941542,
      "grad_norm": 6.78125,
      "learning_rate": 2.873548188600877e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9945054650306702,
      "num_tokens": 482728.0,
      "step": 1096
    },
    {
      "epoch": 0.08437812475963387,
      "grad_norm": 7.65625,
      "learning_rate": 2.8734328128605494e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 483040.0,
      "step": 1097
    },
    {
      "epoch": 0.08445504191985231,
      "grad_norm": 10.0625,
      "learning_rate": 2.8733174371202216e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 483348.0,
      "step": 1098
    },
    {
      "epoch": 0.08453195908007076,
      "grad_norm": 7.15625,
      "learning_rate": 2.873202061379894e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 483678.0,
      "step": 1099
    },
    {
      "epoch": 0.0846088762402892,
      "grad_norm": 11.0625,
      "learning_rate": 2.873086685639566e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 484056.0,
      "step": 1100
    },
    {
      "epoch": 0.08468579340050765,
      "grad_norm": 7.625,
      "learning_rate": 2.8729713098992384e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 484360.0,
      "step": 1101
    },
    {
      "epoch": 0.0847627105607261,
      "grad_norm": 3.109375,
      "learning_rate": 2.872855934158911e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 484920.0,
      "step": 1102
    },
    {
      "epoch": 0.08483962772094454,
      "grad_norm": 5.75,
      "learning_rate": 2.8727405584185833e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 485232.0,
      "step": 1103
    },
    {
      "epoch": 0.08491654488116299,
      "grad_norm": 11.0,
      "learning_rate": 2.8726251826782556e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9919571280479431,
      "num_tokens": 485608.0,
      "step": 1104
    },
    {
      "epoch": 0.08499346204138143,
      "grad_norm": 13.1875,
      "learning_rate": 2.872509806937928e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9906250238418579,
      "num_tokens": 485932.0,
      "step": 1105
    },
    {
      "epoch": 0.08507037920159988,
      "grad_norm": 3.4375,
      "learning_rate": 2.8723944311976e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 486512.0,
      "step": 1106
    },
    {
      "epoch": 0.08514729636181832,
      "grad_norm": 4.0625,
      "learning_rate": 2.872279055457273e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 486862.0,
      "step": 1107
    },
    {
      "epoch": 0.08522421352203677,
      "grad_norm": 4.75,
      "learning_rate": 2.872163679716945e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 487464.0,
      "step": 1108
    },
    {
      "epoch": 0.08530113068225521,
      "grad_norm": 5.125,
      "learning_rate": 2.872048303976617e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 487792.0,
      "step": 1109
    },
    {
      "epoch": 0.08537804784247366,
      "grad_norm": 10.75,
      "learning_rate": 2.8719329282362896e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 488076.0,
      "step": 1110
    },
    {
      "epoch": 0.0854549650026921,
      "grad_norm": 4.96875,
      "learning_rate": 2.871817552495962e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 488600.0,
      "step": 1111
    },
    {
      "epoch": 0.08553188216291055,
      "grad_norm": 3.28125,
      "learning_rate": 2.871702176755634e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 489160.0,
      "step": 1112
    },
    {
      "epoch": 0.085608799323129,
      "grad_norm": 5.75,
      "learning_rate": 2.8715868010153068e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 489508.0,
      "step": 1113
    },
    {
      "epoch": 0.08568571648334744,
      "grad_norm": 2.796875,
      "learning_rate": 2.871471425274979e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 490070.0,
      "step": 1114
    },
    {
      "epoch": 0.08576263364356589,
      "grad_norm": 9.625,
      "learning_rate": 2.8713560495346513e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 490352.0,
      "step": 1115
    },
    {
      "epoch": 0.08583955080378432,
      "grad_norm": 5.6875,
      "learning_rate": 2.8712406737943235e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9940476417541504,
      "num_tokens": 490690.0,
      "step": 1116
    },
    {
      "epoch": 0.08591646796400276,
      "grad_norm": 10.0625,
      "learning_rate": 2.871125298053996e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 490996.0,
      "step": 1117
    },
    {
      "epoch": 0.08599338512422121,
      "grad_norm": 9.0625,
      "learning_rate": 2.8710099223136685e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 491336.0,
      "step": 1118
    },
    {
      "epoch": 0.08607030228443965,
      "grad_norm": 2.25,
      "learning_rate": 2.8708945465733407e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 491898.0,
      "step": 1119
    },
    {
      "epoch": 0.0861472194446581,
      "grad_norm": 5.34375,
      "learning_rate": 2.8707791708330126e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 492478.0,
      "step": 1120
    },
    {
      "epoch": 0.08622413660487654,
      "grad_norm": 13.4375,
      "learning_rate": 2.8706637950926853e-06,
      "loss": 0.0393,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 492754.0,
      "step": 1121
    },
    {
      "epoch": 0.08630105376509499,
      "grad_norm": 10.0,
      "learning_rate": 2.8705484193523575e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 493062.0,
      "step": 1122
    },
    {
      "epoch": 0.08637797092531344,
      "grad_norm": 8.25,
      "learning_rate": 2.8704330436120298e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 493368.0,
      "step": 1123
    },
    {
      "epoch": 0.08645488808553188,
      "grad_norm": 6.46875,
      "learning_rate": 2.8703176678717025e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 493668.0,
      "step": 1124
    },
    {
      "epoch": 0.08653180524575033,
      "grad_norm": 3.640625,
      "learning_rate": 2.8702022921313743e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 494210.0,
      "step": 1125
    },
    {
      "epoch": 0.08660872240596877,
      "grad_norm": 6.3125,
      "learning_rate": 2.870086916391047e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 494544.0,
      "step": 1126
    },
    {
      "epoch": 0.08668563956618722,
      "grad_norm": 3.1875,
      "learning_rate": 2.8699715406507192e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 495106.0,
      "step": 1127
    },
    {
      "epoch": 0.08676255672640566,
      "grad_norm": 6.53125,
      "learning_rate": 2.8698561649103915e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 495456.0,
      "step": 1128
    },
    {
      "epoch": 0.08683947388662411,
      "grad_norm": 3.546875,
      "learning_rate": 2.869740789170064e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 496046.0,
      "step": 1129
    },
    {
      "epoch": 0.08691639104684255,
      "grad_norm": 3.921875,
      "learning_rate": 2.8696254134297364e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 496564.0,
      "step": 1130
    },
    {
      "epoch": 0.086993308207061,
      "grad_norm": 4.6875,
      "learning_rate": 2.8695100376894083e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 496876.0,
      "step": 1131
    },
    {
      "epoch": 0.08707022536727944,
      "grad_norm": 3.90625,
      "learning_rate": 2.869394661949081e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 497416.0,
      "step": 1132
    },
    {
      "epoch": 0.08714714252749789,
      "grad_norm": 4.0625,
      "learning_rate": 2.869279286208753e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 497934.0,
      "step": 1133
    },
    {
      "epoch": 0.08722405968771633,
      "grad_norm": 3.390625,
      "learning_rate": 2.8691639104684255e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 498456.0,
      "step": 1134
    },
    {
      "epoch": 0.08730097684793478,
      "grad_norm": 5.9375,
      "learning_rate": 2.869048534728098e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 498802.0,
      "step": 1135
    },
    {
      "epoch": 0.08737789400815323,
      "grad_norm": 13.6875,
      "learning_rate": 2.86893315898777e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 499034.0,
      "step": 1136
    },
    {
      "epoch": 0.08745481116837166,
      "grad_norm": 9.8125,
      "learning_rate": 2.8688177832474427e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 499292.0,
      "step": 1137
    },
    {
      "epoch": 0.0875317283285901,
      "grad_norm": 2.953125,
      "learning_rate": 2.868702407507115e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 500106.0,
      "step": 1138
    },
    {
      "epoch": 0.08760864548880855,
      "grad_norm": 3.5625,
      "learning_rate": 2.868587031766787e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 500668.0,
      "step": 1139
    },
    {
      "epoch": 0.087685562649027,
      "grad_norm": 12.875,
      "learning_rate": 2.86847165602646e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 500930.0,
      "step": 1140
    },
    {
      "epoch": 0.08776247980924544,
      "grad_norm": 4.75,
      "learning_rate": 2.868356280286132e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 501456.0,
      "step": 1141
    },
    {
      "epoch": 0.08783939696946388,
      "grad_norm": 4.28125,
      "learning_rate": 2.868240904545804e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 501972.0,
      "step": 1142
    },
    {
      "epoch": 0.08791631412968233,
      "grad_norm": 6.46875,
      "learning_rate": 2.8681255288054766e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 502294.0,
      "step": 1143
    },
    {
      "epoch": 0.08799323128990078,
      "grad_norm": 7.4375,
      "learning_rate": 2.868010153065149e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 502640.0,
      "step": 1144
    },
    {
      "epoch": 0.08807014845011922,
      "grad_norm": 5.6875,
      "learning_rate": 2.867894777324821e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 502922.0,
      "step": 1145
    },
    {
      "epoch": 0.08814706561033767,
      "grad_norm": 5.28125,
      "learning_rate": 2.867779401584494e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9894551634788513,
      "num_tokens": 503492.0,
      "step": 1146
    },
    {
      "epoch": 0.08822398277055611,
      "grad_norm": 6.8125,
      "learning_rate": 2.8676640258441657e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 503816.0,
      "step": 1147
    },
    {
      "epoch": 0.08830089993077456,
      "grad_norm": 7.1875,
      "learning_rate": 2.8675486501038383e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 504116.0,
      "step": 1148
    },
    {
      "epoch": 0.088377817090993,
      "grad_norm": 9.625,
      "learning_rate": 2.8674332743635106e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 504444.0,
      "step": 1149
    },
    {
      "epoch": 0.08845473425121145,
      "grad_norm": 3.796875,
      "learning_rate": 2.867317898623183e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 505000.0,
      "step": 1150
    },
    {
      "epoch": 0.08853165141142989,
      "grad_norm": 13.5625,
      "learning_rate": 2.8672025228828555e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9868995547294617,
      "num_tokens": 505230.0,
      "step": 1151
    },
    {
      "epoch": 0.08860856857164834,
      "grad_norm": 8.125,
      "learning_rate": 2.867087147142528e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 505590.0,
      "step": 1152
    },
    {
      "epoch": 0.08868548573186678,
      "grad_norm": 7.40625,
      "learning_rate": 2.8669717714021996e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 505886.0,
      "step": 1153
    },
    {
      "epoch": 0.08876240289208523,
      "grad_norm": 3.765625,
      "learning_rate": 2.8668563956618723e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 506472.0,
      "step": 1154
    },
    {
      "epoch": 0.08883932005230367,
      "grad_norm": 2.75,
      "learning_rate": 2.8667410199215446e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 507032.0,
      "step": 1155
    },
    {
      "epoch": 0.08891623721252212,
      "grad_norm": 5.75,
      "learning_rate": 2.866625644181217e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 507348.0,
      "step": 1156
    },
    {
      "epoch": 0.08899315437274055,
      "grad_norm": 11.625,
      "learning_rate": 2.8665102684408895e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 507656.0,
      "step": 1157
    },
    {
      "epoch": 0.089070071532959,
      "grad_norm": 5.375,
      "learning_rate": 2.8663948927005614e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 508202.0,
      "step": 1158
    },
    {
      "epoch": 0.08914698869317744,
      "grad_norm": 13.125,
      "learning_rate": 2.866279516960234e-06,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 508458.0,
      "step": 1159
    },
    {
      "epoch": 0.08922390585339589,
      "grad_norm": 12.1875,
      "learning_rate": 2.8661641412199063e-06,
      "loss": 0.038,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 508742.0,
      "step": 1160
    },
    {
      "epoch": 0.08930082301361433,
      "grad_norm": 3.640625,
      "learning_rate": 2.8660487654795786e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 509332.0,
      "step": 1161
    },
    {
      "epoch": 0.08937774017383278,
      "grad_norm": 10.375,
      "learning_rate": 2.865933389739251e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 509640.0,
      "step": 1162
    },
    {
      "epoch": 0.08945465733405122,
      "grad_norm": 9.3125,
      "learning_rate": 2.865818013998923e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 509942.0,
      "step": 1163
    },
    {
      "epoch": 0.08953157449426967,
      "grad_norm": 4.25,
      "learning_rate": 2.8657026382585953e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 510262.0,
      "step": 1164
    },
    {
      "epoch": 0.08960849165448811,
      "grad_norm": 5.34375,
      "learning_rate": 2.865587262518268e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 510568.0,
      "step": 1165
    },
    {
      "epoch": 0.08968540881470656,
      "grad_norm": 8.1875,
      "learning_rate": 2.8654718867779403e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 510880.0,
      "step": 1166
    },
    {
      "epoch": 0.089762325974925,
      "grad_norm": 4.28125,
      "learning_rate": 2.8653565110376125e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 511444.0,
      "step": 1167
    },
    {
      "epoch": 0.08983924313514345,
      "grad_norm": 4.875,
      "learning_rate": 2.865241135297285e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 511964.0,
      "step": 1168
    },
    {
      "epoch": 0.0899161602953619,
      "grad_norm": 3.65625,
      "learning_rate": 2.865125759556957e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 512534.0,
      "step": 1169
    },
    {
      "epoch": 0.08999307745558034,
      "grad_norm": 4.9375,
      "learning_rate": 2.8650103838166293e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 513294.0,
      "step": 1170
    },
    {
      "epoch": 0.09006999461579879,
      "grad_norm": 8.75,
      "learning_rate": 2.864895008076302e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 513582.0,
      "step": 1171
    },
    {
      "epoch": 0.09014691177601723,
      "grad_norm": 4.4375,
      "learning_rate": 2.8647796323359742e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 514098.0,
      "step": 1172
    },
    {
      "epoch": 0.09022382893623568,
      "grad_norm": 7.84375,
      "learning_rate": 2.8646642565956465e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 514448.0,
      "step": 1173
    },
    {
      "epoch": 0.09030074609645412,
      "grad_norm": 5.0,
      "learning_rate": 2.8645488808553188e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 514754.0,
      "step": 1174
    },
    {
      "epoch": 0.09037766325667257,
      "grad_norm": 3.734375,
      "learning_rate": 2.864433505114991e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 515338.0,
      "step": 1175
    },
    {
      "epoch": 0.09045458041689101,
      "grad_norm": 6.125,
      "learning_rate": 2.8643181293746637e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 515650.0,
      "step": 1176
    },
    {
      "epoch": 0.09053149757710946,
      "grad_norm": 6.5,
      "learning_rate": 2.864202753634336e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 515968.0,
      "step": 1177
    },
    {
      "epoch": 0.09060841473732789,
      "grad_norm": 3.140625,
      "learning_rate": 2.864087377894008e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 516726.0,
      "step": 1178
    },
    {
      "epoch": 0.09068533189754634,
      "grad_norm": 9.6875,
      "learning_rate": 2.863972002153681e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 517062.0,
      "step": 1179
    },
    {
      "epoch": 0.09076224905776478,
      "grad_norm": 9.0,
      "learning_rate": 2.8638566264133527e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 517318.0,
      "step": 1180
    },
    {
      "epoch": 0.09083916621798323,
      "grad_norm": 3.65625,
      "learning_rate": 2.863741250673025e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 517836.0,
      "step": 1181
    },
    {
      "epoch": 0.09091608337820167,
      "grad_norm": 4.46875,
      "learning_rate": 2.8636258749326977e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 518392.0,
      "step": 1182
    },
    {
      "epoch": 0.09099300053842012,
      "grad_norm": 4.3125,
      "learning_rate": 2.86351049919237e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 518954.0,
      "step": 1183
    },
    {
      "epoch": 0.09106991769863856,
      "grad_norm": 4.46875,
      "learning_rate": 2.863395123452042e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 519770.0,
      "step": 1184
    },
    {
      "epoch": 0.09114683485885701,
      "grad_norm": 8.0,
      "learning_rate": 2.8632797477117144e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 520092.0,
      "step": 1185
    },
    {
      "epoch": 0.09122375201907545,
      "grad_norm": 5.4375,
      "learning_rate": 2.8631643719713867e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 520644.0,
      "step": 1186
    },
    {
      "epoch": 0.0913006691792939,
      "grad_norm": 9.125,
      "learning_rate": 2.8630489962310594e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 520946.0,
      "step": 1187
    },
    {
      "epoch": 0.09137758633951235,
      "grad_norm": 8.75,
      "learning_rate": 2.8629336204907316e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9828178882598877,
      "num_tokens": 521238.0,
      "step": 1188
    },
    {
      "epoch": 0.09145450349973079,
      "grad_norm": 2.6875,
      "learning_rate": 2.862818244750404e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 522020.0,
      "step": 1189
    },
    {
      "epoch": 0.09153142065994924,
      "grad_norm": 9.125,
      "learning_rate": 2.8627028690100766e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 522274.0,
      "step": 1190
    },
    {
      "epoch": 0.09160833782016768,
      "grad_norm": 8.875,
      "learning_rate": 2.8625874932697484e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 522584.0,
      "step": 1191
    },
    {
      "epoch": 0.09168525498038613,
      "grad_norm": 8.8125,
      "learning_rate": 2.8624721175294207e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9826989769935608,
      "num_tokens": 522874.0,
      "step": 1192
    },
    {
      "epoch": 0.09176217214060457,
      "grad_norm": 3.046875,
      "learning_rate": 2.8623567417890934e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 523420.0,
      "step": 1193
    },
    {
      "epoch": 0.09183908930082302,
      "grad_norm": 8.25,
      "learning_rate": 2.8622413660487656e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 523764.0,
      "step": 1194
    },
    {
      "epoch": 0.09191600646104146,
      "grad_norm": 3.4375,
      "learning_rate": 2.862125990308438e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 524316.0,
      "step": 1195
    },
    {
      "epoch": 0.09199292362125991,
      "grad_norm": 8.75,
      "learning_rate": 2.86201061456811e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 524622.0,
      "step": 1196
    },
    {
      "epoch": 0.09206984078147835,
      "grad_norm": 2.453125,
      "learning_rate": 2.8618952388277824e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9973545074462891,
      "num_tokens": 525380.0,
      "step": 1197
    },
    {
      "epoch": 0.09214675794169679,
      "grad_norm": 3.484375,
      "learning_rate": 2.861779863087455e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 525960.0,
      "step": 1198
    },
    {
      "epoch": 0.09222367510191523,
      "grad_norm": 6.34375,
      "learning_rate": 2.8616644873471273e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 526254.0,
      "step": 1199
    },
    {
      "epoch": 0.09230059226213368,
      "grad_norm": 9.25,
      "learning_rate": 2.8615491116067996e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 526554.0,
      "step": 1200
    },
    {
      "epoch": 0.09237750942235212,
      "grad_norm": 6.15625,
      "learning_rate": 2.861433735866472e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 526832.0,
      "step": 1201
    },
    {
      "epoch": 0.09245442658257057,
      "grad_norm": 5.0,
      "learning_rate": 2.861318360126144e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 527410.0,
      "step": 1202
    },
    {
      "epoch": 0.09253134374278901,
      "grad_norm": 8.375,
      "learning_rate": 2.8612029843858164e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9808306694030762,
      "num_tokens": 527724.0,
      "step": 1203
    },
    {
      "epoch": 0.09260826090300746,
      "grad_norm": 3.28125,
      "learning_rate": 2.861087608645489e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 528282.0,
      "step": 1204
    },
    {
      "epoch": 0.0926851780632259,
      "grad_norm": 3.515625,
      "learning_rate": 2.8609722329051613e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 528832.0,
      "step": 1205
    },
    {
      "epoch": 0.09276209522344435,
      "grad_norm": 10.375,
      "learning_rate": 2.8608568571648336e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 529120.0,
      "step": 1206
    },
    {
      "epoch": 0.0928390123836628,
      "grad_norm": 8.5625,
      "learning_rate": 2.860741481424506e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 529482.0,
      "step": 1207
    },
    {
      "epoch": 0.09291592954388124,
      "grad_norm": 6.21875,
      "learning_rate": 2.860626105684178e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 529806.0,
      "step": 1208
    },
    {
      "epoch": 0.09299284670409969,
      "grad_norm": 3.1875,
      "learning_rate": 2.8605107299438508e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 530362.0,
      "step": 1209
    },
    {
      "epoch": 0.09306976386431813,
      "grad_norm": 6.0,
      "learning_rate": 2.860395354203523e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 530666.0,
      "step": 1210
    },
    {
      "epoch": 0.09314668102453658,
      "grad_norm": 4.59375,
      "learning_rate": 2.8602799784631953e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 531182.0,
      "step": 1211
    },
    {
      "epoch": 0.09322359818475502,
      "grad_norm": 3.65625,
      "learning_rate": 2.8601646027228675e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 531734.0,
      "step": 1212
    },
    {
      "epoch": 0.09330051534497347,
      "grad_norm": 3.265625,
      "learning_rate": 2.8600492269825398e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 532522.0,
      "step": 1213
    },
    {
      "epoch": 0.09337743250519191,
      "grad_norm": 7.40625,
      "learning_rate": 2.859933851242212e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 532814.0,
      "step": 1214
    },
    {
      "epoch": 0.09345434966541036,
      "grad_norm": 8.6875,
      "learning_rate": 2.8598184755018847e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 533110.0,
      "step": 1215
    },
    {
      "epoch": 0.0935312668256288,
      "grad_norm": 8.375,
      "learning_rate": 2.859703099761557e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 533396.0,
      "step": 1216
    },
    {
      "epoch": 0.09360818398584725,
      "grad_norm": 3.90625,
      "learning_rate": 2.8595877240212292e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 533946.0,
      "step": 1217
    },
    {
      "epoch": 0.0936851011460657,
      "grad_norm": 7.65625,
      "learning_rate": 2.8594723482809015e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 534254.0,
      "step": 1218
    },
    {
      "epoch": 0.09376201830628413,
      "grad_norm": 3.515625,
      "learning_rate": 2.8593569725405738e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 534824.0,
      "step": 1219
    },
    {
      "epoch": 0.09383893546650257,
      "grad_norm": 4.125,
      "learning_rate": 2.859241596800246e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 535342.0,
      "step": 1220
    },
    {
      "epoch": 0.09391585262672102,
      "grad_norm": 3.8125,
      "learning_rate": 2.8591262210599187e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 535912.0,
      "step": 1221
    },
    {
      "epoch": 0.09399276978693946,
      "grad_norm": 13.0625,
      "learning_rate": 2.859010845319591e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 536240.0,
      "step": 1222
    },
    {
      "epoch": 0.09406968694715791,
      "grad_norm": 3.890625,
      "learning_rate": 2.8588954695792632e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 536820.0,
      "step": 1223
    },
    {
      "epoch": 0.09414660410737635,
      "grad_norm": 4.9375,
      "learning_rate": 2.8587800938389355e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 537400.0,
      "step": 1224
    },
    {
      "epoch": 0.0942235212675948,
      "grad_norm": 7.625,
      "learning_rate": 2.8586647180986077e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 537706.0,
      "step": 1225
    },
    {
      "epoch": 0.09430043842781324,
      "grad_norm": 6.75,
      "learning_rate": 2.8585493423582804e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 538028.0,
      "step": 1226
    },
    {
      "epoch": 0.09437735558803169,
      "grad_norm": 9.9375,
      "learning_rate": 2.8584339666179527e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 538360.0,
      "step": 1227
    },
    {
      "epoch": 0.09445427274825013,
      "grad_norm": 6.6875,
      "learning_rate": 2.8583185908776245e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 538678.0,
      "step": 1228
    },
    {
      "epoch": 0.09453118990846858,
      "grad_norm": 14.125,
      "learning_rate": 2.858203215137297e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 538978.0,
      "step": 1229
    },
    {
      "epoch": 0.09460810706868703,
      "grad_norm": 8.25,
      "learning_rate": 2.8580878393969694e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 539286.0,
      "step": 1230
    },
    {
      "epoch": 0.09468502422890547,
      "grad_norm": 4.59375,
      "learning_rate": 2.8579724636566417e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 539620.0,
      "step": 1231
    },
    {
      "epoch": 0.09476194138912392,
      "grad_norm": 3.046875,
      "learning_rate": 2.8578570879163144e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 540164.0,
      "step": 1232
    },
    {
      "epoch": 0.09483885854934236,
      "grad_norm": 3.71875,
      "learning_rate": 2.8577417121759866e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 540704.0,
      "step": 1233
    },
    {
      "epoch": 0.0949157757095608,
      "grad_norm": 3.546875,
      "learning_rate": 2.857626336435659e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 541290.0,
      "step": 1234
    },
    {
      "epoch": 0.09499269286977925,
      "grad_norm": 11.875,
      "learning_rate": 2.857510960695331e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 541658.0,
      "step": 1235
    },
    {
      "epoch": 0.0950696100299977,
      "grad_norm": 8.9375,
      "learning_rate": 2.8573955849550034e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 542156.0,
      "step": 1236
    },
    {
      "epoch": 0.09514652719021614,
      "grad_norm": 5.71875,
      "learning_rate": 2.857280209214676e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 542500.0,
      "step": 1237
    },
    {
      "epoch": 0.09522344435043459,
      "grad_norm": 4.6875,
      "learning_rate": 2.8571648334743484e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 543016.0,
      "step": 1238
    },
    {
      "epoch": 0.09530036151065302,
      "grad_norm": 3.359375,
      "learning_rate": 2.85704945773402e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 543600.0,
      "step": 1239
    },
    {
      "epoch": 0.09537727867087147,
      "grad_norm": 5.28125,
      "learning_rate": 2.856934081993693e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 543918.0,
      "step": 1240
    },
    {
      "epoch": 0.09545419583108991,
      "grad_norm": 5.125,
      "learning_rate": 2.856818706253365e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 544514.0,
      "step": 1241
    },
    {
      "epoch": 0.09553111299130836,
      "grad_norm": 8.1875,
      "learning_rate": 2.8567033305130374e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 544846.0,
      "step": 1242
    },
    {
      "epoch": 0.0956080301515268,
      "grad_norm": 3.625,
      "learning_rate": 2.85658795477271e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 545628.0,
      "step": 1243
    },
    {
      "epoch": 0.09568494731174525,
      "grad_norm": 7.3125,
      "learning_rate": 2.8564725790323823e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9972899556159973,
      "num_tokens": 546000.0,
      "step": 1244
    },
    {
      "epoch": 0.09576186447196369,
      "grad_norm": 4.65625,
      "learning_rate": 2.8563572032920546e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 546544.0,
      "step": 1245
    },
    {
      "epoch": 0.09583878163218214,
      "grad_norm": 3.484375,
      "learning_rate": 2.856241827551727e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 547564.0,
      "step": 1246
    },
    {
      "epoch": 0.09591569879240058,
      "grad_norm": 13.625,
      "learning_rate": 2.856126451811399e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 547894.0,
      "step": 1247
    },
    {
      "epoch": 0.09599261595261903,
      "grad_norm": 5.53125,
      "learning_rate": 2.8560110760710718e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 548458.0,
      "step": 1248
    },
    {
      "epoch": 0.09606953311283747,
      "grad_norm": 9.375,
      "learning_rate": 2.855895700330744e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 548754.0,
      "step": 1249
    },
    {
      "epoch": 0.09614645027305592,
      "grad_norm": 4.125,
      "learning_rate": 2.855780324590416e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 549324.0,
      "step": 1250
    },
    {
      "epoch": 0.09622336743327436,
      "grad_norm": 9.625,
      "learning_rate": 2.8556649488500886e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 549636.0,
      "step": 1251
    },
    {
      "epoch": 0.09630028459349281,
      "grad_norm": 11.0625,
      "learning_rate": 2.855549573109761e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 549926.0,
      "step": 1252
    },
    {
      "epoch": 0.09637720175371126,
      "grad_norm": 2.203125,
      "learning_rate": 2.855434197369433e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 550750.0,
      "step": 1253
    },
    {
      "epoch": 0.0964541189139297,
      "grad_norm": 8.0625,
      "learning_rate": 2.8553188216291058e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 551088.0,
      "step": 1254
    },
    {
      "epoch": 0.09653103607414815,
      "grad_norm": 4.625,
      "learning_rate": 2.855203445888778e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 551630.0,
      "step": 1255
    },
    {
      "epoch": 0.09660795323436659,
      "grad_norm": 5.03125,
      "learning_rate": 2.8550880701484503e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9942857027053833,
      "num_tokens": 552156.0,
      "step": 1256
    },
    {
      "epoch": 0.09668487039458504,
      "grad_norm": 8.625,
      "learning_rate": 2.8549726944081225e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 552494.0,
      "step": 1257
    },
    {
      "epoch": 0.09676178755480348,
      "grad_norm": 10.6875,
      "learning_rate": 2.8548573186677948e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 552866.0,
      "step": 1258
    },
    {
      "epoch": 0.09683870471502193,
      "grad_norm": 3.0625,
      "learning_rate": 2.8547419429274675e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.993842363357544,
      "num_tokens": 553680.0,
      "step": 1259
    },
    {
      "epoch": 0.09691562187524036,
      "grad_norm": 8.25,
      "learning_rate": 2.8546265671871397e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 553998.0,
      "step": 1260
    },
    {
      "epoch": 0.0969925390354588,
      "grad_norm": 2.875,
      "learning_rate": 2.8545111914468116e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 554812.0,
      "step": 1261
    },
    {
      "epoch": 0.09706945619567725,
      "grad_norm": 3.484375,
      "learning_rate": 2.8543958157064842e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9912126660346985,
      "num_tokens": 555382.0,
      "step": 1262
    },
    {
      "epoch": 0.0971463733558957,
      "grad_norm": 3.171875,
      "learning_rate": 2.8542804399661565e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 555928.0,
      "step": 1263
    },
    {
      "epoch": 0.09722329051611414,
      "grad_norm": 2.953125,
      "learning_rate": 2.8541650642258288e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 556748.0,
      "step": 1264
    },
    {
      "epoch": 0.09730020767633259,
      "grad_norm": 2.96875,
      "learning_rate": 2.8540496884855014e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 557568.0,
      "step": 1265
    },
    {
      "epoch": 0.09737712483655103,
      "grad_norm": 11.0625,
      "learning_rate": 2.8539343127451733e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 557878.0,
      "step": 1266
    },
    {
      "epoch": 0.09745404199676948,
      "grad_norm": 10.25,
      "learning_rate": 2.853818937004846e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9804560542106628,
      "num_tokens": 558186.0,
      "step": 1267
    },
    {
      "epoch": 0.09753095915698792,
      "grad_norm": 4.59375,
      "learning_rate": 2.8537035612645182e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 558704.0,
      "step": 1268
    },
    {
      "epoch": 0.09760787631720637,
      "grad_norm": 6.1875,
      "learning_rate": 2.8535881855241905e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 559038.0,
      "step": 1269
    },
    {
      "epoch": 0.09768479347742481,
      "grad_norm": 11.4375,
      "learning_rate": 2.8534728097838627e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 559318.0,
      "step": 1270
    },
    {
      "epoch": 0.09776171063764326,
      "grad_norm": 4.28125,
      "learning_rate": 2.8533574340435354e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 559836.0,
      "step": 1271
    },
    {
      "epoch": 0.0978386277978617,
      "grad_norm": 12.375,
      "learning_rate": 2.8532420583032072e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 560112.0,
      "step": 1272
    },
    {
      "epoch": 0.09791554495808015,
      "grad_norm": 3.359375,
      "learning_rate": 2.85312668256288e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9908925294876099,
      "num_tokens": 560662.0,
      "step": 1273
    },
    {
      "epoch": 0.0979924621182986,
      "grad_norm": 5.84375,
      "learning_rate": 2.853011306822552e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 561012.0,
      "step": 1274
    },
    {
      "epoch": 0.09806937927851704,
      "grad_norm": 8.375,
      "learning_rate": 2.8528959310822244e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9892086386680603,
      "num_tokens": 561292.0,
      "step": 1275
    },
    {
      "epoch": 0.09814629643873549,
      "grad_norm": 8.5,
      "learning_rate": 2.852780555341897e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 561602.0,
      "step": 1276
    },
    {
      "epoch": 0.09822321359895393,
      "grad_norm": 12.5625,
      "learning_rate": 2.852665179601569e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 561960.0,
      "step": 1277
    },
    {
      "epoch": 0.09830013075917238,
      "grad_norm": 9.4375,
      "learning_rate": 2.8525498038612412e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 562298.0,
      "step": 1278
    },
    {
      "epoch": 0.09837704791939082,
      "grad_norm": 9.0,
      "learning_rate": 2.852434428120914e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 562612.0,
      "step": 1279
    },
    {
      "epoch": 0.09845396507960925,
      "grad_norm": 5.09375,
      "learning_rate": 2.852319052380586e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 563188.0,
      "step": 1280
    },
    {
      "epoch": 0.0985308822398277,
      "grad_norm": 3.78125,
      "learning_rate": 2.8522036766402584e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 563774.0,
      "step": 1281
    },
    {
      "epoch": 0.09860779940004614,
      "grad_norm": 7.6875,
      "learning_rate": 2.852088300899931e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 564064.0,
      "step": 1282
    },
    {
      "epoch": 0.09868471656026459,
      "grad_norm": 7.5625,
      "learning_rate": 2.851972925159603e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9840255379676819,
      "num_tokens": 564378.0,
      "step": 1283
    },
    {
      "epoch": 0.09876163372048304,
      "grad_norm": 5.90625,
      "learning_rate": 2.8518575494192756e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 564724.0,
      "step": 1284
    },
    {
      "epoch": 0.09883855088070148,
      "grad_norm": 13.25,
      "learning_rate": 2.851742173678948e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.978723406791687,
      "num_tokens": 565008.0,
      "step": 1285
    },
    {
      "epoch": 0.09891546804091993,
      "grad_norm": 6.53125,
      "learning_rate": 2.85162679793862e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 565604.0,
      "step": 1286
    },
    {
      "epoch": 0.09899238520113837,
      "grad_norm": 6.59375,
      "learning_rate": 2.851511422198293e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 566144.0,
      "step": 1287
    },
    {
      "epoch": 0.09906930236135682,
      "grad_norm": 10.375,
      "learning_rate": 2.8513960464579646e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 566480.0,
      "step": 1288
    },
    {
      "epoch": 0.09914621952157526,
      "grad_norm": 6.875,
      "learning_rate": 2.851280670717637e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 566976.0,
      "step": 1289
    },
    {
      "epoch": 0.09922313668179371,
      "grad_norm": 3.25,
      "learning_rate": 2.8511652949773096e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 567546.0,
      "step": 1290
    },
    {
      "epoch": 0.09930005384201215,
      "grad_norm": 4.34375,
      "learning_rate": 2.851049919236982e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 567896.0,
      "step": 1291
    },
    {
      "epoch": 0.0993769710022306,
      "grad_norm": 3.796875,
      "learning_rate": 2.850934543496654e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 568444.0,
      "step": 1292
    },
    {
      "epoch": 0.09945388816244904,
      "grad_norm": 5.375,
      "learning_rate": 2.8508191677563268e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 568778.0,
      "step": 1293
    },
    {
      "epoch": 0.09953080532266749,
      "grad_norm": 4.75,
      "learning_rate": 2.8507037920159986e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 569326.0,
      "step": 1294
    },
    {
      "epoch": 0.09960772248288594,
      "grad_norm": 5.03125,
      "learning_rate": 2.8505884162756713e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 569636.0,
      "step": 1295
    },
    {
      "epoch": 0.09968463964310438,
      "grad_norm": 5.4375,
      "learning_rate": 2.8504730405353436e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 569958.0,
      "step": 1296
    },
    {
      "epoch": 0.09976155680332283,
      "grad_norm": 4.6875,
      "learning_rate": 2.850357664795016e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 570280.0,
      "step": 1297
    },
    {
      "epoch": 0.09983847396354127,
      "grad_norm": 10.8125,
      "learning_rate": 2.8502422890546885e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 570566.0,
      "step": 1298
    },
    {
      "epoch": 0.09991539112375972,
      "grad_norm": 7.65625,
      "learning_rate": 2.8501269133143603e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 570932.0,
      "step": 1299
    },
    {
      "epoch": 0.09999230828397816,
      "grad_norm": 5.59375,
      "learning_rate": 2.8500115375740326e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 571266.0,
      "step": 1300
    },
    {
      "epoch": 0.1000692254441966,
      "grad_norm": 4.78125,
      "learning_rate": 2.8498961618337053e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 571782.0,
      "step": 1301
    },
    {
      "epoch": 0.10014614260441504,
      "grad_norm": 9.9375,
      "learning_rate": 2.8497807860933775e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 572096.0,
      "step": 1302
    },
    {
      "epoch": 0.10022305976463348,
      "grad_norm": 3.3125,
      "learning_rate": 2.84966541035305e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 572882.0,
      "step": 1303
    },
    {
      "epoch": 0.10029997692485193,
      "grad_norm": 11.125,
      "learning_rate": 2.849550034612722e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 573144.0,
      "step": 1304
    },
    {
      "epoch": 0.10037689408507038,
      "grad_norm": 3.390625,
      "learning_rate": 2.8494346588723943e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 573706.0,
      "step": 1305
    },
    {
      "epoch": 0.10045381124528882,
      "grad_norm": 3.40625,
      "learning_rate": 2.849319283132067e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 574292.0,
      "step": 1306
    },
    {
      "epoch": 0.10053072840550727,
      "grad_norm": 4.8125,
      "learning_rate": 2.8492039073917392e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 574620.0,
      "step": 1307
    },
    {
      "epoch": 0.10060764556572571,
      "grad_norm": 2.9375,
      "learning_rate": 2.8490885316514115e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 575416.0,
      "step": 1308
    },
    {
      "epoch": 0.10068456272594416,
      "grad_norm": 7.8125,
      "learning_rate": 2.848973155911084e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 575728.0,
      "step": 1309
    },
    {
      "epoch": 0.1007614798861626,
      "grad_norm": 3.890625,
      "learning_rate": 2.848857780170756e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 576284.0,
      "step": 1310
    },
    {
      "epoch": 0.10083839704638105,
      "grad_norm": 2.421875,
      "learning_rate": 2.8487424044304283e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 577110.0,
      "step": 1311
    },
    {
      "epoch": 0.1009153142065995,
      "grad_norm": 10.875,
      "learning_rate": 2.848627028690101e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 577412.0,
      "step": 1312
    },
    {
      "epoch": 0.10099223136681794,
      "grad_norm": 14.375,
      "learning_rate": 2.8485116529497732e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 577670.0,
      "step": 1313
    },
    {
      "epoch": 0.10106914852703638,
      "grad_norm": 7.625,
      "learning_rate": 2.8483962772094455e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 577968.0,
      "step": 1314
    },
    {
      "epoch": 0.10114606568725483,
      "grad_norm": 5.1875,
      "learning_rate": 2.8482809014691177e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 578318.0,
      "step": 1315
    },
    {
      "epoch": 0.10122298284747328,
      "grad_norm": 4.1875,
      "learning_rate": 2.84816552572879e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 578886.0,
      "step": 1316
    },
    {
      "epoch": 0.10129990000769172,
      "grad_norm": 5.15625,
      "learning_rate": 2.8480501499884627e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9893428087234497,
      "num_tokens": 579450.0,
      "step": 1317
    },
    {
      "epoch": 0.10137681716791017,
      "grad_norm": 7.53125,
      "learning_rate": 2.847934774248135e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 579756.0,
      "step": 1318
    },
    {
      "epoch": 0.10145373432812861,
      "grad_norm": 7.5625,
      "learning_rate": 2.847819398507807e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 580054.0,
      "step": 1319
    },
    {
      "epoch": 0.10153065148834706,
      "grad_norm": 2.65625,
      "learning_rate": 2.8477040227674794e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 580868.0,
      "step": 1320
    },
    {
      "epoch": 0.10160756864856549,
      "grad_norm": 3.0,
      "learning_rate": 2.8475886470271517e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 581448.0,
      "step": 1321
    },
    {
      "epoch": 0.10168448580878393,
      "grad_norm": 10.5,
      "learning_rate": 2.847473271286824e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 581800.0,
      "step": 1322
    },
    {
      "epoch": 0.10176140296900238,
      "grad_norm": 3.390625,
      "learning_rate": 2.8473578955464966e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 582324.0,
      "step": 1323
    },
    {
      "epoch": 0.10183832012922082,
      "grad_norm": 9.125,
      "learning_rate": 2.847242519806169e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 582632.0,
      "step": 1324
    },
    {
      "epoch": 0.10191523728943927,
      "grad_norm": 8.1875,
      "learning_rate": 2.847127144065841e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 582976.0,
      "step": 1325
    },
    {
      "epoch": 0.10199215444965772,
      "grad_norm": 3.65625,
      "learning_rate": 2.8470117683255134e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 583500.0,
      "step": 1326
    },
    {
      "epoch": 0.10206907160987616,
      "grad_norm": 3.171875,
      "learning_rate": 2.8468963925851857e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 584062.0,
      "step": 1327
    },
    {
      "epoch": 0.1021459887700946,
      "grad_norm": 2.984375,
      "learning_rate": 2.846781016844858e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 584608.0,
      "step": 1328
    },
    {
      "epoch": 0.10222290593031305,
      "grad_norm": 4.5,
      "learning_rate": 2.8466656411045306e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 585188.0,
      "step": 1329
    },
    {
      "epoch": 0.1022998230905315,
      "grad_norm": 2.828125,
      "learning_rate": 2.846550265364203e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 585970.0,
      "step": 1330
    },
    {
      "epoch": 0.10237674025074994,
      "grad_norm": 6.21875,
      "learning_rate": 2.846434889623875e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 586292.0,
      "step": 1331
    },
    {
      "epoch": 0.10245365741096839,
      "grad_norm": 8.125,
      "learning_rate": 2.8463195138835474e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 586628.0,
      "step": 1332
    },
    {
      "epoch": 0.10253057457118683,
      "grad_norm": 3.90625,
      "learning_rate": 2.8462041381432197e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.990328848361969,
      "num_tokens": 587146.0,
      "step": 1333
    },
    {
      "epoch": 0.10260749173140528,
      "grad_norm": 2.53125,
      "learning_rate": 2.8460887624028923e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 587702.0,
      "step": 1334
    },
    {
      "epoch": 0.10268440889162372,
      "grad_norm": 4.75,
      "learning_rate": 2.8459733866625646e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 588290.0,
      "step": 1335
    },
    {
      "epoch": 0.10276132605184217,
      "grad_norm": 5.65625,
      "learning_rate": 2.845858010922237e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 588634.0,
      "step": 1336
    },
    {
      "epoch": 0.10283824321206061,
      "grad_norm": 3.203125,
      "learning_rate": 2.845742635181909e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 589194.0,
      "step": 1337
    },
    {
      "epoch": 0.10291516037227906,
      "grad_norm": 9.6875,
      "learning_rate": 2.8456272594415814e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9937499761581421,
      "num_tokens": 589518.0,
      "step": 1338
    },
    {
      "epoch": 0.1029920775324975,
      "grad_norm": 10.3125,
      "learning_rate": 2.8455118837012536e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 589824.0,
      "step": 1339
    },
    {
      "epoch": 0.10306899469271595,
      "grad_norm": 3.84375,
      "learning_rate": 2.8453965079609263e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 590396.0,
      "step": 1340
    },
    {
      "epoch": 0.1031459118529344,
      "grad_norm": 8.375,
      "learning_rate": 2.8452811322205986e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 590676.0,
      "step": 1341
    },
    {
      "epoch": 0.10322282901315283,
      "grad_norm": 3.25,
      "learning_rate": 2.845165756480271e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 591458.0,
      "step": 1342
    },
    {
      "epoch": 0.10329974617337127,
      "grad_norm": 7.1875,
      "learning_rate": 2.845050380739943e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 591790.0,
      "step": 1343
    },
    {
      "epoch": 0.10337666333358972,
      "grad_norm": 4.15625,
      "learning_rate": 2.8449350049996153e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 592550.0,
      "step": 1344
    },
    {
      "epoch": 0.10345358049380816,
      "grad_norm": 3.15625,
      "learning_rate": 2.844819629259288e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 593100.0,
      "step": 1345
    },
    {
      "epoch": 0.10353049765402661,
      "grad_norm": 2.984375,
      "learning_rate": 2.8447042535189603e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 593898.0,
      "step": 1346
    },
    {
      "epoch": 0.10360741481424506,
      "grad_norm": 4.125,
      "learning_rate": 2.8445888777786325e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 594462.0,
      "step": 1347
    },
    {
      "epoch": 0.1036843319744635,
      "grad_norm": 11.875,
      "learning_rate": 2.844473502038305e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 594802.0,
      "step": 1348
    },
    {
      "epoch": 0.10376124913468195,
      "grad_norm": 3.828125,
      "learning_rate": 2.844358126297977e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 595326.0,
      "step": 1349
    },
    {
      "epoch": 0.10383816629490039,
      "grad_norm": 8.5625,
      "learning_rate": 2.8442427505576493e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 595630.0,
      "step": 1350
    },
    {
      "epoch": 0.10391508345511884,
      "grad_norm": 2.765625,
      "learning_rate": 2.844127374817322e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 596456.0,
      "step": 1351
    },
    {
      "epoch": 0.10399200061533728,
      "grad_norm": 7.21875,
      "learning_rate": 2.8440119990769942e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 596750.0,
      "step": 1352
    },
    {
      "epoch": 0.10406891777555573,
      "grad_norm": 6.375,
      "learning_rate": 2.8438966233366665e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 597048.0,
      "step": 1353
    },
    {
      "epoch": 0.10414583493577417,
      "grad_norm": 6.03125,
      "learning_rate": 2.8437812475963388e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 597360.0,
      "step": 1354
    },
    {
      "epoch": 0.10422275209599262,
      "grad_norm": 6.0625,
      "learning_rate": 2.843665871856011e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 597668.0,
      "step": 1355
    },
    {
      "epoch": 0.10429966925621106,
      "grad_norm": 2.828125,
      "learning_rate": 2.8435504961156837e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 598464.0,
      "step": 1356
    },
    {
      "epoch": 0.10437658641642951,
      "grad_norm": 7.21875,
      "learning_rate": 2.843435120375356e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 598812.0,
      "step": 1357
    },
    {
      "epoch": 0.10445350357664795,
      "grad_norm": 3.515625,
      "learning_rate": 2.8433197446350282e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 599356.0,
      "step": 1358
    },
    {
      "epoch": 0.1045304207368664,
      "grad_norm": 4.65625,
      "learning_rate": 2.8432043688947005e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 599924.0,
      "step": 1359
    },
    {
      "epoch": 0.10460733789708485,
      "grad_norm": 3.078125,
      "learning_rate": 2.8430889931543727e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 600472.0,
      "step": 1360
    },
    {
      "epoch": 0.10468425505730329,
      "grad_norm": 5.75,
      "learning_rate": 2.842973617414045e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 600798.0,
      "step": 1361
    },
    {
      "epoch": 0.10476117221752172,
      "grad_norm": 8.1875,
      "learning_rate": 2.8428582416737177e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 601132.0,
      "step": 1362
    },
    {
      "epoch": 0.10483808937774017,
      "grad_norm": 6.5625,
      "learning_rate": 2.84274286593339e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 601416.0,
      "step": 1363
    },
    {
      "epoch": 0.10491500653795861,
      "grad_norm": 4.1875,
      "learning_rate": 2.842627490193062e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 601942.0,
      "step": 1364
    },
    {
      "epoch": 0.10499192369817706,
      "grad_norm": 3.484375,
      "learning_rate": 2.8425121144527345e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 602486.0,
      "step": 1365
    },
    {
      "epoch": 0.1050688408583955,
      "grad_norm": 3.828125,
      "learning_rate": 2.8423967387124067e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 603074.0,
      "step": 1366
    },
    {
      "epoch": 0.10514575801861395,
      "grad_norm": 4.5,
      "learning_rate": 2.8422813629720794e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 603590.0,
      "step": 1367
    },
    {
      "epoch": 0.1052226751788324,
      "grad_norm": 7.75,
      "learning_rate": 2.8421659872317516e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 603920.0,
      "step": 1368
    },
    {
      "epoch": 0.10529959233905084,
      "grad_norm": 3.46875,
      "learning_rate": 2.8420506114914235e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 604454.0,
      "step": 1369
    },
    {
      "epoch": 0.10537650949926929,
      "grad_norm": 6.0,
      "learning_rate": 2.841935235751096e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 604748.0,
      "step": 1370
    },
    {
      "epoch": 0.10545342665948773,
      "grad_norm": 6.96875,
      "learning_rate": 2.8418198600107684e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 605088.0,
      "step": 1371
    },
    {
      "epoch": 0.10553034381970618,
      "grad_norm": 8.25,
      "learning_rate": 2.8417044842704407e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 605448.0,
      "step": 1372
    },
    {
      "epoch": 0.10560726097992462,
      "grad_norm": 4.34375,
      "learning_rate": 2.8415891085301134e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 605968.0,
      "step": 1373
    },
    {
      "epoch": 0.10568417814014307,
      "grad_norm": 5.71875,
      "learning_rate": 2.8414737327897856e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 606274.0,
      "step": 1374
    },
    {
      "epoch": 0.10576109530036151,
      "grad_norm": 9.5,
      "learning_rate": 2.841358357049458e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 606610.0,
      "step": 1375
    },
    {
      "epoch": 0.10583801246057996,
      "grad_norm": 8.0625,
      "learning_rate": 2.84124298130913e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 606934.0,
      "step": 1376
    },
    {
      "epoch": 0.1059149296207984,
      "grad_norm": 7.71875,
      "learning_rate": 2.8411276055688024e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 607240.0,
      "step": 1377
    },
    {
      "epoch": 0.10599184678101685,
      "grad_norm": 14.0,
      "learning_rate": 2.8410122298284747e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 607472.0,
      "step": 1378
    },
    {
      "epoch": 0.1060687639412353,
      "grad_norm": 4.03125,
      "learning_rate": 2.8408968540881473e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 608020.0,
      "step": 1379
    },
    {
      "epoch": 0.10614568110145374,
      "grad_norm": 3.25,
      "learning_rate": 2.840781478347819e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 608538.0,
      "step": 1380
    },
    {
      "epoch": 0.10622259826167219,
      "grad_norm": 7.75,
      "learning_rate": 2.840666102607492e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 608834.0,
      "step": 1381
    },
    {
      "epoch": 0.10629951542189062,
      "grad_norm": 7.21875,
      "learning_rate": 2.840550726867164e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 609158.0,
      "step": 1382
    },
    {
      "epoch": 0.10637643258210906,
      "grad_norm": 5.53125,
      "learning_rate": 2.8404353511268364e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 609464.0,
      "step": 1383
    },
    {
      "epoch": 0.10645334974232751,
      "grad_norm": 2.71875,
      "learning_rate": 2.840319975386509e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 609982.0,
      "step": 1384
    },
    {
      "epoch": 0.10653026690254595,
      "grad_norm": 3.609375,
      "learning_rate": 2.8402045996461813e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 610790.0,
      "step": 1385
    },
    {
      "epoch": 0.1066071840627644,
      "grad_norm": 4.3125,
      "learning_rate": 2.840089223905853e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 611358.0,
      "step": 1386
    },
    {
      "epoch": 0.10668410122298284,
      "grad_norm": 7.0,
      "learning_rate": 2.839973848165526e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9927007555961609,
      "num_tokens": 611634.0,
      "step": 1387
    },
    {
      "epoch": 0.10676101838320129,
      "grad_norm": 7.5,
      "learning_rate": 2.839858472425198e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 611944.0,
      "step": 1388
    },
    {
      "epoch": 0.10683793554341973,
      "grad_norm": 8.5,
      "learning_rate": 2.8397430966848703e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9927007555961609,
      "num_tokens": 612220.0,
      "step": 1389
    },
    {
      "epoch": 0.10691485270363818,
      "grad_norm": 6.09375,
      "learning_rate": 2.839627720944543e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 612520.0,
      "step": 1390
    },
    {
      "epoch": 0.10699176986385663,
      "grad_norm": 3.140625,
      "learning_rate": 2.839512345204215e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9936061501502991,
      "num_tokens": 613304.0,
      "step": 1391
    },
    {
      "epoch": 0.10706868702407507,
      "grad_norm": 6.1875,
      "learning_rate": 2.8393969694638875e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 613632.0,
      "step": 1392
    },
    {
      "epoch": 0.10714560418429352,
      "grad_norm": 7.3125,
      "learning_rate": 2.83928159372356e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 613962.0,
      "step": 1393
    },
    {
      "epoch": 0.10722252134451196,
      "grad_norm": 3.46875,
      "learning_rate": 2.839166217983232e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 614526.0,
      "step": 1394
    },
    {
      "epoch": 0.10729943850473041,
      "grad_norm": 3.1875,
      "learning_rate": 2.8390508422429047e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 615570.0,
      "step": 1395
    },
    {
      "epoch": 0.10737635566494885,
      "grad_norm": 7.53125,
      "learning_rate": 2.838935466502577e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 615876.0,
      "step": 1396
    },
    {
      "epoch": 0.1074532728251673,
      "grad_norm": 3.140625,
      "learning_rate": 2.838820090762249e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 616444.0,
      "step": 1397
    },
    {
      "epoch": 0.10753018998538574,
      "grad_norm": 8.75,
      "learning_rate": 2.8387047150219215e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 616736.0,
      "step": 1398
    },
    {
      "epoch": 0.10760710714560419,
      "grad_norm": 10.125,
      "learning_rate": 2.8385893392815938e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 617020.0,
      "step": 1399
    },
    {
      "epoch": 0.10768402430582263,
      "grad_norm": 6.8125,
      "learning_rate": 2.838473963541266e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 617338.0,
      "step": 1400
    },
    {
      "epoch": 0.10776094146604108,
      "grad_norm": 4.71875,
      "learning_rate": 2.8383585878009387e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 617666.0,
      "step": 1401
    },
    {
      "epoch": 0.10783785862625953,
      "grad_norm": 8.4375,
      "learning_rate": 2.8382432120606105e-06,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 617992.0,
      "step": 1402
    },
    {
      "epoch": 0.10791477578647796,
      "grad_norm": 5.4375,
      "learning_rate": 2.8381278363202832e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 618512.0,
      "step": 1403
    },
    {
      "epoch": 0.1079916929466964,
      "grad_norm": 9.5625,
      "learning_rate": 2.8380124605799555e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 618790.0,
      "step": 1404
    },
    {
      "epoch": 0.10806861010691485,
      "grad_norm": 5.0,
      "learning_rate": 2.8378970848396277e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 619314.0,
      "step": 1405
    },
    {
      "epoch": 0.10814552726713329,
      "grad_norm": 9.1875,
      "learning_rate": 2.8377817090993004e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 619630.0,
      "step": 1406
    },
    {
      "epoch": 0.10822244442735174,
      "grad_norm": 4.40625,
      "learning_rate": 2.8376663333589723e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 620200.0,
      "step": 1407
    },
    {
      "epoch": 0.10829936158757018,
      "grad_norm": 7.0625,
      "learning_rate": 2.8375509576186445e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9937499761581421,
      "num_tokens": 620524.0,
      "step": 1408
    },
    {
      "epoch": 0.10837627874778863,
      "grad_norm": 2.65625,
      "learning_rate": 2.837435581878317e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 621070.0,
      "step": 1409
    },
    {
      "epoch": 0.10845319590800707,
      "grad_norm": 5.90625,
      "learning_rate": 2.8373202061379895e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 621392.0,
      "step": 1410
    },
    {
      "epoch": 0.10853011306822552,
      "grad_norm": 1.9453125,
      "learning_rate": 2.8372048303976617e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 621960.0,
      "step": 1411
    },
    {
      "epoch": 0.10860703022844397,
      "grad_norm": 7.9375,
      "learning_rate": 2.8370894546573344e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 622266.0,
      "step": 1412
    },
    {
      "epoch": 0.10868394738866241,
      "grad_norm": 9.1875,
      "learning_rate": 2.8369740789170062e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 622546.0,
      "step": 1413
    },
    {
      "epoch": 0.10876086454888086,
      "grad_norm": 6.5,
      "learning_rate": 2.836858703176679e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 622870.0,
      "step": 1414
    },
    {
      "epoch": 0.1088377817090993,
      "grad_norm": 8.5625,
      "learning_rate": 2.836743327436351e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 623178.0,
      "step": 1415
    },
    {
      "epoch": 0.10891469886931775,
      "grad_norm": 6.0625,
      "learning_rate": 2.8366279516960234e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 623508.0,
      "step": 1416
    },
    {
      "epoch": 0.10899161602953619,
      "grad_norm": 4.28125,
      "learning_rate": 2.836512575955696e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 624076.0,
      "step": 1417
    },
    {
      "epoch": 0.10906853318975464,
      "grad_norm": 8.375,
      "learning_rate": 2.836397200215368e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 624354.0,
      "step": 1418
    },
    {
      "epoch": 0.10914545034997308,
      "grad_norm": 5.34375,
      "learning_rate": 2.83628182447504e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 624950.0,
      "step": 1419
    },
    {
      "epoch": 0.10922236751019153,
      "grad_norm": 3.625,
      "learning_rate": 2.836166448734713e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 625502.0,
      "step": 1420
    },
    {
      "epoch": 0.10929928467040997,
      "grad_norm": 6.0625,
      "learning_rate": 2.836051072994385e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 625786.0,
      "step": 1421
    },
    {
      "epoch": 0.10937620183062842,
      "grad_norm": 8.75,
      "learning_rate": 2.8359356972540574e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 626102.0,
      "step": 1422
    },
    {
      "epoch": 0.10945311899084685,
      "grad_norm": 3.78125,
      "learning_rate": 2.83582032151373e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 626658.0,
      "step": 1423
    },
    {
      "epoch": 0.1095300361510653,
      "grad_norm": 6.875,
      "learning_rate": 2.835704945773402e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 627006.0,
      "step": 1424
    },
    {
      "epoch": 0.10960695331128374,
      "grad_norm": 8.1875,
      "learning_rate": 2.8355895700330746e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 627342.0,
      "step": 1425
    },
    {
      "epoch": 0.10968387047150219,
      "grad_norm": 8.5,
      "learning_rate": 2.835474194292747e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 627642.0,
      "step": 1426
    },
    {
      "epoch": 0.10976078763172063,
      "grad_norm": 4.65625,
      "learning_rate": 2.835358818552419e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 628242.0,
      "step": 1427
    },
    {
      "epoch": 0.10983770479193908,
      "grad_norm": 4.125,
      "learning_rate": 2.8352434428120914e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 628792.0,
      "step": 1428
    },
    {
      "epoch": 0.10991462195215752,
      "grad_norm": 2.734375,
      "learning_rate": 2.8351280670717636e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 629352.0,
      "step": 1429
    },
    {
      "epoch": 0.10999153911237597,
      "grad_norm": 6.4375,
      "learning_rate": 2.835012691331436e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 629658.0,
      "step": 1430
    },
    {
      "epoch": 0.11006845627259441,
      "grad_norm": 5.8125,
      "learning_rate": 2.8348973155911086e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 630152.0,
      "step": 1431
    },
    {
      "epoch": 0.11014537343281286,
      "grad_norm": 3.609375,
      "learning_rate": 2.834781939850781e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 630700.0,
      "step": 1432
    },
    {
      "epoch": 0.1102222905930313,
      "grad_norm": 3.46875,
      "learning_rate": 2.834666564110453e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 631300.0,
      "step": 1433
    },
    {
      "epoch": 0.11029920775324975,
      "grad_norm": 5.71875,
      "learning_rate": 2.8345511883701258e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 631870.0,
      "step": 1434
    },
    {
      "epoch": 0.1103761249134682,
      "grad_norm": 5.0,
      "learning_rate": 2.8344358126297976e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 632364.0,
      "step": 1435
    },
    {
      "epoch": 0.11045304207368664,
      "grad_norm": 6.90625,
      "learning_rate": 2.83432043688947e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 632676.0,
      "step": 1436
    },
    {
      "epoch": 0.11052995923390509,
      "grad_norm": 2.890625,
      "learning_rate": 2.8342050611491425e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 633226.0,
      "step": 1437
    },
    {
      "epoch": 0.11060687639412353,
      "grad_norm": 7.84375,
      "learning_rate": 2.834089685408815e-06,
      "loss": 0.0361,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 633510.0,
      "step": 1438
    },
    {
      "epoch": 0.11068379355434198,
      "grad_norm": 7.5625,
      "learning_rate": 2.833974309668487e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 633802.0,
      "step": 1439
    },
    {
      "epoch": 0.11076071071456042,
      "grad_norm": 4.0625,
      "learning_rate": 2.8338589339281593e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 634372.0,
      "step": 1440
    },
    {
      "epoch": 0.11083762787477887,
      "grad_norm": 4.96875,
      "learning_rate": 2.8337435581878316e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 634968.0,
      "step": 1441
    },
    {
      "epoch": 0.11091454503499731,
      "grad_norm": 9.125,
      "learning_rate": 2.8336281824475043e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 635240.0,
      "step": 1442
    },
    {
      "epoch": 0.11099146219521576,
      "grad_norm": 8.625,
      "learning_rate": 2.8335128067071765e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 635572.0,
      "step": 1443
    },
    {
      "epoch": 0.11106837935543419,
      "grad_norm": 3.015625,
      "learning_rate": 2.8333974309668488e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 636118.0,
      "step": 1444
    },
    {
      "epoch": 0.11114529651565264,
      "grad_norm": 2.984375,
      "learning_rate": 2.833282055226521e-06,
      "loss": 0.0053,
      "mean_token_accuracy": 0.996874988079071,
      "num_tokens": 637402.0,
      "step": 1445
    },
    {
      "epoch": 0.11122221367587108,
      "grad_norm": 2.65625,
      "learning_rate": 2.8331666794861933e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9974874258041382,
      "num_tokens": 638200.0,
      "step": 1446
    },
    {
      "epoch": 0.11129913083608953,
      "grad_norm": 2.265625,
      "learning_rate": 2.8330513037458655e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 638734.0,
      "step": 1447
    },
    {
      "epoch": 0.11137604799630797,
      "grad_norm": 6.625,
      "learning_rate": 2.8329359280055382e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 639298.0,
      "step": 1448
    },
    {
      "epoch": 0.11145296515652642,
      "grad_norm": 7.84375,
      "learning_rate": 2.8328205522652105e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9837133288383484,
      "num_tokens": 639606.0,
      "step": 1449
    },
    {
      "epoch": 0.11152988231674486,
      "grad_norm": 11.0625,
      "learning_rate": 2.8327051765248827e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 639912.0,
      "step": 1450
    },
    {
      "epoch": 0.11160679947696331,
      "grad_norm": 7.59375,
      "learning_rate": 2.832589800784555e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 640224.0,
      "step": 1451
    },
    {
      "epoch": 0.11168371663718175,
      "grad_norm": 9.5,
      "learning_rate": 2.8324744250442273e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 640534.0,
      "step": 1452
    },
    {
      "epoch": 0.1117606337974002,
      "grad_norm": 6.21875,
      "learning_rate": 2.8323590493039e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 640858.0,
      "step": 1453
    },
    {
      "epoch": 0.11183755095761864,
      "grad_norm": 7.53125,
      "learning_rate": 2.832243673563572e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 641220.0,
      "step": 1454
    },
    {
      "epoch": 0.11191446811783709,
      "grad_norm": 7.4375,
      "learning_rate": 2.8321282978232445e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 641528.0,
      "step": 1455
    },
    {
      "epoch": 0.11199138527805554,
      "grad_norm": 7.75,
      "learning_rate": 2.8320129220829167e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 641806.0,
      "step": 1456
    },
    {
      "epoch": 0.11206830243827398,
      "grad_norm": 9.9375,
      "learning_rate": 2.831897546342589e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 642184.0,
      "step": 1457
    },
    {
      "epoch": 0.11214521959849243,
      "grad_norm": 6.21875,
      "learning_rate": 2.8317821706022612e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 642756.0,
      "step": 1458
    },
    {
      "epoch": 0.11222213675871087,
      "grad_norm": 10.0,
      "learning_rate": 2.831666794861934e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 643082.0,
      "step": 1459
    },
    {
      "epoch": 0.11229905391892932,
      "grad_norm": 6.25,
      "learning_rate": 2.831551419121606e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 643378.0,
      "step": 1460
    },
    {
      "epoch": 0.11237597107914776,
      "grad_norm": 6.875,
      "learning_rate": 2.8314360433812784e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 643656.0,
      "step": 1461
    },
    {
      "epoch": 0.11245288823936621,
      "grad_norm": 4.1875,
      "learning_rate": 2.8313206676409507e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 644202.0,
      "step": 1462
    },
    {
      "epoch": 0.11252980539958465,
      "grad_norm": 15.5625,
      "learning_rate": 2.831205291900623e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9773463010787964,
      "num_tokens": 644514.0,
      "step": 1463
    },
    {
      "epoch": 0.11260672255980309,
      "grad_norm": 7.96875,
      "learning_rate": 2.8310899161602956e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 644784.0,
      "step": 1464
    },
    {
      "epoch": 0.11268363972002153,
      "grad_norm": 7.6875,
      "learning_rate": 2.830974540419968e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 645116.0,
      "step": 1465
    },
    {
      "epoch": 0.11276055688023998,
      "grad_norm": 9.625,
      "learning_rate": 2.83085916467964e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 645424.0,
      "step": 1466
    },
    {
      "epoch": 0.11283747404045842,
      "grad_norm": 3.078125,
      "learning_rate": 2.8307437889393124e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 645982.0,
      "step": 1467
    },
    {
      "epoch": 0.11291439120067687,
      "grad_norm": 11.1875,
      "learning_rate": 2.8306284131989847e-06,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 646264.0,
      "step": 1468
    },
    {
      "epoch": 0.11299130836089531,
      "grad_norm": 8.0,
      "learning_rate": 2.830513037458657e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 646536.0,
      "step": 1469
    },
    {
      "epoch": 0.11306822552111376,
      "grad_norm": 5.0625,
      "learning_rate": 2.8303976617183296e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 647100.0,
      "step": 1470
    },
    {
      "epoch": 0.1131451426813322,
      "grad_norm": 7.9375,
      "learning_rate": 2.830282285978002e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 647384.0,
      "step": 1471
    },
    {
      "epoch": 0.11322205984155065,
      "grad_norm": 9.1875,
      "learning_rate": 2.830166910237674e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 647660.0,
      "step": 1472
    },
    {
      "epoch": 0.1132989770017691,
      "grad_norm": 3.984375,
      "learning_rate": 2.8300515344973464e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 647978.0,
      "step": 1473
    },
    {
      "epoch": 0.11337589416198754,
      "grad_norm": 7.46875,
      "learning_rate": 2.8299361587570186e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 648278.0,
      "step": 1474
    },
    {
      "epoch": 0.11345281132220598,
      "grad_norm": 5.90625,
      "learning_rate": 2.8298207830166913e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 648582.0,
      "step": 1475
    },
    {
      "epoch": 0.11352972848242443,
      "grad_norm": 5.9375,
      "learning_rate": 2.8297054072763636e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 649156.0,
      "step": 1476
    },
    {
      "epoch": 0.11360664564264288,
      "grad_norm": 6.46875,
      "learning_rate": 2.829590031536036e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 649476.0,
      "step": 1477
    },
    {
      "epoch": 0.11368356280286132,
      "grad_norm": 11.875,
      "learning_rate": 2.829474655795708e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 649806.0,
      "step": 1478
    },
    {
      "epoch": 0.11376047996307977,
      "grad_norm": 7.96875,
      "learning_rate": 2.8293592800553803e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 650062.0,
      "step": 1479
    },
    {
      "epoch": 0.11383739712329821,
      "grad_norm": 3.734375,
      "learning_rate": 2.8292439043150526e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 650622.0,
      "step": 1480
    },
    {
      "epoch": 0.11391431428351666,
      "grad_norm": 4.03125,
      "learning_rate": 2.8291285285747253e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 651432.0,
      "step": 1481
    },
    {
      "epoch": 0.1139912314437351,
      "grad_norm": 7.09375,
      "learning_rate": 2.8290131528343975e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 651758.0,
      "step": 1482
    },
    {
      "epoch": 0.11406814860395355,
      "grad_norm": 9.0,
      "learning_rate": 2.82889777709407e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 652096.0,
      "step": 1483
    },
    {
      "epoch": 0.114145065764172,
      "grad_norm": 7.28125,
      "learning_rate": 2.828782401353742e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 652406.0,
      "step": 1484
    },
    {
      "epoch": 0.11422198292439042,
      "grad_norm": 4.0,
      "learning_rate": 2.8286670256134143e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 652980.0,
      "step": 1485
    },
    {
      "epoch": 0.11429890008460887,
      "grad_norm": 8.5625,
      "learning_rate": 2.828551649873087e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 653332.0,
      "step": 1486
    },
    {
      "epoch": 0.11437581724482732,
      "grad_norm": 7.21875,
      "learning_rate": 2.8284362741327593e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 653698.0,
      "step": 1487
    },
    {
      "epoch": 0.11445273440504576,
      "grad_norm": 12.625,
      "learning_rate": 2.8283208983924315e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 653954.0,
      "step": 1488
    },
    {
      "epoch": 0.1145296515652642,
      "grad_norm": 3.84375,
      "learning_rate": 2.8282055226521038e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 654278.0,
      "step": 1489
    },
    {
      "epoch": 0.11460656872548265,
      "grad_norm": 5.375,
      "learning_rate": 2.828090146911776e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 654814.0,
      "step": 1490
    },
    {
      "epoch": 0.1146834858857011,
      "grad_norm": 4.1875,
      "learning_rate": 2.8279747711714483e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 655398.0,
      "step": 1491
    },
    {
      "epoch": 0.11476040304591954,
      "grad_norm": 12.125,
      "learning_rate": 2.827859395431121e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 655698.0,
      "step": 1492
    },
    {
      "epoch": 0.11483732020613799,
      "grad_norm": 6.65625,
      "learning_rate": 2.8277440196907932e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 655996.0,
      "step": 1493
    },
    {
      "epoch": 0.11491423736635643,
      "grad_norm": 10.875,
      "learning_rate": 2.8276286439504655e-06,
      "loss": 0.0398,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 656274.0,
      "step": 1494
    },
    {
      "epoch": 0.11499115452657488,
      "grad_norm": 4.9375,
      "learning_rate": 2.8275132682101377e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 656816.0,
      "step": 1495
    },
    {
      "epoch": 0.11506807168679332,
      "grad_norm": 4.03125,
      "learning_rate": 2.82739789246981e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 657374.0,
      "step": 1496
    },
    {
      "epoch": 0.11514498884701177,
      "grad_norm": 2.703125,
      "learning_rate": 2.8272825167294823e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 658184.0,
      "step": 1497
    },
    {
      "epoch": 0.11522190600723022,
      "grad_norm": 7.78125,
      "learning_rate": 2.827167140989155e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 658488.0,
      "step": 1498
    },
    {
      "epoch": 0.11529882316744866,
      "grad_norm": 5.03125,
      "learning_rate": 2.827051765248827e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 658800.0,
      "step": 1499
    },
    {
      "epoch": 0.1153757403276671,
      "grad_norm": 5.09375,
      "learning_rate": 2.8269363895084995e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 659108.0,
      "step": 1500
    },
    {
      "epoch": 0.11545265748788555,
      "grad_norm": 9.1875,
      "learning_rate": 2.8268210137681717e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 659426.0,
      "step": 1501
    },
    {
      "epoch": 0.115529574648104,
      "grad_norm": 5.53125,
      "learning_rate": 2.826705638027844e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 659704.0,
      "step": 1502
    },
    {
      "epoch": 0.11560649180832244,
      "grad_norm": 9.8125,
      "learning_rate": 2.8265902622875167e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 660014.0,
      "step": 1503
    },
    {
      "epoch": 0.11568340896854089,
      "grad_norm": 5.90625,
      "learning_rate": 2.826474886547189e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 660296.0,
      "step": 1504
    },
    {
      "epoch": 0.11576032612875932,
      "grad_norm": 3.609375,
      "learning_rate": 2.8263595108068607e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 660842.0,
      "step": 1505
    },
    {
      "epoch": 0.11583724328897776,
      "grad_norm": 3.71875,
      "learning_rate": 2.8262441350665334e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 661388.0,
      "step": 1506
    },
    {
      "epoch": 0.11591416044919621,
      "grad_norm": 3.796875,
      "learning_rate": 2.8261287593262057e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 661910.0,
      "step": 1507
    },
    {
      "epoch": 0.11599107760941466,
      "grad_norm": 4.53125,
      "learning_rate": 2.826013383585878e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 662496.0,
      "step": 1508
    },
    {
      "epoch": 0.1160679947696331,
      "grad_norm": 5.0,
      "learning_rate": 2.8258980078455506e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 663068.0,
      "step": 1509
    },
    {
      "epoch": 0.11614491192985155,
      "grad_norm": 10.9375,
      "learning_rate": 2.8257826321052225e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 663432.0,
      "step": 1510
    },
    {
      "epoch": 0.11622182909006999,
      "grad_norm": 2.359375,
      "learning_rate": 2.825667256364895e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 663978.0,
      "step": 1511
    },
    {
      "epoch": 0.11629874625028844,
      "grad_norm": 6.53125,
      "learning_rate": 2.8255518806245674e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 664300.0,
      "step": 1512
    },
    {
      "epoch": 0.11637566341050688,
      "grad_norm": 3.3125,
      "learning_rate": 2.8254365048842397e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 664856.0,
      "step": 1513
    },
    {
      "epoch": 0.11645258057072533,
      "grad_norm": 8.125,
      "learning_rate": 2.8253211291439123e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 665174.0,
      "step": 1514
    },
    {
      "epoch": 0.11652949773094377,
      "grad_norm": 3.265625,
      "learning_rate": 2.8252057534035846e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 665742.0,
      "step": 1515
    },
    {
      "epoch": 0.11660641489116222,
      "grad_norm": 4.59375,
      "learning_rate": 2.8250903776632564e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 666284.0,
      "step": 1516
    },
    {
      "epoch": 0.11668333205138066,
      "grad_norm": 4.71875,
      "learning_rate": 2.824975001922929e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 666778.0,
      "step": 1517
    },
    {
      "epoch": 0.11676024921159911,
      "grad_norm": 6.96875,
      "learning_rate": 2.8248596261826014e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 667096.0,
      "step": 1518
    },
    {
      "epoch": 0.11683716637181756,
      "grad_norm": 4.8125,
      "learning_rate": 2.8247442504422736e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 667666.0,
      "step": 1519
    },
    {
      "epoch": 0.116914083532036,
      "grad_norm": 2.671875,
      "learning_rate": 2.8246288747019463e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 668474.0,
      "step": 1520
    },
    {
      "epoch": 0.11699100069225445,
      "grad_norm": 6.28125,
      "learning_rate": 2.824513498961618e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 668802.0,
      "step": 1521
    },
    {
      "epoch": 0.11706791785247289,
      "grad_norm": 8.875,
      "learning_rate": 2.824398123221291e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 669082.0,
      "step": 1522
    },
    {
      "epoch": 0.11714483501269134,
      "grad_norm": 6.1875,
      "learning_rate": 2.824282747480963e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 669376.0,
      "step": 1523
    },
    {
      "epoch": 0.11722175217290978,
      "grad_norm": 9.3125,
      "learning_rate": 2.8241673717406353e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 669694.0,
      "step": 1524
    },
    {
      "epoch": 0.11729866933312823,
      "grad_norm": 6.78125,
      "learning_rate": 2.824051996000308e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 669976.0,
      "step": 1525
    },
    {
      "epoch": 0.11737558649334666,
      "grad_norm": 4.4375,
      "learning_rate": 2.8239366202599803e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 670544.0,
      "step": 1526
    },
    {
      "epoch": 0.1174525036535651,
      "grad_norm": 5.75,
      "learning_rate": 2.823821244519652e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 670866.0,
      "step": 1527
    },
    {
      "epoch": 0.11752942081378355,
      "grad_norm": 4.3125,
      "learning_rate": 2.823705868779325e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 671382.0,
      "step": 1528
    },
    {
      "epoch": 0.117606337974002,
      "grad_norm": 5.15625,
      "learning_rate": 2.823590493038997e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 671694.0,
      "step": 1529
    },
    {
      "epoch": 0.11768325513422044,
      "grad_norm": 6.96875,
      "learning_rate": 2.8234751172986693e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 672004.0,
      "step": 1530
    },
    {
      "epoch": 0.11776017229443889,
      "grad_norm": 2.828125,
      "learning_rate": 2.823359741558342e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 672824.0,
      "step": 1531
    },
    {
      "epoch": 0.11783708945465733,
      "grad_norm": 3.265625,
      "learning_rate": 2.823244365818014e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 673364.0,
      "step": 1532
    },
    {
      "epoch": 0.11791400661487578,
      "grad_norm": 5.1875,
      "learning_rate": 2.8231289900776865e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 673676.0,
      "step": 1533
    },
    {
      "epoch": 0.11799092377509422,
      "grad_norm": 5.21875,
      "learning_rate": 2.8230136143373588e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 674248.0,
      "step": 1534
    },
    {
      "epoch": 0.11806784093531267,
      "grad_norm": 8.9375,
      "learning_rate": 2.822898238597031e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 674588.0,
      "step": 1535
    },
    {
      "epoch": 0.11814475809553111,
      "grad_norm": 6.125,
      "learning_rate": 2.8227828628567037e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 674866.0,
      "step": 1536
    },
    {
      "epoch": 0.11822167525574956,
      "grad_norm": 3.25,
      "learning_rate": 2.822667487116376e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 675406.0,
      "step": 1537
    },
    {
      "epoch": 0.118298592415968,
      "grad_norm": 2.53125,
      "learning_rate": 2.822552111376048e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 676192.0,
      "step": 1538
    },
    {
      "epoch": 0.11837550957618645,
      "grad_norm": 4.125,
      "learning_rate": 2.8224367356357205e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 676686.0,
      "step": 1539
    },
    {
      "epoch": 0.1184524267364049,
      "grad_norm": 8.875,
      "learning_rate": 2.8223213598953927e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 676964.0,
      "step": 1540
    },
    {
      "epoch": 0.11852934389662334,
      "grad_norm": 8.25,
      "learning_rate": 2.822205984155065e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 677326.0,
      "step": 1541
    },
    {
      "epoch": 0.11860626105684179,
      "grad_norm": 3.9375,
      "learning_rate": 2.8220906084147377e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 677916.0,
      "step": 1542
    },
    {
      "epoch": 0.11868317821706023,
      "grad_norm": 8.0625,
      "learning_rate": 2.8219752326744095e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 678276.0,
      "step": 1543
    },
    {
      "epoch": 0.11876009537727868,
      "grad_norm": 7.6875,
      "learning_rate": 2.821859856934082e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.992094874382019,
      "num_tokens": 678530.0,
      "step": 1544
    },
    {
      "epoch": 0.11883701253749712,
      "grad_norm": 4.3125,
      "learning_rate": 2.8217444811937545e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 678824.0,
      "step": 1545
    },
    {
      "epoch": 0.11891392969771555,
      "grad_norm": 9.625,
      "learning_rate": 2.8216291054534267e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9841772317886353,
      "num_tokens": 679142.0,
      "step": 1546
    },
    {
      "epoch": 0.118990846857934,
      "grad_norm": 2.984375,
      "learning_rate": 2.821513729713099e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 680192.0,
      "step": 1547
    },
    {
      "epoch": 0.11906776401815244,
      "grad_norm": 5.28125,
      "learning_rate": 2.8213983539727712e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.993127167224884,
      "num_tokens": 680484.0,
      "step": 1548
    },
    {
      "epoch": 0.11914468117837089,
      "grad_norm": 3.484375,
      "learning_rate": 2.8212829782324435e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 681052.0,
      "step": 1549
    },
    {
      "epoch": 0.11922159833858934,
      "grad_norm": 5.40625,
      "learning_rate": 2.821167602492116e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 681358.0,
      "step": 1550
    },
    {
      "epoch": 0.11929851549880778,
      "grad_norm": 2.671875,
      "learning_rate": 2.8210522267517884e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 681920.0,
      "step": 1551
    },
    {
      "epoch": 0.11937543265902623,
      "grad_norm": 8.0625,
      "learning_rate": 2.8209368510114607e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 682258.0,
      "step": 1552
    },
    {
      "epoch": 0.11945234981924467,
      "grad_norm": 2.96875,
      "learning_rate": 2.8208214752711334e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 682802.0,
      "step": 1553
    },
    {
      "epoch": 0.11952926697946312,
      "grad_norm": 7.46875,
      "learning_rate": 2.820706099530805e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 683088.0,
      "step": 1554
    },
    {
      "epoch": 0.11960618413968156,
      "grad_norm": 5.75,
      "learning_rate": 2.8205907237904775e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 683392.0,
      "step": 1555
    },
    {
      "epoch": 0.11968310129990001,
      "grad_norm": 3.828125,
      "learning_rate": 2.82047534805015e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 683956.0,
      "step": 1556
    },
    {
      "epoch": 0.11976001846011845,
      "grad_norm": 3.625,
      "learning_rate": 2.8203599723098224e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 684530.0,
      "step": 1557
    },
    {
      "epoch": 0.1198369356203369,
      "grad_norm": 8.4375,
      "learning_rate": 2.8202445965694947e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 684848.0,
      "step": 1558
    },
    {
      "epoch": 0.11991385278055534,
      "grad_norm": 4.3125,
      "learning_rate": 2.820129220829167e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 685170.0,
      "step": 1559
    },
    {
      "epoch": 0.11999076994077379,
      "grad_norm": 7.3125,
      "learning_rate": 2.820013845088839e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 685668.0,
      "step": 1560
    },
    {
      "epoch": 0.12006768710099223,
      "grad_norm": 3.859375,
      "learning_rate": 2.819898469348512e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 686256.0,
      "step": 1561
    },
    {
      "epoch": 0.12014460426121068,
      "grad_norm": 6.09375,
      "learning_rate": 2.819783093608184e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 686542.0,
      "step": 1562
    },
    {
      "epoch": 0.12022152142142913,
      "grad_norm": 10.5625,
      "learning_rate": 2.8196677178678564e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 686818.0,
      "step": 1563
    },
    {
      "epoch": 0.12029843858164757,
      "grad_norm": 4.65625,
      "learning_rate": 2.819552342127529e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 687420.0,
      "step": 1564
    },
    {
      "epoch": 0.12037535574186602,
      "grad_norm": 8.3125,
      "learning_rate": 2.819436966387201e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 687752.0,
      "step": 1565
    },
    {
      "epoch": 0.12045227290208446,
      "grad_norm": 8.5625,
      "learning_rate": 2.819321590646873e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 688062.0,
      "step": 1566
    },
    {
      "epoch": 0.1205291900623029,
      "grad_norm": 7.84375,
      "learning_rate": 2.819206214906546e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 688390.0,
      "step": 1567
    },
    {
      "epoch": 0.12060610722252134,
      "grad_norm": 2.671875,
      "learning_rate": 2.819090839166218e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 689434.0,
      "step": 1568
    },
    {
      "epoch": 0.12068302438273978,
      "grad_norm": 5.625,
      "learning_rate": 2.8189754634258904e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 689954.0,
      "step": 1569
    },
    {
      "epoch": 0.12075994154295823,
      "grad_norm": 5.8125,
      "learning_rate": 2.8188600876855626e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 690272.0,
      "step": 1570
    },
    {
      "epoch": 0.12083685870317667,
      "grad_norm": 5.875,
      "learning_rate": 2.818744711945235e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 690606.0,
      "step": 1571
    },
    {
      "epoch": 0.12091377586339512,
      "grad_norm": 3.703125,
      "learning_rate": 2.8186293362049075e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 691180.0,
      "step": 1572
    },
    {
      "epoch": 0.12099069302361357,
      "grad_norm": 3.203125,
      "learning_rate": 2.81851396046458e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 691748.0,
      "step": 1573
    },
    {
      "epoch": 0.12106761018383201,
      "grad_norm": 9.5625,
      "learning_rate": 2.818398584724252e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 692034.0,
      "step": 1574
    },
    {
      "epoch": 0.12114452734405046,
      "grad_norm": 4.78125,
      "learning_rate": 2.8182832089839243e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 692354.0,
      "step": 1575
    },
    {
      "epoch": 0.1212214445042689,
      "grad_norm": 10.9375,
      "learning_rate": 2.8181678332435966e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 692634.0,
      "step": 1576
    },
    {
      "epoch": 0.12129836166448735,
      "grad_norm": 7.28125,
      "learning_rate": 2.818052457503269e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 692934.0,
      "step": 1577
    },
    {
      "epoch": 0.12137527882470579,
      "grad_norm": 5.1875,
      "learning_rate": 2.8179370817629415e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 693256.0,
      "step": 1578
    },
    {
      "epoch": 0.12145219598492424,
      "grad_norm": 2.40625,
      "learning_rate": 2.8178217060226138e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 693814.0,
      "step": 1579
    },
    {
      "epoch": 0.12152911314514268,
      "grad_norm": 8.375,
      "learning_rate": 2.817706330282286e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 1.0,
      "num_tokens": 694094.0,
      "step": 1580
    },
    {
      "epoch": 0.12160603030536113,
      "grad_norm": 6.46875,
      "learning_rate": 2.8175909545419583e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 694450.0,
      "step": 1581
    },
    {
      "epoch": 0.12168294746557957,
      "grad_norm": 8.1875,
      "learning_rate": 2.8174755788016306e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 694748.0,
      "step": 1582
    },
    {
      "epoch": 0.12175986462579802,
      "grad_norm": 7.1875,
      "learning_rate": 2.8173602030613032e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 695080.0,
      "step": 1583
    },
    {
      "epoch": 0.12183678178601647,
      "grad_norm": 6.46875,
      "learning_rate": 2.8172448273209755e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 695364.0,
      "step": 1584
    },
    {
      "epoch": 0.12191369894623491,
      "grad_norm": 9.5,
      "learning_rate": 2.8171294515806478e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 695704.0,
      "step": 1585
    },
    {
      "epoch": 0.12199061610645336,
      "grad_norm": 2.1875,
      "learning_rate": 2.81701407584032e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 696262.0,
      "step": 1586
    },
    {
      "epoch": 0.12206753326667179,
      "grad_norm": 3.203125,
      "learning_rate": 2.8168987000999923e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 696780.0,
      "step": 1587
    },
    {
      "epoch": 0.12214445042689023,
      "grad_norm": 7.46875,
      "learning_rate": 2.8167833243596645e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 697090.0,
      "step": 1588
    },
    {
      "epoch": 0.12222136758710868,
      "grad_norm": 4.5625,
      "learning_rate": 2.816667948619337e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 697664.0,
      "step": 1589
    },
    {
      "epoch": 0.12229828474732712,
      "grad_norm": 5.84375,
      "learning_rate": 2.8165525728790095e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 697982.0,
      "step": 1590
    },
    {
      "epoch": 0.12237520190754557,
      "grad_norm": 3.4375,
      "learning_rate": 2.8164371971386817e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 698556.0,
      "step": 1591
    },
    {
      "epoch": 0.12245211906776401,
      "grad_norm": 5.03125,
      "learning_rate": 2.816321821398354e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 698876.0,
      "step": 1592
    },
    {
      "epoch": 0.12252903622798246,
      "grad_norm": 6.9375,
      "learning_rate": 2.8162064456580262e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 699160.0,
      "step": 1593
    },
    {
      "epoch": 0.1226059533882009,
      "grad_norm": 8.5625,
      "learning_rate": 2.816091069917699e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9824561476707458,
      "num_tokens": 699446.0,
      "step": 1594
    },
    {
      "epoch": 0.12268287054841935,
      "grad_norm": 6.0625,
      "learning_rate": 2.815975694177371e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 699776.0,
      "step": 1595
    },
    {
      "epoch": 0.1227597877086378,
      "grad_norm": 5.03125,
      "learning_rate": 2.8158603184370434e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 700070.0,
      "step": 1596
    },
    {
      "epoch": 0.12283670486885624,
      "grad_norm": 4.65625,
      "learning_rate": 2.8157449426967157e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 700416.0,
      "step": 1597
    },
    {
      "epoch": 0.12291362202907469,
      "grad_norm": 4.21875,
      "learning_rate": 2.815629566956388e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 700724.0,
      "step": 1598
    },
    {
      "epoch": 0.12299053918929313,
      "grad_norm": 8.875,
      "learning_rate": 2.81551419121606e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 701054.0,
      "step": 1599
    },
    {
      "epoch": 0.12306745634951158,
      "grad_norm": 6.375,
      "learning_rate": 2.815398815475733e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 701382.0,
      "step": 1600
    },
    {
      "epoch": 0.12314437350973002,
      "grad_norm": 7.3125,
      "learning_rate": 2.815283439735405e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 701660.0,
      "step": 1601
    },
    {
      "epoch": 0.12322129066994847,
      "grad_norm": 6.71875,
      "learning_rate": 2.8151680639950774e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 701978.0,
      "step": 1602
    },
    {
      "epoch": 0.12329820783016691,
      "grad_norm": 12.0,
      "learning_rate": 2.8150526882547497e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 702328.0,
      "step": 1603
    },
    {
      "epoch": 0.12337512499038536,
      "grad_norm": 6.15625,
      "learning_rate": 2.814937312514422e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 702640.0,
      "step": 1604
    },
    {
      "epoch": 0.1234520421506038,
      "grad_norm": 3.734375,
      "learning_rate": 2.814821936774094e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 703438.0,
      "step": 1605
    },
    {
      "epoch": 0.12352895931082225,
      "grad_norm": 6.03125,
      "learning_rate": 2.814706561033767e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 703768.0,
      "step": 1606
    },
    {
      "epoch": 0.1236058764710407,
      "grad_norm": 3.3125,
      "learning_rate": 2.814591185293439e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 704310.0,
      "step": 1607
    },
    {
      "epoch": 0.12368279363125913,
      "grad_norm": 5.625,
      "learning_rate": 2.8144758095531114e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 704852.0,
      "step": 1608
    },
    {
      "epoch": 0.12375971079147757,
      "grad_norm": 6.875,
      "learning_rate": 2.8143604338127836e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 705208.0,
      "step": 1609
    },
    {
      "epoch": 0.12383662795169602,
      "grad_norm": 5.40625,
      "learning_rate": 2.814245058072456e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 705780.0,
      "step": 1610
    },
    {
      "epoch": 0.12391354511191446,
      "grad_norm": 10.5625,
      "learning_rate": 2.8141296823321286e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 706052.0,
      "step": 1611
    },
    {
      "epoch": 0.12399046227213291,
      "grad_norm": 5.03125,
      "learning_rate": 2.814014306591801e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 706652.0,
      "step": 1612
    },
    {
      "epoch": 0.12406737943235135,
      "grad_norm": 2.3125,
      "learning_rate": 2.8138989308514727e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 707238.0,
      "step": 1613
    },
    {
      "epoch": 0.1241442965925698,
      "grad_norm": 4.84375,
      "learning_rate": 2.8137835551111454e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 707578.0,
      "step": 1614
    },
    {
      "epoch": 0.12422121375278825,
      "grad_norm": 10.125,
      "learning_rate": 2.8136681793708176e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 707884.0,
      "step": 1615
    },
    {
      "epoch": 0.12429813091300669,
      "grad_norm": 8.0,
      "learning_rate": 2.81355280363049e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 708186.0,
      "step": 1616
    },
    {
      "epoch": 0.12437504807322514,
      "grad_norm": 5.40625,
      "learning_rate": 2.8134374278901626e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 708510.0,
      "step": 1617
    },
    {
      "epoch": 0.12445196523344358,
      "grad_norm": 4.90625,
      "learning_rate": 2.813322052149835e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 708826.0,
      "step": 1618
    },
    {
      "epoch": 0.12452888239366203,
      "grad_norm": 6.0,
      "learning_rate": 2.813206676409507e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 709162.0,
      "step": 1619
    },
    {
      "epoch": 0.12460579955388047,
      "grad_norm": 5.4375,
      "learning_rate": 2.8130913006691793e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 709528.0,
      "step": 1620
    },
    {
      "epoch": 0.12468271671409892,
      "grad_norm": 10.9375,
      "learning_rate": 2.8129759249288516e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 709818.0,
      "step": 1621
    },
    {
      "epoch": 0.12475963387431736,
      "grad_norm": 3.859375,
      "learning_rate": 2.8128605491885243e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 710174.0,
      "step": 1622
    },
    {
      "epoch": 0.12483655103453581,
      "grad_norm": 12.9375,
      "learning_rate": 2.8127451734481965e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 710454.0,
      "step": 1623
    },
    {
      "epoch": 0.12491346819475425,
      "grad_norm": 5.40625,
      "learning_rate": 2.8126297977078684e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 710764.0,
      "step": 1624
    },
    {
      "epoch": 0.1249903853549727,
      "grad_norm": 5.90625,
      "learning_rate": 2.812514421967541e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 711284.0,
      "step": 1625
    },
    {
      "epoch": 0.12506730251519113,
      "grad_norm": 4.625,
      "learning_rate": 2.8123990462272133e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 711802.0,
      "step": 1626
    },
    {
      "epoch": 0.1251442196754096,
      "grad_norm": 4.875,
      "learning_rate": 2.8122836704868856e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 712148.0,
      "step": 1627
    },
    {
      "epoch": 0.12522113683562802,
      "grad_norm": 9.8125,
      "learning_rate": 2.8121682947465582e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 712438.0,
      "step": 1628
    },
    {
      "epoch": 0.12529805399584648,
      "grad_norm": 15.3125,
      "learning_rate": 2.8120529190062305e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.9825327396392822,
      "num_tokens": 712668.0,
      "step": 1629
    },
    {
      "epoch": 0.1253749711560649,
      "grad_norm": 6.40625,
      "learning_rate": 2.8119375432659028e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 712974.0,
      "step": 1630
    },
    {
      "epoch": 0.12545188831628337,
      "grad_norm": 3.359375,
      "learning_rate": 2.811822167525575e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 713520.0,
      "step": 1631
    },
    {
      "epoch": 0.1255288054765018,
      "grad_norm": 8.75,
      "learning_rate": 2.8117067917852473e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9862637519836426,
      "num_tokens": 713886.0,
      "step": 1632
    },
    {
      "epoch": 0.12560572263672026,
      "grad_norm": 5.75,
      "learning_rate": 2.81159141604492e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9897785186767578,
      "num_tokens": 714474.0,
      "step": 1633
    },
    {
      "epoch": 0.1256826397969387,
      "grad_norm": 3.515625,
      "learning_rate": 2.811476040304592e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 715024.0,
      "step": 1634
    },
    {
      "epoch": 0.12575955695715715,
      "grad_norm": 10.3125,
      "learning_rate": 2.811360664564264e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9961089491844177,
      "num_tokens": 715282.0,
      "step": 1635
    },
    {
      "epoch": 0.12583647411737559,
      "grad_norm": 3.234375,
      "learning_rate": 2.8112452888239367e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9973614811897278,
      "num_tokens": 716042.0,
      "step": 1636
    },
    {
      "epoch": 0.12591339127759404,
      "grad_norm": 3.09375,
      "learning_rate": 2.811129913083609e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9908925294876099,
      "num_tokens": 716592.0,
      "step": 1637
    },
    {
      "epoch": 0.12599030843781248,
      "grad_norm": 3.828125,
      "learning_rate": 2.8110145373432812e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 717150.0,
      "step": 1638
    },
    {
      "epoch": 0.1260672255980309,
      "grad_norm": 2.875,
      "learning_rate": 2.810899161602954e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9952152967453003,
      "num_tokens": 718198.0,
      "step": 1639
    },
    {
      "epoch": 0.12614414275824937,
      "grad_norm": 4.09375,
      "learning_rate": 2.810783785862626e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9913644194602966,
      "num_tokens": 718778.0,
      "step": 1640
    },
    {
      "epoch": 0.1262210599184678,
      "grad_norm": 10.25,
      "learning_rate": 2.8106684101222984e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 719108.0,
      "step": 1641
    },
    {
      "epoch": 0.12629797707868626,
      "grad_norm": 8.0,
      "learning_rate": 2.8105530343819707e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 719448.0,
      "step": 1642
    },
    {
      "epoch": 0.1263748942389047,
      "grad_norm": 3.890625,
      "learning_rate": 2.810437658641643e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9976019263267517,
      "num_tokens": 720284.0,
      "step": 1643
    },
    {
      "epoch": 0.12645181139912315,
      "grad_norm": 4.71875,
      "learning_rate": 2.8103222829013156e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 720884.0,
      "step": 1644
    },
    {
      "epoch": 0.12652872855934158,
      "grad_norm": 8.1875,
      "learning_rate": 2.810206907160988e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 721180.0,
      "step": 1645
    },
    {
      "epoch": 0.12660564571956004,
      "grad_norm": 3.453125,
      "learning_rate": 2.8100915314206597e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 721698.0,
      "step": 1646
    },
    {
      "epoch": 0.12668256287977847,
      "grad_norm": 7.78125,
      "learning_rate": 2.8099761556803324e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 721996.0,
      "step": 1647
    },
    {
      "epoch": 0.12675948003999693,
      "grad_norm": 3.4375,
      "learning_rate": 2.8098607799400047e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 722564.0,
      "step": 1648
    },
    {
      "epoch": 0.12683639720021536,
      "grad_norm": 6.1875,
      "learning_rate": 2.809745404199677e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 722858.0,
      "step": 1649
    },
    {
      "epoch": 0.12691331436043382,
      "grad_norm": 7.59375,
      "learning_rate": 2.8096300284593496e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 723142.0,
      "step": 1650
    },
    {
      "epoch": 0.12699023152065225,
      "grad_norm": 5.9375,
      "learning_rate": 2.8095146527190214e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 723470.0,
      "step": 1651
    },
    {
      "epoch": 0.1270671486808707,
      "grad_norm": 8.5,
      "learning_rate": 2.809399276978694e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 723784.0,
      "step": 1652
    },
    {
      "epoch": 0.12714406584108914,
      "grad_norm": 6.15625,
      "learning_rate": 2.8092839012383664e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 724078.0,
      "step": 1653
    },
    {
      "epoch": 0.1272209830013076,
      "grad_norm": 8.5625,
      "learning_rate": 2.8091685254980386e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 724388.0,
      "step": 1654
    },
    {
      "epoch": 0.12729790016152603,
      "grad_norm": 3.46875,
      "learning_rate": 2.809053149757711e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 724956.0,
      "step": 1655
    },
    {
      "epoch": 0.1273748173217445,
      "grad_norm": 14.125,
      "learning_rate": 2.8089377740173836e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 725190.0,
      "step": 1656
    },
    {
      "epoch": 0.12745173448196292,
      "grad_norm": 7.375,
      "learning_rate": 2.8088223982770554e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 725518.0,
      "step": 1657
    },
    {
      "epoch": 0.12752865164218138,
      "grad_norm": 2.59375,
      "learning_rate": 2.808707022536728e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 726064.0,
      "step": 1658
    },
    {
      "epoch": 0.12760556880239982,
      "grad_norm": 3.09375,
      "learning_rate": 2.8085916467964004e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 726632.0,
      "step": 1659
    },
    {
      "epoch": 0.12768248596261825,
      "grad_norm": 2.546875,
      "learning_rate": 2.8084762710560726e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 727442.0,
      "step": 1660
    },
    {
      "epoch": 0.1277594031228367,
      "grad_norm": 11.4375,
      "learning_rate": 2.8083608953157453e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 727744.0,
      "step": 1661
    },
    {
      "epoch": 0.12783632028305514,
      "grad_norm": 7.46875,
      "learning_rate": 2.808245519575417e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 728020.0,
      "step": 1662
    },
    {
      "epoch": 0.1279132374432736,
      "grad_norm": 6.21875,
      "learning_rate": 2.8081301438350894e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 728344.0,
      "step": 1663
    },
    {
      "epoch": 0.12799015460349203,
      "grad_norm": 6.09375,
      "learning_rate": 2.808014768094762e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 728672.0,
      "step": 1664
    },
    {
      "epoch": 0.1280670717637105,
      "grad_norm": 6.21875,
      "learning_rate": 2.8078993923544343e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 728996.0,
      "step": 1665
    },
    {
      "epoch": 0.12814398892392892,
      "grad_norm": 7.4375,
      "learning_rate": 2.8077840166141066e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 729348.0,
      "step": 1666
    },
    {
      "epoch": 0.12822090608414738,
      "grad_norm": 8.3125,
      "learning_rate": 2.8076686408737793e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 729650.0,
      "step": 1667
    },
    {
      "epoch": 0.1282978232443658,
      "grad_norm": 3.359375,
      "learning_rate": 2.807553265133451e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 730234.0,
      "step": 1668
    },
    {
      "epoch": 0.12837474040458427,
      "grad_norm": 3.171875,
      "learning_rate": 2.8074378893931238e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 730798.0,
      "step": 1669
    },
    {
      "epoch": 0.1284516575648027,
      "grad_norm": 5.53125,
      "learning_rate": 2.807322513652796e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 731150.0,
      "step": 1670
    },
    {
      "epoch": 0.12852857472502116,
      "grad_norm": 4.5625,
      "learning_rate": 2.8072071379124683e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 731472.0,
      "step": 1671
    },
    {
      "epoch": 0.1286054918852396,
      "grad_norm": 8.5625,
      "learning_rate": 2.807091762172141e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 731790.0,
      "step": 1672
    },
    {
      "epoch": 0.12868240904545805,
      "grad_norm": 3.21875,
      "learning_rate": 2.806976386431813e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 732352.0,
      "step": 1673
    },
    {
      "epoch": 0.12875932620567648,
      "grad_norm": 7.15625,
      "learning_rate": 2.806861010691485e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 732646.0,
      "step": 1674
    },
    {
      "epoch": 0.12883624336589494,
      "grad_norm": 11.3125,
      "learning_rate": 2.8067456349511578e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 732950.0,
      "step": 1675
    },
    {
      "epoch": 0.12891316052611337,
      "grad_norm": 12.625,
      "learning_rate": 2.80663025921083e-06,
      "loss": 0.0409,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 733252.0,
      "step": 1676
    },
    {
      "epoch": 0.12899007768633183,
      "grad_norm": 5.1875,
      "learning_rate": 2.8065148834705023e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 733796.0,
      "step": 1677
    },
    {
      "epoch": 0.12906699484655026,
      "grad_norm": 5.59375,
      "learning_rate": 2.8063995077301745e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 734116.0,
      "step": 1678
    },
    {
      "epoch": 0.1291439120067687,
      "grad_norm": 5.46875,
      "learning_rate": 2.8062841319898468e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 734462.0,
      "step": 1679
    },
    {
      "epoch": 0.12922082916698716,
      "grad_norm": 5.28125,
      "learning_rate": 2.8061687562495195e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 734784.0,
      "step": 1680
    },
    {
      "epoch": 0.1292977463272056,
      "grad_norm": 3.015625,
      "learning_rate": 2.8060533805091917e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 735566.0,
      "step": 1681
    },
    {
      "epoch": 0.12937466348742405,
      "grad_norm": 3.8125,
      "learning_rate": 2.805938004768864e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 736130.0,
      "step": 1682
    },
    {
      "epoch": 0.12945158064764248,
      "grad_norm": 8.5625,
      "learning_rate": 2.8058226290285367e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 736412.0,
      "step": 1683
    },
    {
      "epoch": 0.12952849780786094,
      "grad_norm": 9.5,
      "learning_rate": 2.8057072532882085e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 736702.0,
      "step": 1684
    },
    {
      "epoch": 0.12960541496807937,
      "grad_norm": 5.8125,
      "learning_rate": 2.8055918775478808e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9875665903091431,
      "num_tokens": 737266.0,
      "step": 1685
    },
    {
      "epoch": 0.12968233212829783,
      "grad_norm": 2.953125,
      "learning_rate": 2.8054765018075534e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9938119053840637,
      "num_tokens": 738076.0,
      "step": 1686
    },
    {
      "epoch": 0.12975924928851626,
      "grad_norm": 5.90625,
      "learning_rate": 2.8053611260672257e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 738388.0,
      "step": 1687
    },
    {
      "epoch": 0.12983616644873472,
      "grad_norm": 6.90625,
      "learning_rate": 2.805245750326898e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 738692.0,
      "step": 1688
    },
    {
      "epoch": 0.12991308360895315,
      "grad_norm": 6.9375,
      "learning_rate": 2.8051303745865702e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 739006.0,
      "step": 1689
    },
    {
      "epoch": 0.1299900007691716,
      "grad_norm": 4.8125,
      "learning_rate": 2.8050149988462425e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 739284.0,
      "step": 1690
    },
    {
      "epoch": 0.13006691792939004,
      "grad_norm": 5.0,
      "learning_rate": 2.804899623105915e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 739880.0,
      "step": 1691
    },
    {
      "epoch": 0.1301438350896085,
      "grad_norm": 8.0,
      "learning_rate": 2.8047842473655874e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 740216.0,
      "step": 1692
    },
    {
      "epoch": 0.13022075224982693,
      "grad_norm": 8.875,
      "learning_rate": 2.8046688716252597e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 740552.0,
      "step": 1693
    },
    {
      "epoch": 0.1302976694100454,
      "grad_norm": 6.0,
      "learning_rate": 2.8045534958849324e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 740838.0,
      "step": 1694
    },
    {
      "epoch": 0.13037458657026382,
      "grad_norm": 2.203125,
      "learning_rate": 2.804438120144604e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 741388.0,
      "step": 1695
    },
    {
      "epoch": 0.13045150373048228,
      "grad_norm": 3.734375,
      "learning_rate": 2.8043227444042764e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 741976.0,
      "step": 1696
    },
    {
      "epoch": 0.1305284208907007,
      "grad_norm": 5.375,
      "learning_rate": 2.804207368663949e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 742284.0,
      "step": 1697
    },
    {
      "epoch": 0.13060533805091917,
      "grad_norm": 4.875,
      "learning_rate": 2.8040919929236214e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 742826.0,
      "step": 1698
    },
    {
      "epoch": 0.1306822552111376,
      "grad_norm": 3.140625,
      "learning_rate": 2.8039766171832936e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 743410.0,
      "step": 1699
    },
    {
      "epoch": 0.13075917237135604,
      "grad_norm": 4.90625,
      "learning_rate": 2.803861241442966e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 743716.0,
      "step": 1700
    },
    {
      "epoch": 0.1308360895315745,
      "grad_norm": 7.0625,
      "learning_rate": 2.803745865702638e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 744050.0,
      "step": 1701
    },
    {
      "epoch": 0.13091300669179293,
      "grad_norm": 2.984375,
      "learning_rate": 2.803630489962311e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 744610.0,
      "step": 1702
    },
    {
      "epoch": 0.1309899238520114,
      "grad_norm": 3.546875,
      "learning_rate": 2.803515114221983e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 745172.0,
      "step": 1703
    },
    {
      "epoch": 0.13106684101222982,
      "grad_norm": 2.0,
      "learning_rate": 2.8033997384816554e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 745720.0,
      "step": 1704
    },
    {
      "epoch": 0.13114375817244828,
      "grad_norm": 6.03125,
      "learning_rate": 2.8032843627413276e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 746002.0,
      "step": 1705
    },
    {
      "epoch": 0.1312206753326667,
      "grad_norm": 3.578125,
      "learning_rate": 2.803168987001e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 746564.0,
      "step": 1706
    },
    {
      "epoch": 0.13129759249288517,
      "grad_norm": 4.9375,
      "learning_rate": 2.803053611260672e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 746884.0,
      "step": 1707
    },
    {
      "epoch": 0.1313745096531036,
      "grad_norm": 8.4375,
      "learning_rate": 2.802938235520345e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 747202.0,
      "step": 1708
    },
    {
      "epoch": 0.13145142681332206,
      "grad_norm": 6.53125,
      "learning_rate": 2.802822859780017e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 747526.0,
      "step": 1709
    },
    {
      "epoch": 0.1315283439735405,
      "grad_norm": 4.875,
      "learning_rate": 2.8027074840396893e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 748074.0,
      "step": 1710
    },
    {
      "epoch": 0.13160526113375895,
      "grad_norm": 11.625,
      "learning_rate": 2.8025921082993616e-06,
      "loss": 0.0454,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 748330.0,
      "step": 1711
    },
    {
      "epoch": 0.13168217829397738,
      "grad_norm": 11.3125,
      "learning_rate": 2.802476732559034e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 748648.0,
      "step": 1712
    },
    {
      "epoch": 0.13175909545419584,
      "grad_norm": 4.1875,
      "learning_rate": 2.802361356818706e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 749234.0,
      "step": 1713
    },
    {
      "epoch": 0.13183601261441427,
      "grad_norm": 3.421875,
      "learning_rate": 2.8022459810783788e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 749814.0,
      "step": 1714
    },
    {
      "epoch": 0.13191292977463273,
      "grad_norm": 2.8125,
      "learning_rate": 2.802130605338051e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 750378.0,
      "step": 1715
    },
    {
      "epoch": 0.13198984693485116,
      "grad_norm": 3.03125,
      "learning_rate": 2.8020152295977233e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 751160.0,
      "step": 1716
    },
    {
      "epoch": 0.13206676409506962,
      "grad_norm": 3.484375,
      "learning_rate": 2.8018998538573956e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 751684.0,
      "step": 1717
    },
    {
      "epoch": 0.13214368125528805,
      "grad_norm": 7.0,
      "learning_rate": 2.801784478117068e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 752036.0,
      "step": 1718
    },
    {
      "epoch": 0.1322205984155065,
      "grad_norm": 2.6875,
      "learning_rate": 2.8016691023767405e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 752584.0,
      "step": 1719
    },
    {
      "epoch": 0.13229751557572494,
      "grad_norm": 2.203125,
      "learning_rate": 2.8015537266364128e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 753146.0,
      "step": 1720
    },
    {
      "epoch": 0.13237443273594338,
      "grad_norm": 3.96875,
      "learning_rate": 2.801438350896085e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 753718.0,
      "step": 1721
    },
    {
      "epoch": 0.13245134989616184,
      "grad_norm": 5.3125,
      "learning_rate": 2.8013229751557573e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 754282.0,
      "step": 1722
    },
    {
      "epoch": 0.13252826705638027,
      "grad_norm": 11.1875,
      "learning_rate": 2.8012075994154295e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 754610.0,
      "step": 1723
    },
    {
      "epoch": 0.13260518421659873,
      "grad_norm": 5.25,
      "learning_rate": 2.801092223675102e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 755174.0,
      "step": 1724
    },
    {
      "epoch": 0.13268210137681716,
      "grad_norm": 8.75,
      "learning_rate": 2.8009768479347745e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 755436.0,
      "step": 1725
    },
    {
      "epoch": 0.13275901853703562,
      "grad_norm": 6.21875,
      "learning_rate": 2.8008614721944467e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 755734.0,
      "step": 1726
    },
    {
      "epoch": 0.13283593569725405,
      "grad_norm": 9.0,
      "learning_rate": 2.800746096454119e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 756004.0,
      "step": 1727
    },
    {
      "epoch": 0.1329128528574725,
      "grad_norm": 8.1875,
      "learning_rate": 2.8006307207137912e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 756342.0,
      "step": 1728
    },
    {
      "epoch": 0.13298977001769094,
      "grad_norm": 4.96875,
      "learning_rate": 2.8005153449734635e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 756906.0,
      "step": 1729
    },
    {
      "epoch": 0.1330666871779094,
      "grad_norm": 2.953125,
      "learning_rate": 2.800399969233136e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 757452.0,
      "step": 1730
    },
    {
      "epoch": 0.13314360433812783,
      "grad_norm": 2.796875,
      "learning_rate": 2.8002845934928084e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 757970.0,
      "step": 1731
    },
    {
      "epoch": 0.1332205214983463,
      "grad_norm": 7.09375,
      "learning_rate": 2.8001692177524807e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9926199316978455,
      "num_tokens": 758242.0,
      "step": 1732
    },
    {
      "epoch": 0.13329743865856472,
      "grad_norm": 3.09375,
      "learning_rate": 2.800053842012153e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 758998.0,
      "step": 1733
    },
    {
      "epoch": 0.13337435581878318,
      "grad_norm": 4.375,
      "learning_rate": 2.7999384662718252e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 759550.0,
      "step": 1734
    },
    {
      "epoch": 0.1334512729790016,
      "grad_norm": 9.4375,
      "learning_rate": 2.7998230905314975e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 759862.0,
      "step": 1735
    },
    {
      "epoch": 0.13352819013922007,
      "grad_norm": 3.109375,
      "learning_rate": 2.79970771479117e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 760682.0,
      "step": 1736
    },
    {
      "epoch": 0.1336051072994385,
      "grad_norm": 6.4375,
      "learning_rate": 2.7995923390508424e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 760994.0,
      "step": 1737
    },
    {
      "epoch": 0.13368202445965696,
      "grad_norm": 3.59375,
      "learning_rate": 2.7994769633105147e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 761556.0,
      "step": 1738
    },
    {
      "epoch": 0.1337589416198754,
      "grad_norm": 2.4375,
      "learning_rate": 2.799361587570187e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 762102.0,
      "step": 1739
    },
    {
      "epoch": 0.13383585878009385,
      "grad_norm": 3.328125,
      "learning_rate": 2.799246211829859e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 762670.0,
      "step": 1740
    },
    {
      "epoch": 0.13391277594031228,
      "grad_norm": 2.703125,
      "learning_rate": 2.799130836089532e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 763480.0,
      "step": 1741
    },
    {
      "epoch": 0.13398969310053072,
      "grad_norm": 7.96875,
      "learning_rate": 2.799015460349204e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 763808.0,
      "step": 1742
    },
    {
      "epoch": 0.13406661026074917,
      "grad_norm": 5.4375,
      "learning_rate": 2.7989000846088764e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 764154.0,
      "step": 1743
    },
    {
      "epoch": 0.1341435274209676,
      "grad_norm": 2.96875,
      "learning_rate": 2.7987847088685486e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 764738.0,
      "step": 1744
    },
    {
      "epoch": 0.13422044458118607,
      "grad_norm": 2.5,
      "learning_rate": 2.798669333128221e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 765296.0,
      "step": 1745
    },
    {
      "epoch": 0.1342973617414045,
      "grad_norm": 2.46875,
      "learning_rate": 2.798553957387893e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 765836.0,
      "step": 1746
    },
    {
      "epoch": 0.13437427890162296,
      "grad_norm": 3.328125,
      "learning_rate": 2.798438581647566e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 766380.0,
      "step": 1747
    },
    {
      "epoch": 0.1344511960618414,
      "grad_norm": 1.7734375,
      "learning_rate": 2.798323205907238e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 767188.0,
      "step": 1748
    },
    {
      "epoch": 0.13452811322205985,
      "grad_norm": 8.25,
      "learning_rate": 2.7982078301669104e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 767498.0,
      "step": 1749
    },
    {
      "epoch": 0.13460503038227828,
      "grad_norm": 6.78125,
      "learning_rate": 2.7980924544265826e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 767854.0,
      "step": 1750
    },
    {
      "epoch": 0.13468194754249674,
      "grad_norm": 6.46875,
      "learning_rate": 2.797977078686255e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 768214.0,
      "step": 1751
    },
    {
      "epoch": 0.13475886470271517,
      "grad_norm": 5.78125,
      "learning_rate": 2.7978617029459276e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 768536.0,
      "step": 1752
    },
    {
      "epoch": 0.13483578186293363,
      "grad_norm": 8.4375,
      "learning_rate": 2.7977463272056e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 768834.0,
      "step": 1753
    },
    {
      "epoch": 0.13491269902315206,
      "grad_norm": 5.59375,
      "learning_rate": 2.7976309514652717e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 769190.0,
      "step": 1754
    },
    {
      "epoch": 0.13498961618337052,
      "grad_norm": 3.765625,
      "learning_rate": 2.7975155757249443e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 769768.0,
      "step": 1755
    },
    {
      "epoch": 0.13506653334358895,
      "grad_norm": 4.84375,
      "learning_rate": 2.7974001999846166e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 770292.0,
      "step": 1756
    },
    {
      "epoch": 0.1351434505038074,
      "grad_norm": 8.4375,
      "learning_rate": 2.797284824244289e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 770574.0,
      "step": 1757
    },
    {
      "epoch": 0.13522036766402584,
      "grad_norm": 3.859375,
      "learning_rate": 2.7971694485039615e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 770920.0,
      "step": 1758
    },
    {
      "epoch": 0.1352972848242443,
      "grad_norm": 6.25,
      "learning_rate": 2.797054072763634e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 771258.0,
      "step": 1759
    },
    {
      "epoch": 0.13537420198446273,
      "grad_norm": 10.4375,
      "learning_rate": 2.796938697023306e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.996632993221283,
      "num_tokens": 771558.0,
      "step": 1760
    },
    {
      "epoch": 0.13545111914468116,
      "grad_norm": 8.625,
      "learning_rate": 2.7968233212829783e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 771854.0,
      "step": 1761
    },
    {
      "epoch": 0.13552803630489962,
      "grad_norm": 12.9375,
      "learning_rate": 2.7967079455426506e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 772112.0,
      "step": 1762
    },
    {
      "epoch": 0.13560495346511806,
      "grad_norm": 608.0,
      "learning_rate": 2.796592569802323e-06,
      "loss": 0.0691,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 772422.0,
      "step": 1763
    },
    {
      "epoch": 0.13568187062533651,
      "grad_norm": 5.71875,
      "learning_rate": 2.7964771940619955e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 772722.0,
      "step": 1764
    },
    {
      "epoch": 0.13575878778555495,
      "grad_norm": 3.375,
      "learning_rate": 2.7963618183216673e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 773280.0,
      "step": 1765
    },
    {
      "epoch": 0.1358357049457734,
      "grad_norm": 3.40625,
      "learning_rate": 2.79624644258134e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 773824.0,
      "step": 1766
    },
    {
      "epoch": 0.13591262210599184,
      "grad_norm": 3.53125,
      "learning_rate": 2.7961310668410123e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 774368.0,
      "step": 1767
    },
    {
      "epoch": 0.1359895392662103,
      "grad_norm": 5.375,
      "learning_rate": 2.7960156911006845e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 774680.0,
      "step": 1768
    },
    {
      "epoch": 0.13606645642642873,
      "grad_norm": 5.0,
      "learning_rate": 2.7959003153603572e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 774964.0,
      "step": 1769
    },
    {
      "epoch": 0.1361433735866472,
      "grad_norm": 6.84375,
      "learning_rate": 2.7957849396200295e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 775294.0,
      "step": 1770
    },
    {
      "epoch": 0.13622029074686562,
      "grad_norm": 4.0,
      "learning_rate": 2.7956695638797013e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 775842.0,
      "step": 1771
    },
    {
      "epoch": 0.13629720790708408,
      "grad_norm": 4.15625,
      "learning_rate": 2.795554188139374e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 776434.0,
      "step": 1772
    },
    {
      "epoch": 0.1363741250673025,
      "grad_norm": 6.75,
      "learning_rate": 2.7954388123990462e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.993127167224884,
      "num_tokens": 776726.0,
      "step": 1773
    },
    {
      "epoch": 0.13645104222752097,
      "grad_norm": 6.21875,
      "learning_rate": 2.7953234366587185e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 777054.0,
      "step": 1774
    },
    {
      "epoch": 0.1365279593877394,
      "grad_norm": 10.125,
      "learning_rate": 2.795208060918391e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 777318.0,
      "step": 1775
    },
    {
      "epoch": 0.13660487654795786,
      "grad_norm": 7.21875,
      "learning_rate": 2.795092685178063e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 777598.0,
      "step": 1776
    },
    {
      "epoch": 0.1366817937081763,
      "grad_norm": 56.75,
      "learning_rate": 2.7949773094377357e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 777920.0,
      "step": 1777
    },
    {
      "epoch": 0.13675871086839475,
      "grad_norm": 8.625,
      "learning_rate": 2.794861933697408e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 778264.0,
      "step": 1778
    },
    {
      "epoch": 0.13683562802861318,
      "grad_norm": 3.359375,
      "learning_rate": 2.7947465579570802e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 778810.0,
      "step": 1779
    },
    {
      "epoch": 0.13691254518883164,
      "grad_norm": 7.90625,
      "learning_rate": 2.794631182216753e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9825072884559631,
      "num_tokens": 779154.0,
      "step": 1780
    },
    {
      "epoch": 0.13698946234905007,
      "grad_norm": 2.796875,
      "learning_rate": 2.7945158064764247e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9908925294876099,
      "num_tokens": 779704.0,
      "step": 1781
    },
    {
      "epoch": 0.1370663795092685,
      "grad_norm": 3.875,
      "learning_rate": 2.794400430736097e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 780260.0,
      "step": 1782
    },
    {
      "epoch": 0.13714329666948696,
      "grad_norm": 10.1875,
      "learning_rate": 2.7942850549957697e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.989276111125946,
      "num_tokens": 780636.0,
      "step": 1783
    },
    {
      "epoch": 0.1372202138297054,
      "grad_norm": 4.625,
      "learning_rate": 2.794169679255442e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 780920.0,
      "step": 1784
    },
    {
      "epoch": 0.13729713098992385,
      "grad_norm": 5.0,
      "learning_rate": 2.794054303515114e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 781214.0,
      "step": 1785
    },
    {
      "epoch": 0.13737404815014229,
      "grad_norm": 7.875,
      "learning_rate": 2.793938927774787e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 781518.0,
      "step": 1786
    },
    {
      "epoch": 0.13745096531036075,
      "grad_norm": 8.9375,
      "learning_rate": 2.7938235520344587e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 781850.0,
      "step": 1787
    },
    {
      "epoch": 0.13752788247057918,
      "grad_norm": 2.421875,
      "learning_rate": 2.7937081762941314e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 782396.0,
      "step": 1788
    },
    {
      "epoch": 0.13760479963079764,
      "grad_norm": 4.9375,
      "learning_rate": 2.7935928005538036e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 782722.0,
      "step": 1789
    },
    {
      "epoch": 0.13768171679101607,
      "grad_norm": 4.625,
      "learning_rate": 2.793477424813476e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 783242.0,
      "step": 1790
    },
    {
      "epoch": 0.13775863395123453,
      "grad_norm": 3.578125,
      "learning_rate": 2.7933620490731486e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 783564.0,
      "step": 1791
    },
    {
      "epoch": 0.13783555111145296,
      "grad_norm": 3.625,
      "learning_rate": 2.7932466733328204e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 784150.0,
      "step": 1792
    },
    {
      "epoch": 0.13791246827167142,
      "grad_norm": 7.84375,
      "learning_rate": 2.7931312975924927e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 784466.0,
      "step": 1793
    },
    {
      "epoch": 0.13798938543188985,
      "grad_norm": 3.671875,
      "learning_rate": 2.7930159218521654e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 785010.0,
      "step": 1794
    },
    {
      "epoch": 0.1380663025921083,
      "grad_norm": 6.625,
      "learning_rate": 2.7929005461118376e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 785348.0,
      "step": 1795
    },
    {
      "epoch": 0.13814321975232674,
      "grad_norm": 2.640625,
      "learning_rate": 2.79278517037151e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 786172.0,
      "step": 1796
    },
    {
      "epoch": 0.1382201369125452,
      "grad_norm": 6.15625,
      "learning_rate": 2.7926697946311826e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 786464.0,
      "step": 1797
    },
    {
      "epoch": 0.13829705407276363,
      "grad_norm": 7.40625,
      "learning_rate": 2.7925544188908544e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 786768.0,
      "step": 1798
    },
    {
      "epoch": 0.1383739712329821,
      "grad_norm": 3.796875,
      "learning_rate": 2.792439043150527e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 787338.0,
      "step": 1799
    },
    {
      "epoch": 0.13845088839320052,
      "grad_norm": 3.5,
      "learning_rate": 2.7923236674101993e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 787658.0,
      "step": 1800
    },
    {
      "epoch": 0.13852780555341898,
      "grad_norm": 8.8125,
      "learning_rate": 2.7922082916698716e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 787948.0,
      "step": 1801
    },
    {
      "epoch": 0.1386047227136374,
      "grad_norm": 7.09375,
      "learning_rate": 2.7920929159295443e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 788446.0,
      "step": 1802
    },
    {
      "epoch": 0.13868163987385584,
      "grad_norm": 7.8125,
      "learning_rate": 2.791977540189216e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 788742.0,
      "step": 1803
    },
    {
      "epoch": 0.1387585570340743,
      "grad_norm": 4.78125,
      "learning_rate": 2.7918621644488884e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 789052.0,
      "step": 1804
    },
    {
      "epoch": 0.13883547419429274,
      "grad_norm": 3.390625,
      "learning_rate": 2.791746788708561e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 789626.0,
      "step": 1805
    },
    {
      "epoch": 0.1389123913545112,
      "grad_norm": 4.5,
      "learning_rate": 2.7916314129682333e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 789924.0,
      "step": 1806
    },
    {
      "epoch": 0.13898930851472963,
      "grad_norm": 3.421875,
      "learning_rate": 2.7915160372279056e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 790472.0,
      "step": 1807
    },
    {
      "epoch": 0.13906622567494809,
      "grad_norm": 6.65625,
      "learning_rate": 2.7914006614875782e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 790838.0,
      "step": 1808
    },
    {
      "epoch": 0.13914314283516652,
      "grad_norm": 4.15625,
      "learning_rate": 2.79128528574725e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 791148.0,
      "step": 1809
    },
    {
      "epoch": 0.13922005999538498,
      "grad_norm": 2.578125,
      "learning_rate": 2.7911699100069228e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 791670.0,
      "step": 1810
    },
    {
      "epoch": 0.1392969771556034,
      "grad_norm": 8.8125,
      "learning_rate": 2.791054534266595e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 791926.0,
      "step": 1811
    },
    {
      "epoch": 0.13937389431582187,
      "grad_norm": 5.46875,
      "learning_rate": 2.7909391585262673e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 792208.0,
      "step": 1812
    },
    {
      "epoch": 0.1394508114760403,
      "grad_norm": 7.0,
      "learning_rate": 2.7908237827859395e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 792488.0,
      "step": 1813
    },
    {
      "epoch": 0.13952772863625876,
      "grad_norm": 2.484375,
      "learning_rate": 2.790708407045612e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 793302.0,
      "step": 1814
    },
    {
      "epoch": 0.1396046457964772,
      "grad_norm": 7.53125,
      "learning_rate": 2.790593031305284e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 793640.0,
      "step": 1815
    },
    {
      "epoch": 0.13968156295669565,
      "grad_norm": 4.78125,
      "learning_rate": 2.7904776555649567e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 793952.0,
      "step": 1816
    },
    {
      "epoch": 0.13975848011691408,
      "grad_norm": 6.84375,
      "learning_rate": 2.790362279824629e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 794234.0,
      "step": 1817
    },
    {
      "epoch": 0.13983539727713254,
      "grad_norm": 4.9375,
      "learning_rate": 2.7902469040843013e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 794560.0,
      "step": 1818
    },
    {
      "epoch": 0.13991231443735097,
      "grad_norm": 7.25,
      "learning_rate": 2.7901315283439735e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 794900.0,
      "step": 1819
    },
    {
      "epoch": 0.13998923159756943,
      "grad_norm": 2.125,
      "learning_rate": 2.7900161526036458e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 795450.0,
      "step": 1820
    },
    {
      "epoch": 0.14006614875778786,
      "grad_norm": 3.078125,
      "learning_rate": 2.789900776863318e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 796274.0,
      "step": 1821
    },
    {
      "epoch": 0.14014306591800632,
      "grad_norm": 8.25,
      "learning_rate": 2.7897854011229907e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 796602.0,
      "step": 1822
    },
    {
      "epoch": 0.14021998307822475,
      "grad_norm": 3.65625,
      "learning_rate": 2.789670025382663e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 797144.0,
      "step": 1823
    },
    {
      "epoch": 0.14029690023844318,
      "grad_norm": 8.1875,
      "learning_rate": 2.7895546496423352e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 797484.0,
      "step": 1824
    },
    {
      "epoch": 0.14037381739866164,
      "grad_norm": 3.609375,
      "learning_rate": 2.7894392739020075e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 798038.0,
      "step": 1825
    },
    {
      "epoch": 0.14045073455888007,
      "grad_norm": 8.875,
      "learning_rate": 2.7893238981616797e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 798292.0,
      "step": 1826
    },
    {
      "epoch": 0.14052765171909853,
      "grad_norm": 5.0,
      "learning_rate": 2.7892085224213524e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 798864.0,
      "step": 1827
    },
    {
      "epoch": 0.14060456887931697,
      "grad_norm": 5.4375,
      "learning_rate": 2.7890931466810247e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 799192.0,
      "step": 1828
    },
    {
      "epoch": 0.14068148603953542,
      "grad_norm": 5.3125,
      "learning_rate": 2.788977770940697e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 799512.0,
      "step": 1829
    },
    {
      "epoch": 0.14075840319975386,
      "grad_norm": 4.15625,
      "learning_rate": 2.788862395200369e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 800030.0,
      "step": 1830
    },
    {
      "epoch": 0.14083532035997232,
      "grad_norm": 2.703125,
      "learning_rate": 2.7887470194600415e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 800850.0,
      "step": 1831
    },
    {
      "epoch": 0.14091223752019075,
      "grad_norm": 6.53125,
      "learning_rate": 2.7886316437197137e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 801156.0,
      "step": 1832
    },
    {
      "epoch": 0.1409891546804092,
      "grad_norm": 2.359375,
      "learning_rate": 2.7885162679793864e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 801976.0,
      "step": 1833
    },
    {
      "epoch": 0.14106607184062764,
      "grad_norm": 10.5625,
      "learning_rate": 2.7884008922390587e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 802294.0,
      "step": 1834
    },
    {
      "epoch": 0.1411429890008461,
      "grad_norm": 10.0,
      "learning_rate": 2.788285516498731e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 802596.0,
      "step": 1835
    },
    {
      "epoch": 0.14121990616106453,
      "grad_norm": 7.25,
      "learning_rate": 2.788170140758403e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 802926.0,
      "step": 1836
    },
    {
      "epoch": 0.141296823321283,
      "grad_norm": 7.875,
      "learning_rate": 2.7880547650180754e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 803214.0,
      "step": 1837
    },
    {
      "epoch": 0.14137374048150142,
      "grad_norm": 8.375,
      "learning_rate": 2.787939389277748e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9972899556159973,
      "num_tokens": 803586.0,
      "step": 1838
    },
    {
      "epoch": 0.14145065764171988,
      "grad_norm": 2.5,
      "learning_rate": 2.7878240135374204e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 804630.0,
      "step": 1839
    },
    {
      "epoch": 0.1415275748019383,
      "grad_norm": 3.5,
      "learning_rate": 2.7877086377970926e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 805176.0,
      "step": 1840
    },
    {
      "epoch": 0.14160449196215677,
      "grad_norm": 3.828125,
      "learning_rate": 2.787593262056765e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 805744.0,
      "step": 1841
    },
    {
      "epoch": 0.1416814091223752,
      "grad_norm": 9.5625,
      "learning_rate": 2.787477886316437e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 806054.0,
      "step": 1842
    },
    {
      "epoch": 0.14175832628259363,
      "grad_norm": 8.4375,
      "learning_rate": 2.7873625105761094e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 806340.0,
      "step": 1843
    },
    {
      "epoch": 0.1418352434428121,
      "grad_norm": 3.03125,
      "learning_rate": 2.787247134835782e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 806884.0,
      "step": 1844
    },
    {
      "epoch": 0.14191216060303052,
      "grad_norm": 8.0625,
      "learning_rate": 2.7871317590954543e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 807190.0,
      "step": 1845
    },
    {
      "epoch": 0.14198907776324898,
      "grad_norm": 2.859375,
      "learning_rate": 2.7870163833551266e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 807946.0,
      "step": 1846
    },
    {
      "epoch": 0.14206599492346741,
      "grad_norm": 5.03125,
      "learning_rate": 2.786901007614799e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 808254.0,
      "step": 1847
    },
    {
      "epoch": 0.14214291208368587,
      "grad_norm": 5.375,
      "learning_rate": 2.786785631874471e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 808564.0,
      "step": 1848
    },
    {
      "epoch": 0.1422198292439043,
      "grad_norm": 10.9375,
      "learning_rate": 2.786670256134144e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 808936.0,
      "step": 1849
    },
    {
      "epoch": 0.14229674640412276,
      "grad_norm": 2.59375,
      "learning_rate": 2.786554880393816e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 809454.0,
      "step": 1850
    },
    {
      "epoch": 0.1423736635643412,
      "grad_norm": 3.125,
      "learning_rate": 2.7864395046534883e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 810010.0,
      "step": 1851
    },
    {
      "epoch": 0.14245058072455966,
      "grad_norm": 3.296875,
      "learning_rate": 2.7863241289131606e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 810566.0,
      "step": 1852
    },
    {
      "epoch": 0.1425274978847781,
      "grad_norm": 5.8125,
      "learning_rate": 2.786208753172833e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 810890.0,
      "step": 1853
    },
    {
      "epoch": 0.14260441504499655,
      "grad_norm": 4.9375,
      "learning_rate": 2.786093377432505e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 811206.0,
      "step": 1854
    },
    {
      "epoch": 0.14268133220521498,
      "grad_norm": 8.125,
      "learning_rate": 2.7859780016921778e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 811544.0,
      "step": 1855
    },
    {
      "epoch": 0.14275824936543344,
      "grad_norm": 5.78125,
      "learning_rate": 2.78586262595185e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 811874.0,
      "step": 1856
    },
    {
      "epoch": 0.14283516652565187,
      "grad_norm": 3.859375,
      "learning_rate": 2.7857472502115223e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 812442.0,
      "step": 1857
    },
    {
      "epoch": 0.14291208368587033,
      "grad_norm": 4.09375,
      "learning_rate": 2.7856318744711945e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 813014.0,
      "step": 1858
    },
    {
      "epoch": 0.14298900084608876,
      "grad_norm": 9.1875,
      "learning_rate": 2.785516498730867e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 813302.0,
      "step": 1859
    },
    {
      "epoch": 0.14306591800630722,
      "grad_norm": 5.28125,
      "learning_rate": 2.7854011229905395e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 813586.0,
      "step": 1860
    },
    {
      "epoch": 0.14314283516652565,
      "grad_norm": 3.15625,
      "learning_rate": 2.7852857472502117e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 814144.0,
      "step": 1861
    },
    {
      "epoch": 0.1432197523267441,
      "grad_norm": 2.84375,
      "learning_rate": 2.785170371509884e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 815204.0,
      "step": 1862
    },
    {
      "epoch": 0.14329666948696254,
      "grad_norm": 2.75,
      "learning_rate": 2.7850549957695563e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 815762.0,
      "step": 1863
    },
    {
      "epoch": 0.14337358664718097,
      "grad_norm": 7.03125,
      "learning_rate": 2.7849396200292285e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 816066.0,
      "step": 1864
    },
    {
      "epoch": 0.14345050380739943,
      "grad_norm": 5.25,
      "learning_rate": 2.7848242442889008e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 816372.0,
      "step": 1865
    },
    {
      "epoch": 0.14352742096761786,
      "grad_norm": 5.9375,
      "learning_rate": 2.7847088685485735e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 816654.0,
      "step": 1866
    },
    {
      "epoch": 0.14360433812783632,
      "grad_norm": 3.953125,
      "learning_rate": 2.7845934928082457e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 817218.0,
      "step": 1867
    },
    {
      "epoch": 0.14368125528805475,
      "grad_norm": 2.46875,
      "learning_rate": 2.784478117067918e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 818004.0,
      "step": 1868
    },
    {
      "epoch": 0.1437581724482732,
      "grad_norm": 6.15625,
      "learning_rate": 2.7843627413275902e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 818304.0,
      "step": 1869
    },
    {
      "epoch": 0.14383508960849165,
      "grad_norm": 3.8125,
      "learning_rate": 2.7842473655872625e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 818876.0,
      "step": 1870
    },
    {
      "epoch": 0.1439120067687101,
      "grad_norm": 8.25,
      "learning_rate": 2.784131989846935e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 819192.0,
      "step": 1871
    },
    {
      "epoch": 0.14398892392892854,
      "grad_norm": 7.84375,
      "learning_rate": 2.7840166141066074e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 819504.0,
      "step": 1872
    },
    {
      "epoch": 0.144065841089147,
      "grad_norm": 4.9375,
      "learning_rate": 2.7839012383662797e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 819828.0,
      "step": 1873
    },
    {
      "epoch": 0.14414275824936543,
      "grad_norm": 8.1875,
      "learning_rate": 2.783785862625952e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 820136.0,
      "step": 1874
    },
    {
      "epoch": 0.1442196754095839,
      "grad_norm": 7.125,
      "learning_rate": 2.783670486885624e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 820488.0,
      "step": 1875
    },
    {
      "epoch": 0.14429659256980232,
      "grad_norm": 5.53125,
      "learning_rate": 2.7835551111452965e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 820796.0,
      "step": 1876
    },
    {
      "epoch": 0.14437350973002078,
      "grad_norm": 2.84375,
      "learning_rate": 2.783439735404969e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 821380.0,
      "step": 1877
    },
    {
      "epoch": 0.1444504268902392,
      "grad_norm": 2.5,
      "learning_rate": 2.7833243596646414e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 822160.0,
      "step": 1878
    },
    {
      "epoch": 0.14452734405045767,
      "grad_norm": 7.09375,
      "learning_rate": 2.7832089839243137e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 822484.0,
      "step": 1879
    },
    {
      "epoch": 0.1446042612106761,
      "grad_norm": 2.96875,
      "learning_rate": 2.783093608183986e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 823070.0,
      "step": 1880
    },
    {
      "epoch": 0.14468117837089456,
      "grad_norm": 4.28125,
      "learning_rate": 2.782978232443658e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 823642.0,
      "step": 1881
    },
    {
      "epoch": 0.144758095531113,
      "grad_norm": 6.25,
      "learning_rate": 2.7828628567033304e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 823998.0,
      "step": 1882
    },
    {
      "epoch": 0.14483501269133145,
      "grad_norm": 2.53125,
      "learning_rate": 2.782747480963003e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 824546.0,
      "step": 1883
    },
    {
      "epoch": 0.14491192985154988,
      "grad_norm": 3.671875,
      "learning_rate": 2.7826321052226754e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 825068.0,
      "step": 1884
    },
    {
      "epoch": 0.1449888470117683,
      "grad_norm": 5.96875,
      "learning_rate": 2.7825167294823476e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 825356.0,
      "step": 1885
    },
    {
      "epoch": 0.14506576417198677,
      "grad_norm": 4.375,
      "learning_rate": 2.78240135374202e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 825706.0,
      "step": 1886
    },
    {
      "epoch": 0.1451426813322052,
      "grad_norm": 3.578125,
      "learning_rate": 2.782285978001692e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 826224.0,
      "step": 1887
    },
    {
      "epoch": 0.14521959849242366,
      "grad_norm": 6.9375,
      "learning_rate": 2.782170602261365e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 826570.0,
      "step": 1888
    },
    {
      "epoch": 0.1452965156526421,
      "grad_norm": 7.3125,
      "learning_rate": 2.782055226521037e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 826880.0,
      "step": 1889
    },
    {
      "epoch": 0.14537343281286055,
      "grad_norm": 5.21875,
      "learning_rate": 2.781939850780709e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 827190.0,
      "step": 1890
    },
    {
      "epoch": 0.14545034997307899,
      "grad_norm": 6.4375,
      "learning_rate": 2.7818244750403816e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9795918464660645,
      "num_tokens": 827534.0,
      "step": 1891
    },
    {
      "epoch": 0.14552726713329744,
      "grad_norm": 5.75,
      "learning_rate": 2.781709099300054e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 827864.0,
      "step": 1892
    },
    {
      "epoch": 0.14560418429351588,
      "grad_norm": 3.453125,
      "learning_rate": 2.781593723559726e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 828382.0,
      "step": 1893
    },
    {
      "epoch": 0.14568110145373434,
      "grad_norm": 8.875,
      "learning_rate": 2.781478347819399e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 828644.0,
      "step": 1894
    },
    {
      "epoch": 0.14575801861395277,
      "grad_norm": 5.03125,
      "learning_rate": 2.7813629720790706e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 828988.0,
      "step": 1895
    },
    {
      "epoch": 0.14583493577417123,
      "grad_norm": 4.6875,
      "learning_rate": 2.7812475963387433e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 829310.0,
      "step": 1896
    },
    {
      "epoch": 0.14591185293438966,
      "grad_norm": 7.71875,
      "learning_rate": 2.7811322205984156e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 829622.0,
      "step": 1897
    },
    {
      "epoch": 0.14598877009460812,
      "grad_norm": 5.625,
      "learning_rate": 2.781016844858088e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 829940.0,
      "step": 1898
    },
    {
      "epoch": 0.14606568725482655,
      "grad_norm": 3.890625,
      "learning_rate": 2.7809014691177605e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 830504.0,
      "step": 1899
    },
    {
      "epoch": 0.146142604415045,
      "grad_norm": 3.46875,
      "learning_rate": 2.7807860933774328e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 831078.0,
      "step": 1900
    },
    {
      "epoch": 0.14621952157526344,
      "grad_norm": 2.25,
      "learning_rate": 2.7806707176371046e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 831640.0,
      "step": 1901
    },
    {
      "epoch": 0.1462964387354819,
      "grad_norm": 3.796875,
      "learning_rate": 2.7805553418967773e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 831962.0,
      "step": 1902
    },
    {
      "epoch": 0.14637335589570033,
      "grad_norm": 7.40625,
      "learning_rate": 2.7804399661564495e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 832254.0,
      "step": 1903
    },
    {
      "epoch": 0.1464502730559188,
      "grad_norm": 4.9375,
      "learning_rate": 2.780324590416122e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 832774.0,
      "step": 1904
    },
    {
      "epoch": 0.14652719021613722,
      "grad_norm": 3.78125,
      "learning_rate": 2.7802092146757945e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 833590.0,
      "step": 1905
    },
    {
      "epoch": 0.14660410737635565,
      "grad_norm": 2.5625,
      "learning_rate": 2.7800938389354663e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 834150.0,
      "step": 1906
    },
    {
      "epoch": 0.1466810245365741,
      "grad_norm": 3.015625,
      "learning_rate": 2.779978463195139e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 834938.0,
      "step": 1907
    },
    {
      "epoch": 0.14675794169679254,
      "grad_norm": 3.0625,
      "learning_rate": 2.7798630874548113e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 835534.0,
      "step": 1908
    },
    {
      "epoch": 0.146834858857011,
      "grad_norm": 8.125,
      "learning_rate": 2.7797477117144835e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 835820.0,
      "step": 1909
    },
    {
      "epoch": 0.14691177601722943,
      "grad_norm": 4.21875,
      "learning_rate": 2.779632335974156e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 836398.0,
      "step": 1910
    },
    {
      "epoch": 0.1469886931774479,
      "grad_norm": 8.125,
      "learning_rate": 2.7795169602338285e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 836702.0,
      "step": 1911
    },
    {
      "epoch": 0.14706561033766632,
      "grad_norm": 3.171875,
      "learning_rate": 2.7794015844935003e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 837272.0,
      "step": 1912
    },
    {
      "epoch": 0.14714252749788478,
      "grad_norm": 4.0,
      "learning_rate": 2.779286208753173e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 837792.0,
      "step": 1913
    },
    {
      "epoch": 0.14721944465810322,
      "grad_norm": 7.9375,
      "learning_rate": 2.7791708330128452e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 838094.0,
      "step": 1914
    },
    {
      "epoch": 0.14729636181832167,
      "grad_norm": 6.6875,
      "learning_rate": 2.7790554572725175e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 838402.0,
      "step": 1915
    },
    {
      "epoch": 0.1473732789785401,
      "grad_norm": 4.0,
      "learning_rate": 2.77894008153219e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 838982.0,
      "step": 1916
    },
    {
      "epoch": 0.14745019613875857,
      "grad_norm": 10.125,
      "learning_rate": 2.778824705791862e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 839324.0,
      "step": 1917
    },
    {
      "epoch": 0.147527113298977,
      "grad_norm": 6.96875,
      "learning_rate": 2.7787093300515347e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 839684.0,
      "step": 1918
    },
    {
      "epoch": 0.14760403045919546,
      "grad_norm": 8.0625,
      "learning_rate": 2.778593954311207e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 839954.0,
      "step": 1919
    },
    {
      "epoch": 0.1476809476194139,
      "grad_norm": 5.6875,
      "learning_rate": 2.778478578570879e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 840282.0,
      "step": 1920
    },
    {
      "epoch": 0.14775786477963235,
      "grad_norm": 6.59375,
      "learning_rate": 2.778363202830552e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 840588.0,
      "step": 1921
    },
    {
      "epoch": 0.14783478193985078,
      "grad_norm": 4.96875,
      "learning_rate": 2.7782478270902237e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 840934.0,
      "step": 1922
    },
    {
      "epoch": 0.14791169910006924,
      "grad_norm": 2.640625,
      "learning_rate": 2.778132451349896e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 841954.0,
      "step": 1923
    },
    {
      "epoch": 0.14798861626028767,
      "grad_norm": 3.359375,
      "learning_rate": 2.7780170756095687e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 842512.0,
      "step": 1924
    },
    {
      "epoch": 0.1480655334205061,
      "grad_norm": 3.390625,
      "learning_rate": 2.777901699869241e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 843030.0,
      "step": 1925
    },
    {
      "epoch": 0.14814245058072456,
      "grad_norm": 6.75,
      "learning_rate": 2.777786324128913e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 843332.0,
      "step": 1926
    },
    {
      "epoch": 0.148219367740943,
      "grad_norm": 2.640625,
      "learning_rate": 2.777670948388586e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 843878.0,
      "step": 1927
    },
    {
      "epoch": 0.14829628490116145,
      "grad_norm": 7.25,
      "learning_rate": 2.7775555726482577e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 844240.0,
      "step": 1928
    },
    {
      "epoch": 0.14837320206137988,
      "grad_norm": 4.71875,
      "learning_rate": 2.7774401969079304e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 844736.0,
      "step": 1929
    },
    {
      "epoch": 0.14845011922159834,
      "grad_norm": 7.6875,
      "learning_rate": 2.7773248211676026e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 845032.0,
      "step": 1930
    },
    {
      "epoch": 0.14852703638181677,
      "grad_norm": 3.046875,
      "learning_rate": 2.777209445427275e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 845580.0,
      "step": 1931
    },
    {
      "epoch": 0.14860395354203523,
      "grad_norm": 7.15625,
      "learning_rate": 2.777094069686947e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 845912.0,
      "step": 1932
    },
    {
      "epoch": 0.14868087070225366,
      "grad_norm": 6.84375,
      "learning_rate": 2.7769786939466194e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 846246.0,
      "step": 1933
    },
    {
      "epoch": 0.14875778786247212,
      "grad_norm": 3.609375,
      "learning_rate": 2.7768633182062917e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 847058.0,
      "step": 1934
    },
    {
      "epoch": 0.14883470502269056,
      "grad_norm": 7.46875,
      "learning_rate": 2.7767479424659643e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 847402.0,
      "step": 1935
    },
    {
      "epoch": 0.14891162218290901,
      "grad_norm": 7.21875,
      "learning_rate": 2.7766325667256366e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 847730.0,
      "step": 1936
    },
    {
      "epoch": 0.14898853934312745,
      "grad_norm": 6.28125,
      "learning_rate": 2.776517190985309e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9933554530143738,
      "num_tokens": 848032.0,
      "step": 1937
    },
    {
      "epoch": 0.1490654565033459,
      "grad_norm": 8.75,
      "learning_rate": 2.7764018152449815e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 848288.0,
      "step": 1938
    },
    {
      "epoch": 0.14914237366356434,
      "grad_norm": 2.5,
      "learning_rate": 2.7762864395046534e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 849096.0,
      "step": 1939
    },
    {
      "epoch": 0.1492192908237828,
      "grad_norm": 4.125,
      "learning_rate": 2.7761710637643256e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9913644194602966,
      "num_tokens": 849676.0,
      "step": 1940
    },
    {
      "epoch": 0.14929620798400123,
      "grad_norm": 4.46875,
      "learning_rate": 2.7760556880239983e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 850216.0,
      "step": 1941
    },
    {
      "epoch": 0.1493731251442197,
      "grad_norm": 9.75,
      "learning_rate": 2.7759403122836706e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 850544.0,
      "step": 1942
    },
    {
      "epoch": 0.14945004230443812,
      "grad_norm": 8.1875,
      "learning_rate": 2.775824936543343e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 850900.0,
      "step": 1943
    },
    {
      "epoch": 0.14952695946465658,
      "grad_norm": 2.59375,
      "learning_rate": 2.775709560803015e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 851486.0,
      "step": 1944
    },
    {
      "epoch": 0.149603876624875,
      "grad_norm": 5.90625,
      "learning_rate": 2.7755941850626873e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 851792.0,
      "step": 1945
    },
    {
      "epoch": 0.14968079378509344,
      "grad_norm": 5.28125,
      "learning_rate": 2.77547880932236e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 852126.0,
      "step": 1946
    },
    {
      "epoch": 0.1497577109453119,
      "grad_norm": 3.109375,
      "learning_rate": 2.7753634335820323e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 852666.0,
      "step": 1947
    },
    {
      "epoch": 0.14983462810553033,
      "grad_norm": 9.0625,
      "learning_rate": 2.7752480578417045e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9819819927215576,
      "num_tokens": 853000.0,
      "step": 1948
    },
    {
      "epoch": 0.1499115452657488,
      "grad_norm": 7.65625,
      "learning_rate": 2.7751326821013772e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9845201373100281,
      "num_tokens": 853324.0,
      "step": 1949
    },
    {
      "epoch": 0.14998846242596722,
      "grad_norm": 2.734375,
      "learning_rate": 2.775017306361049e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 853910.0,
      "step": 1950
    },
    {
      "epoch": 0.15006537958618568,
      "grad_norm": 4.15625,
      "learning_rate": 2.7749019306207213e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 854256.0,
      "step": 1951
    },
    {
      "epoch": 0.1501422967464041,
      "grad_norm": 7.53125,
      "learning_rate": 2.774786554880394e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 854592.0,
      "step": 1952
    },
    {
      "epoch": 0.15021921390662257,
      "grad_norm": 5.75,
      "learning_rate": 2.7746711791400663e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 854918.0,
      "step": 1953
    },
    {
      "epoch": 0.150296131066841,
      "grad_norm": 3.09375,
      "learning_rate": 2.7745558033997385e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 855464.0,
      "step": 1954
    },
    {
      "epoch": 0.15037304822705946,
      "grad_norm": 6.71875,
      "learning_rate": 2.7744404276594108e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 855794.0,
      "step": 1955
    },
    {
      "epoch": 0.1504499653872779,
      "grad_norm": 7.8125,
      "learning_rate": 2.774325051919083e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 856090.0,
      "step": 1956
    },
    {
      "epoch": 0.15052688254749635,
      "grad_norm": 3.34375,
      "learning_rate": 2.7742096761787557e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 856646.0,
      "step": 1957
    },
    {
      "epoch": 0.15060379970771479,
      "grad_norm": 5.78125,
      "learning_rate": 2.774094300438428e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 856976.0,
      "step": 1958
    },
    {
      "epoch": 0.15068071686793325,
      "grad_norm": 5.75,
      "learning_rate": 2.7739789246981002e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 857306.0,
      "step": 1959
    },
    {
      "epoch": 0.15075763402815168,
      "grad_norm": 5.5625,
      "learning_rate": 2.7738635489577725e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 857666.0,
      "step": 1960
    },
    {
      "epoch": 0.15083455118837014,
      "grad_norm": 4.5625,
      "learning_rate": 2.7737481732174447e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 858188.0,
      "step": 1961
    },
    {
      "epoch": 0.15091146834858857,
      "grad_norm": 9.125,
      "learning_rate": 2.773632797477117e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 858478.0,
      "step": 1962
    },
    {
      "epoch": 0.15098838550880703,
      "grad_norm": 9.0,
      "learning_rate": 2.7735174217367897e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 858764.0,
      "step": 1963
    },
    {
      "epoch": 0.15106530266902546,
      "grad_norm": 2.765625,
      "learning_rate": 2.773402045996462e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 859314.0,
      "step": 1964
    },
    {
      "epoch": 0.15114221982924392,
      "grad_norm": 9.875,
      "learning_rate": 2.773286670256134e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 859670.0,
      "step": 1965
    },
    {
      "epoch": 0.15121913698946235,
      "grad_norm": 4.375,
      "learning_rate": 2.7731712945158065e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 860242.0,
      "step": 1966
    },
    {
      "epoch": 0.15129605414968078,
      "grad_norm": 5.25,
      "learning_rate": 2.7730559187754787e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 860608.0,
      "step": 1967
    },
    {
      "epoch": 0.15137297130989924,
      "grad_norm": 9.0,
      "learning_rate": 2.7729405430351514e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 860948.0,
      "step": 1968
    },
    {
      "epoch": 0.15144988847011767,
      "grad_norm": 5.9375,
      "learning_rate": 2.7728251672948237e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 861304.0,
      "step": 1969
    },
    {
      "epoch": 0.15152680563033613,
      "grad_norm": 3.90625,
      "learning_rate": 2.772709791554496e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 861874.0,
      "step": 1970
    },
    {
      "epoch": 0.15160372279055456,
      "grad_norm": 7.1875,
      "learning_rate": 2.772594415814168e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 862166.0,
      "step": 1971
    },
    {
      "epoch": 0.15168063995077302,
      "grad_norm": 5.0625,
      "learning_rate": 2.7724790400738404e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 862662.0,
      "step": 1972
    },
    {
      "epoch": 0.15175755711099145,
      "grad_norm": 2.6875,
      "learning_rate": 2.7723636643335127e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 863472.0,
      "step": 1973
    },
    {
      "epoch": 0.1518344742712099,
      "grad_norm": 8.3125,
      "learning_rate": 2.7722482885931854e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 863750.0,
      "step": 1974
    },
    {
      "epoch": 0.15191139143142834,
      "grad_norm": 4.0625,
      "learning_rate": 2.7721329128528576e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 864326.0,
      "step": 1975
    },
    {
      "epoch": 0.1519883085916468,
      "grad_norm": 8.75,
      "learning_rate": 2.77201753711253e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 864608.0,
      "step": 1976
    },
    {
      "epoch": 0.15206522575186524,
      "grad_norm": 3.828125,
      "learning_rate": 2.771902161372202e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 865124.0,
      "step": 1977
    },
    {
      "epoch": 0.1521421429120837,
      "grad_norm": 3.53125,
      "learning_rate": 2.7717867856318744e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 865710.0,
      "step": 1978
    },
    {
      "epoch": 0.15221906007230213,
      "grad_norm": 6.5625,
      "learning_rate": 2.771671409891547e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 866038.0,
      "step": 1979
    },
    {
      "epoch": 0.15229597723252059,
      "grad_norm": 9.8125,
      "learning_rate": 2.7715560341512193e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 866350.0,
      "step": 1980
    },
    {
      "epoch": 0.15237289439273902,
      "grad_norm": 3.25,
      "learning_rate": 2.7714406584108916e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 866908.0,
      "step": 1981
    },
    {
      "epoch": 0.15244981155295748,
      "grad_norm": 5.46875,
      "learning_rate": 2.771325282670564e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 867258.0,
      "step": 1982
    },
    {
      "epoch": 0.1525267287131759,
      "grad_norm": 3.421875,
      "learning_rate": 2.771209906930236e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 867820.0,
      "step": 1983
    },
    {
      "epoch": 0.15260364587339437,
      "grad_norm": 4.34375,
      "learning_rate": 2.7710945311899084e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 868146.0,
      "step": 1984
    },
    {
      "epoch": 0.1526805630336128,
      "grad_norm": 3.109375,
      "learning_rate": 2.770979155449581e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 868686.0,
      "step": 1985
    },
    {
      "epoch": 0.15275748019383123,
      "grad_norm": 5.09375,
      "learning_rate": 2.7708637797092533e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 868984.0,
      "step": 1986
    },
    {
      "epoch": 0.1528343973540497,
      "grad_norm": 7.4375,
      "learning_rate": 2.7707484039689256e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 869314.0,
      "step": 1987
    },
    {
      "epoch": 0.15291131451426812,
      "grad_norm": 4.5625,
      "learning_rate": 2.770633028228598e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 869634.0,
      "step": 1988
    },
    {
      "epoch": 0.15298823167448658,
      "grad_norm": 7.59375,
      "learning_rate": 2.77051765248827e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 869964.0,
      "step": 1989
    },
    {
      "epoch": 0.153065148834705,
      "grad_norm": 6.21875,
      "learning_rate": 2.7704022767479424e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 870296.0,
      "step": 1990
    },
    {
      "epoch": 0.15314206599492347,
      "grad_norm": 13.6875,
      "learning_rate": 2.770286901007615e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9883720874786377,
      "num_tokens": 870644.0,
      "step": 1991
    },
    {
      "epoch": 0.1532189831551419,
      "grad_norm": 8.4375,
      "learning_rate": 2.7701715252672873e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 870924.0,
      "step": 1992
    },
    {
      "epoch": 0.15329590031536036,
      "grad_norm": 3.296875,
      "learning_rate": 2.7700561495269595e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 871466.0,
      "step": 1993
    },
    {
      "epoch": 0.1533728174755788,
      "grad_norm": 2.65625,
      "learning_rate": 2.769940773786632e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 872012.0,
      "step": 1994
    },
    {
      "epoch": 0.15344973463579725,
      "grad_norm": 8.1875,
      "learning_rate": 2.769825398046304e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 872318.0,
      "step": 1995
    },
    {
      "epoch": 0.15352665179601568,
      "grad_norm": 3.265625,
      "learning_rate": 2.7697100223059767e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 872866.0,
      "step": 1996
    },
    {
      "epoch": 0.15360356895623414,
      "grad_norm": 2.640625,
      "learning_rate": 2.769594646565649e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 873690.0,
      "step": 1997
    },
    {
      "epoch": 0.15368048611645257,
      "grad_norm": 6.59375,
      "learning_rate": 2.769479270825321e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 874026.0,
      "step": 1998
    },
    {
      "epoch": 0.15375740327667103,
      "grad_norm": 6.21875,
      "learning_rate": 2.7693638950849935e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 874308.0,
      "step": 1999
    },
    {
      "epoch": 0.15383432043688947,
      "grad_norm": 6.59375,
      "learning_rate": 2.7692485193446658e-06,
      "loss": 0.0222,
      "step": 2000
    },
    {
      "epoch": 0.15383432043688947,
      "eval_loss": 0.021004218608140945,
      "eval_mean_token_accuracy": 0.991381471157074,
      "eval_num_tokens": 874806.0,
      "eval_runtime": 21.0984,
      "eval_samples_per_second": 9.479,
      "eval_steps_per_second": 1.185,
      "step": 2000
    },
    {
      "epoch": 0.15391123759710792,
      "grad_norm": 4.28125,
      "learning_rate": 2.769133143604338e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9913029670715332,
      "num_tokens": 875128.0,
      "step": 2001
    },
    {
      "epoch": 0.15398815475732636,
      "grad_norm": 4.0,
      "learning_rate": 2.7690177678640107e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 875622.0,
      "step": 2002
    },
    {
      "epoch": 0.15406507191754482,
      "grad_norm": 8.3125,
      "learning_rate": 2.768902392123683e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 875990.0,
      "step": 2003
    },
    {
      "epoch": 0.15414198907776325,
      "grad_norm": 5.4375,
      "learning_rate": 2.7687870163833552e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 876302.0,
      "step": 2004
    },
    {
      "epoch": 0.1542189062379817,
      "grad_norm": 2.890625,
      "learning_rate": 2.7686716406430275e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 876854.0,
      "step": 2005
    },
    {
      "epoch": 0.15429582339820014,
      "grad_norm": 7.15625,
      "learning_rate": 2.7685562649026998e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 877196.0,
      "step": 2006
    },
    {
      "epoch": 0.15437274055841857,
      "grad_norm": 8.125,
      "learning_rate": 2.7684408891623724e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 877552.0,
      "step": 2007
    },
    {
      "epoch": 0.15444965771863703,
      "grad_norm": 3.40625,
      "learning_rate": 2.7683255134220447e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 878350.0,
      "step": 2008
    },
    {
      "epoch": 0.15452657487885546,
      "grad_norm": 6.84375,
      "learning_rate": 2.7682101376817165e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 878672.0,
      "step": 2009
    },
    {
      "epoch": 0.15460349203907392,
      "grad_norm": 9.9375,
      "learning_rate": 2.768094761941389e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 878952.0,
      "step": 2010
    },
    {
      "epoch": 0.15468040919929235,
      "grad_norm": 4.8125,
      "learning_rate": 2.7679793862010615e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 879252.0,
      "step": 2011
    },
    {
      "epoch": 0.1547573263595108,
      "grad_norm": 2.53125,
      "learning_rate": 2.7678640104607337e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 879798.0,
      "step": 2012
    },
    {
      "epoch": 0.15483424351972924,
      "grad_norm": 3.296875,
      "learning_rate": 2.7677486347204064e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 880602.0,
      "step": 2013
    },
    {
      "epoch": 0.1549111606799477,
      "grad_norm": 8.75,
      "learning_rate": 2.7676332589800787e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 880904.0,
      "step": 2014
    },
    {
      "epoch": 0.15498807784016613,
      "grad_norm": 4.25,
      "learning_rate": 2.767517883239751e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 881424.0,
      "step": 2015
    },
    {
      "epoch": 0.1550649950003846,
      "grad_norm": 3.921875,
      "learning_rate": 2.767402507499423e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 881708.0,
      "step": 2016
    },
    {
      "epoch": 0.15514191216060302,
      "grad_norm": 6.8125,
      "learning_rate": 2.7672871317590954e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 882008.0,
      "step": 2017
    },
    {
      "epoch": 0.15521882932082148,
      "grad_norm": 2.328125,
      "learning_rate": 2.767171756018768e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 882812.0,
      "step": 2018
    },
    {
      "epoch": 0.15529574648103991,
      "grad_norm": 3.0,
      "learning_rate": 2.7670563802784404e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 883376.0,
      "step": 2019
    },
    {
      "epoch": 0.15537266364125837,
      "grad_norm": 3.015625,
      "learning_rate": 2.766941004538112e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 883920.0,
      "step": 2020
    },
    {
      "epoch": 0.1554495808014768,
      "grad_norm": 7.1875,
      "learning_rate": 2.766825628797785e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 884236.0,
      "step": 2021
    },
    {
      "epoch": 0.15552649796169526,
      "grad_norm": 5.0,
      "learning_rate": 2.766710253057457e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 884536.0,
      "step": 2022
    },
    {
      "epoch": 0.1556034151219137,
      "grad_norm": 5.71875,
      "learning_rate": 2.7665948773171294e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 884848.0,
      "step": 2023
    },
    {
      "epoch": 0.15568033228213216,
      "grad_norm": 1.9921875,
      "learning_rate": 2.766479501576802e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 885606.0,
      "step": 2024
    },
    {
      "epoch": 0.1557572494423506,
      "grad_norm": 2.78125,
      "learning_rate": 2.766364125836474e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 886150.0,
      "step": 2025
    },
    {
      "epoch": 0.15583416660256905,
      "grad_norm": 2.421875,
      "learning_rate": 2.7662487500961466e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 886708.0,
      "step": 2026
    },
    {
      "epoch": 0.15591108376278748,
      "grad_norm": 2.0,
      "learning_rate": 2.766133374355819e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 887250.0,
      "step": 2027
    },
    {
      "epoch": 0.1559880009230059,
      "grad_norm": 7.34375,
      "learning_rate": 2.766017998615491e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 887510.0,
      "step": 2028
    },
    {
      "epoch": 0.15606491808322437,
      "grad_norm": 2.734375,
      "learning_rate": 2.765902622875164e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 888062.0,
      "step": 2029
    },
    {
      "epoch": 0.1561418352434428,
      "grad_norm": 2.15625,
      "learning_rate": 2.765787247134836e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 888870.0,
      "step": 2030
    },
    {
      "epoch": 0.15621875240366126,
      "grad_norm": 2.953125,
      "learning_rate": 2.765671871394508e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 889456.0,
      "step": 2031
    },
    {
      "epoch": 0.1562956695638797,
      "grad_norm": 4.125,
      "learning_rate": 2.7655564956541806e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 889998.0,
      "step": 2032
    },
    {
      "epoch": 0.15637258672409815,
      "grad_norm": 2.8125,
      "learning_rate": 2.765441119913853e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 890584.0,
      "step": 2033
    },
    {
      "epoch": 0.15644950388431658,
      "grad_norm": 6.125,
      "learning_rate": 2.765325744173525e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 890928.0,
      "step": 2034
    },
    {
      "epoch": 0.15652642104453504,
      "grad_norm": 6.34375,
      "learning_rate": 2.7652103684331978e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 891234.0,
      "step": 2035
    },
    {
      "epoch": 0.15660333820475347,
      "grad_norm": 6.1875,
      "learning_rate": 2.7650949926928696e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 891522.0,
      "step": 2036
    },
    {
      "epoch": 0.15668025536497193,
      "grad_norm": 3.75,
      "learning_rate": 2.7649796169525423e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 892044.0,
      "step": 2037
    },
    {
      "epoch": 0.15675717252519036,
      "grad_norm": 8.0625,
      "learning_rate": 2.7648642412122146e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 892328.0,
      "step": 2038
    },
    {
      "epoch": 0.15683408968540882,
      "grad_norm": 2.765625,
      "learning_rate": 2.764748865471887e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 892876.0,
      "step": 2039
    },
    {
      "epoch": 0.15691100684562725,
      "grad_norm": 8.0,
      "learning_rate": 2.764633489731559e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 893244.0,
      "step": 2040
    },
    {
      "epoch": 0.1569879240058457,
      "grad_norm": 5.1875,
      "learning_rate": 2.7645181139912317e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 893542.0,
      "step": 2041
    },
    {
      "epoch": 0.15706484116606415,
      "grad_norm": 5.09375,
      "learning_rate": 2.7644027382509036e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 893864.0,
      "step": 2042
    },
    {
      "epoch": 0.1571417583262826,
      "grad_norm": 6.28125,
      "learning_rate": 2.7642873625105763e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 894168.0,
      "step": 2043
    },
    {
      "epoch": 0.15721867548650104,
      "grad_norm": 2.671875,
      "learning_rate": 2.7641719867702485e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 894752.0,
      "step": 2044
    },
    {
      "epoch": 0.1572955926467195,
      "grad_norm": 7.21875,
      "learning_rate": 2.7640566110299208e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9929577708244324,
      "num_tokens": 895038.0,
      "step": 2045
    },
    {
      "epoch": 0.15737250980693793,
      "grad_norm": 2.78125,
      "learning_rate": 2.7639412352895935e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 895596.0,
      "step": 2046
    },
    {
      "epoch": 0.1574494269671564,
      "grad_norm": 2.828125,
      "learning_rate": 2.7638258595492653e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 896154.0,
      "step": 2047
    },
    {
      "epoch": 0.15752634412737482,
      "grad_norm": 8.3125,
      "learning_rate": 2.7637104838089376e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 896442.0,
      "step": 2048
    },
    {
      "epoch": 0.15760326128759325,
      "grad_norm": 2.84375,
      "learning_rate": 2.7635951080686102e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 897016.0,
      "step": 2049
    },
    {
      "epoch": 0.1576801784478117,
      "grad_norm": 8.5625,
      "learning_rate": 2.7634797323282825e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 897294.0,
      "step": 2050
    },
    {
      "epoch": 0.15775709560803014,
      "grad_norm": 6.0625,
      "learning_rate": 2.7633643565879548e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 897612.0,
      "step": 2051
    },
    {
      "epoch": 0.1578340127682486,
      "grad_norm": 5.8125,
      "learning_rate": 2.7632489808476274e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 897952.0,
      "step": 2052
    },
    {
      "epoch": 0.15791092992846703,
      "grad_norm": 12.5,
      "learning_rate": 2.7631336051072993e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 898204.0,
      "step": 2053
    },
    {
      "epoch": 0.1579878470886855,
      "grad_norm": 6.46875,
      "learning_rate": 2.763018229366972e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 898536.0,
      "step": 2054
    },
    {
      "epoch": 0.15806476424890392,
      "grad_norm": 7.0625,
      "learning_rate": 2.762902853626644e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 898838.0,
      "step": 2055
    },
    {
      "epoch": 0.15814168140912238,
      "grad_norm": 4.375,
      "learning_rate": 2.7627874778863165e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 899162.0,
      "step": 2056
    },
    {
      "epoch": 0.1582185985693408,
      "grad_norm": 6.09375,
      "learning_rate": 2.762672102145989e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 899464.0,
      "step": 2057
    },
    {
      "epoch": 0.15829551572955927,
      "grad_norm": 6.0625,
      "learning_rate": 2.762556726405661e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 899760.0,
      "step": 2058
    },
    {
      "epoch": 0.1583724328897777,
      "grad_norm": 7.375,
      "learning_rate": 2.7624413506653332e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 900042.0,
      "step": 2059
    },
    {
      "epoch": 0.15844935004999616,
      "grad_norm": 6.78125,
      "learning_rate": 2.762325974925006e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 900344.0,
      "step": 2060
    },
    {
      "epoch": 0.1585262672102146,
      "grad_norm": 7.96875,
      "learning_rate": 2.762210599184678e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 900624.0,
      "step": 2061
    },
    {
      "epoch": 0.15860318437043305,
      "grad_norm": 6.21875,
      "learning_rate": 2.7620952234443504e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 900930.0,
      "step": 2062
    },
    {
      "epoch": 0.15868010153065149,
      "grad_norm": 2.75,
      "learning_rate": 2.7619798477040227e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 901478.0,
      "step": 2063
    },
    {
      "epoch": 0.15875701869086994,
      "grad_norm": 4.75,
      "learning_rate": 2.761864471963695e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 901760.0,
      "step": 2064
    },
    {
      "epoch": 0.15883393585108838,
      "grad_norm": 4.9375,
      "learning_rate": 2.7617490962233676e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 902066.0,
      "step": 2065
    },
    {
      "epoch": 0.15891085301130684,
      "grad_norm": 5.78125,
      "learning_rate": 2.76163372048304e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 902352.0,
      "step": 2066
    },
    {
      "epoch": 0.15898777017152527,
      "grad_norm": 7.84375,
      "learning_rate": 2.761518344742712e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 902696.0,
      "step": 2067
    },
    {
      "epoch": 0.1590646873317437,
      "grad_norm": 3.8125,
      "learning_rate": 2.761402969002385e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 903276.0,
      "step": 2068
    },
    {
      "epoch": 0.15914160449196216,
      "grad_norm": 9.75,
      "learning_rate": 2.7612875932620567e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 903638.0,
      "step": 2069
    },
    {
      "epoch": 0.1592185216521806,
      "grad_norm": 4.71875,
      "learning_rate": 2.761172217521729e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 903960.0,
      "step": 2070
    },
    {
      "epoch": 0.15929543881239905,
      "grad_norm": 3.59375,
      "learning_rate": 2.7610568417814016e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 904524.0,
      "step": 2071
    },
    {
      "epoch": 0.15937235597261748,
      "grad_norm": 6.09375,
      "learning_rate": 2.760941466041074e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 904808.0,
      "step": 2072
    },
    {
      "epoch": 0.15944927313283594,
      "grad_norm": 3.359375,
      "learning_rate": 2.760826090300746e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 905628.0,
      "step": 2073
    },
    {
      "epoch": 0.15952619029305437,
      "grad_norm": 6.15625,
      "learning_rate": 2.7607107145604184e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 905952.0,
      "step": 2074
    },
    {
      "epoch": 0.15960310745327283,
      "grad_norm": 4.375,
      "learning_rate": 2.7605953388200906e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 906302.0,
      "step": 2075
    },
    {
      "epoch": 0.15968002461349126,
      "grad_norm": 6.21875,
      "learning_rate": 2.7604799630797633e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 906580.0,
      "step": 2076
    },
    {
      "epoch": 0.15975694177370972,
      "grad_norm": 5.84375,
      "learning_rate": 2.7603645873394356e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 906902.0,
      "step": 2077
    },
    {
      "epoch": 0.15983385893392815,
      "grad_norm": 3.1875,
      "learning_rate": 2.760249211599108e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 907486.0,
      "step": 2078
    },
    {
      "epoch": 0.1599107760941466,
      "grad_norm": 3.4375,
      "learning_rate": 2.7601338358587805e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 907812.0,
      "step": 2079
    },
    {
      "epoch": 0.15998769325436504,
      "grad_norm": 5.5,
      "learning_rate": 2.7600184601184524e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 908136.0,
      "step": 2080
    },
    {
      "epoch": 0.1600646104145835,
      "grad_norm": 6.34375,
      "learning_rate": 2.7599030843781246e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 908474.0,
      "step": 2081
    },
    {
      "epoch": 0.16014152757480193,
      "grad_norm": 7.90625,
      "learning_rate": 2.7597877086377973e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 908774.0,
      "step": 2082
    },
    {
      "epoch": 0.1602184447350204,
      "grad_norm": 3.75,
      "learning_rate": 2.7596723328974696e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 909532.0,
      "step": 2083
    },
    {
      "epoch": 0.16029536189523882,
      "grad_norm": 10.6875,
      "learning_rate": 2.759556957157142e-06,
      "loss": 0.0393,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 909788.0,
      "step": 2084
    },
    {
      "epoch": 0.16037227905545728,
      "grad_norm": 8.0625,
      "learning_rate": 2.759441581416814e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 910090.0,
      "step": 2085
    },
    {
      "epoch": 0.16044919621567572,
      "grad_norm": 6.78125,
      "learning_rate": 2.7593262056764863e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 910390.0,
      "step": 2086
    },
    {
      "epoch": 0.16052611337589417,
      "grad_norm": 4.21875,
      "learning_rate": 2.759210829936159e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 910688.0,
      "step": 2087
    },
    {
      "epoch": 0.1606030305361126,
      "grad_norm": 6.78125,
      "learning_rate": 2.7590954541958313e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 911006.0,
      "step": 2088
    },
    {
      "epoch": 0.16067994769633104,
      "grad_norm": 3.828125,
      "learning_rate": 2.7589800784555035e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 911582.0,
      "step": 2089
    },
    {
      "epoch": 0.1607568648565495,
      "grad_norm": 3.796875,
      "learning_rate": 2.7588647027151758e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 911906.0,
      "step": 2090
    },
    {
      "epoch": 0.16083378201676793,
      "grad_norm": 6.1875,
      "learning_rate": 2.758749326974848e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 912220.0,
      "step": 2091
    },
    {
      "epoch": 0.1609106991769864,
      "grad_norm": 4.53125,
      "learning_rate": 2.7586339512345203e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 912538.0,
      "step": 2092
    },
    {
      "epoch": 0.16098761633720482,
      "grad_norm": 6.0625,
      "learning_rate": 2.758518575494193e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9895833134651184,
      "num_tokens": 912828.0,
      "step": 2093
    },
    {
      "epoch": 0.16106453349742328,
      "grad_norm": 4.125,
      "learning_rate": 2.7584031997538652e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 913372.0,
      "step": 2094
    },
    {
      "epoch": 0.1611414506576417,
      "grad_norm": 4.8125,
      "learning_rate": 2.7582878240135375e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 913944.0,
      "step": 2095
    },
    {
      "epoch": 0.16121836781786017,
      "grad_norm": 7.4375,
      "learning_rate": 2.7581724482732098e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 914310.0,
      "step": 2096
    },
    {
      "epoch": 0.1612952849780786,
      "grad_norm": 4.125,
      "learning_rate": 2.758057072532882e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 914884.0,
      "step": 2097
    },
    {
      "epoch": 0.16137220213829706,
      "grad_norm": 8.375,
      "learning_rate": 2.7579416967925543e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 915166.0,
      "step": 2098
    },
    {
      "epoch": 0.1614491192985155,
      "grad_norm": 3.015625,
      "learning_rate": 2.757826321052227e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 915712.0,
      "step": 2099
    },
    {
      "epoch": 0.16152603645873395,
      "grad_norm": 2.015625,
      "learning_rate": 2.7577109453118992e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 916276.0,
      "step": 2100
    },
    {
      "epoch": 0.16160295361895238,
      "grad_norm": 4.0,
      "learning_rate": 2.7575955695715715e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9915966391563416,
      "num_tokens": 916872.0,
      "step": 2101
    },
    {
      "epoch": 0.16167987077917084,
      "grad_norm": 6.34375,
      "learning_rate": 2.7574801938312437e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 917158.0,
      "step": 2102
    },
    {
      "epoch": 0.16175678793938927,
      "grad_norm": 6.3125,
      "learning_rate": 2.757364818090916e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 917464.0,
      "step": 2103
    },
    {
      "epoch": 0.16183370509960773,
      "grad_norm": 4.9375,
      "learning_rate": 2.7572494423505887e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9943661689758301,
      "num_tokens": 917820.0,
      "step": 2104
    },
    {
      "epoch": 0.16191062225982616,
      "grad_norm": 4.09375,
      "learning_rate": 2.757134066610261e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 918142.0,
      "step": 2105
    },
    {
      "epoch": 0.16198753942004462,
      "grad_norm": 5.09375,
      "learning_rate": 2.757018690869933e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 918436.0,
      "step": 2106
    },
    {
      "epoch": 0.16206445658026306,
      "grad_norm": 2.4375,
      "learning_rate": 2.7569033151296054e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 919260.0,
      "step": 2107
    },
    {
      "epoch": 0.16214137374048151,
      "grad_norm": 8.25,
      "learning_rate": 2.7567879393892777e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 919566.0,
      "step": 2108
    },
    {
      "epoch": 0.16221829090069995,
      "grad_norm": 6.15625,
      "learning_rate": 2.75667256364895e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 919902.0,
      "step": 2109
    },
    {
      "epoch": 0.16229520806091838,
      "grad_norm": 2.953125,
      "learning_rate": 2.7565571879086226e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 920454.0,
      "step": 2110
    },
    {
      "epoch": 0.16237212522113684,
      "grad_norm": 2.25,
      "learning_rate": 2.756441812168295e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 921274.0,
      "step": 2111
    },
    {
      "epoch": 0.16244904238135527,
      "grad_norm": 4.5625,
      "learning_rate": 2.756326436427967e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 921618.0,
      "step": 2112
    },
    {
      "epoch": 0.16252595954157373,
      "grad_norm": 3.546875,
      "learning_rate": 2.7562110606876394e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 921940.0,
      "step": 2113
    },
    {
      "epoch": 0.16260287670179216,
      "grad_norm": 2.109375,
      "learning_rate": 2.7560956849473117e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 922726.0,
      "step": 2114
    },
    {
      "epoch": 0.16267979386201062,
      "grad_norm": 10.3125,
      "learning_rate": 2.7559803092069844e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 922984.0,
      "step": 2115
    },
    {
      "epoch": 0.16275671102222905,
      "grad_norm": 2.015625,
      "learning_rate": 2.7558649334666566e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 923770.0,
      "step": 2116
    },
    {
      "epoch": 0.1628336281824475,
      "grad_norm": 2.90625,
      "learning_rate": 2.755749557726329e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 924354.0,
      "step": 2117
    },
    {
      "epoch": 0.16291054534266594,
      "grad_norm": 8.0,
      "learning_rate": 2.755634181986001e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 924640.0,
      "step": 2118
    },
    {
      "epoch": 0.1629874625028844,
      "grad_norm": 4.78125,
      "learning_rate": 2.7555188062456734e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 924950.0,
      "step": 2119
    },
    {
      "epoch": 0.16306437966310283,
      "grad_norm": 4.40625,
      "learning_rate": 2.7554034305053456e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 925296.0,
      "step": 2120
    },
    {
      "epoch": 0.1631412968233213,
      "grad_norm": 9.0625,
      "learning_rate": 2.7552880547650183e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 925574.0,
      "step": 2121
    },
    {
      "epoch": 0.16321821398353972,
      "grad_norm": 7.75,
      "learning_rate": 2.7551726790246906e-06,
      "loss": 0.0371,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 925870.0,
      "step": 2122
    },
    {
      "epoch": 0.16329513114375818,
      "grad_norm": 3.359375,
      "learning_rate": 2.755057303284363e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 926450.0,
      "step": 2123
    },
    {
      "epoch": 0.1633720483039766,
      "grad_norm": 8.125,
      "learning_rate": 2.754941927544035e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 926736.0,
      "step": 2124
    },
    {
      "epoch": 0.16344896546419507,
      "grad_norm": 2.046875,
      "learning_rate": 2.7548265518037074e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 927516.0,
      "step": 2125
    },
    {
      "epoch": 0.1635258826244135,
      "grad_norm": 2.90625,
      "learning_rate": 2.75471117606338e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 928062.0,
      "step": 2126
    },
    {
      "epoch": 0.16360279978463196,
      "grad_norm": 1.7578125,
      "learning_rate": 2.7545958003230523e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 928632.0,
      "step": 2127
    },
    {
      "epoch": 0.1636797169448504,
      "grad_norm": 5.6875,
      "learning_rate": 2.754480424582724e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 928920.0,
      "step": 2128
    },
    {
      "epoch": 0.16375663410506885,
      "grad_norm": 3.703125,
      "learning_rate": 2.754365048842397e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 929500.0,
      "step": 2129
    },
    {
      "epoch": 0.16383355126528729,
      "grad_norm": 7.375,
      "learning_rate": 2.754249673102069e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 929796.0,
      "step": 2130
    },
    {
      "epoch": 0.16391046842550572,
      "grad_norm": 3.75,
      "learning_rate": 2.7541342973617413e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 930312.0,
      "step": 2131
    },
    {
      "epoch": 0.16398738558572418,
      "grad_norm": 2.140625,
      "learning_rate": 2.754018921621414e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 931096.0,
      "step": 2132
    },
    {
      "epoch": 0.1640643027459426,
      "grad_norm": 4.625,
      "learning_rate": 2.7539035458810863e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 931418.0,
      "step": 2133
    },
    {
      "epoch": 0.16414121990616107,
      "grad_norm": 7.0,
      "learning_rate": 2.7537881701407585e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 931766.0,
      "step": 2134
    },
    {
      "epoch": 0.1642181370663795,
      "grad_norm": 3.234375,
      "learning_rate": 2.7536727944004308e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 932342.0,
      "step": 2135
    },
    {
      "epoch": 0.16429505422659796,
      "grad_norm": 4.0,
      "learning_rate": 2.753557418660103e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 932836.0,
      "step": 2136
    },
    {
      "epoch": 0.1643719713868164,
      "grad_norm": 6.875,
      "learning_rate": 2.7534420429197757e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 933152.0,
      "step": 2137
    },
    {
      "epoch": 0.16444888854703485,
      "grad_norm": 3.65625,
      "learning_rate": 2.753326667179448e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 933724.0,
      "step": 2138
    },
    {
      "epoch": 0.16452580570725328,
      "grad_norm": 5.90625,
      "learning_rate": 2.75321129143912e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 934030.0,
      "step": 2139
    },
    {
      "epoch": 0.16460272286747174,
      "grad_norm": 4.6875,
      "learning_rate": 2.7530959156987925e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 934356.0,
      "step": 2140
    },
    {
      "epoch": 0.16467964002769017,
      "grad_norm": 5.90625,
      "learning_rate": 2.7529805399584648e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 934688.0,
      "step": 2141
    },
    {
      "epoch": 0.16475655718790863,
      "grad_norm": 3.453125,
      "learning_rate": 2.752865164218137e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 935206.0,
      "step": 2142
    },
    {
      "epoch": 0.16483347434812706,
      "grad_norm": 7.625,
      "learning_rate": 2.7527497884778097e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 935540.0,
      "step": 2143
    },
    {
      "epoch": 0.16491039150834552,
      "grad_norm": 3.5625,
      "learning_rate": 2.752634412737482e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 936096.0,
      "step": 2144
    },
    {
      "epoch": 0.16498730866856395,
      "grad_norm": 4.21875,
      "learning_rate": 2.7525190369971542e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 936430.0,
      "step": 2145
    },
    {
      "epoch": 0.1650642258287824,
      "grad_norm": 4.34375,
      "learning_rate": 2.7524036612568265e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 936950.0,
      "step": 2146
    },
    {
      "epoch": 0.16514114298900084,
      "grad_norm": 3.765625,
      "learning_rate": 2.7522882855164987e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 937470.0,
      "step": 2147
    },
    {
      "epoch": 0.1652180601492193,
      "grad_norm": 4.28125,
      "learning_rate": 2.752172909776171e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 937796.0,
      "step": 2148
    },
    {
      "epoch": 0.16529497730943774,
      "grad_norm": 4.375,
      "learning_rate": 2.7520575340358437e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 938376.0,
      "step": 2149
    },
    {
      "epoch": 0.16537189446965617,
      "grad_norm": 4.875,
      "learning_rate": 2.7519421582955155e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 938726.0,
      "step": 2150
    },
    {
      "epoch": 0.16544881162987463,
      "grad_norm": 2.34375,
      "learning_rate": 2.751826782555188e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 939272.0,
      "step": 2151
    },
    {
      "epoch": 0.16552572879009306,
      "grad_norm": 8.3125,
      "learning_rate": 2.7517114068148604e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 939572.0,
      "step": 2152
    },
    {
      "epoch": 0.16560264595031152,
      "grad_norm": 8.4375,
      "learning_rate": 2.7515960310745327e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 939832.0,
      "step": 2153
    },
    {
      "epoch": 0.16567956311052995,
      "grad_norm": 4.5625,
      "learning_rate": 2.7514806553342054e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 940126.0,
      "step": 2154
    },
    {
      "epoch": 0.1657564802707484,
      "grad_norm": 4.0,
      "learning_rate": 2.7513652795938776e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 940470.0,
      "step": 2155
    },
    {
      "epoch": 0.16583339743096684,
      "grad_norm": 6.90625,
      "learning_rate": 2.7512499038535495e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 940776.0,
      "step": 2156
    },
    {
      "epoch": 0.1659103145911853,
      "grad_norm": 2.734375,
      "learning_rate": 2.751134528113222e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9971292018890381,
      "num_tokens": 941824.0,
      "step": 2157
    },
    {
      "epoch": 0.16598723175140373,
      "grad_norm": 5.90625,
      "learning_rate": 2.7510191523728944e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 942102.0,
      "step": 2158
    },
    {
      "epoch": 0.1660641489116222,
      "grad_norm": 4.0625,
      "learning_rate": 2.7509037766325667e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 942702.0,
      "step": 2159
    },
    {
      "epoch": 0.16614106607184062,
      "grad_norm": 3.34375,
      "learning_rate": 2.7507884008922394e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 943462.0,
      "step": 2160
    },
    {
      "epoch": 0.16621798323205908,
      "grad_norm": 2.3125,
      "learning_rate": 2.750673025151911e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 944008.0,
      "step": 2161
    },
    {
      "epoch": 0.1662949003922775,
      "grad_norm": 5.1875,
      "learning_rate": 2.750557649411584e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 944314.0,
      "step": 2162
    },
    {
      "epoch": 0.16637181755249597,
      "grad_norm": 6.46875,
      "learning_rate": 2.750442273671256e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 944620.0,
      "step": 2163
    },
    {
      "epoch": 0.1664487347127144,
      "grad_norm": 7.5,
      "learning_rate": 2.7503268979309284e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 944976.0,
      "step": 2164
    },
    {
      "epoch": 0.16652565187293286,
      "grad_norm": 5.5,
      "learning_rate": 2.750211522190601e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 945258.0,
      "step": 2165
    },
    {
      "epoch": 0.1666025690331513,
      "grad_norm": 6.09375,
      "learning_rate": 2.750096146450273e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 945588.0,
      "step": 2166
    },
    {
      "epoch": 0.16667948619336975,
      "grad_norm": 5.34375,
      "learning_rate": 2.749980770709945e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 945910.0,
      "step": 2167
    },
    {
      "epoch": 0.16675640335358818,
      "grad_norm": 4.375,
      "learning_rate": 2.749865394969618e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 946436.0,
      "step": 2168
    },
    {
      "epoch": 0.16683332051380664,
      "grad_norm": 5.59375,
      "learning_rate": 2.74975001922929e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 946752.0,
      "step": 2169
    },
    {
      "epoch": 0.16691023767402507,
      "grad_norm": 7.6875,
      "learning_rate": 2.7496346434889624e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 947068.0,
      "step": 2170
    },
    {
      "epoch": 0.1669871548342435,
      "grad_norm": 4.90625,
      "learning_rate": 2.749519267748635e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 947418.0,
      "step": 2171
    },
    {
      "epoch": 0.16706407199446197,
      "grad_norm": 7.71875,
      "learning_rate": 2.749403892008307e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 947674.0,
      "step": 2172
    },
    {
      "epoch": 0.1671409891546804,
      "grad_norm": 4.1875,
      "learning_rate": 2.7492885162679796e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 947992.0,
      "step": 2173
    },
    {
      "epoch": 0.16721790631489886,
      "grad_norm": 4.46875,
      "learning_rate": 2.749173140527652e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 948312.0,
      "step": 2174
    },
    {
      "epoch": 0.1672948234751173,
      "grad_norm": 4.65625,
      "learning_rate": 2.749057764787324e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 948658.0,
      "step": 2175
    },
    {
      "epoch": 0.16737174063533575,
      "grad_norm": 7.03125,
      "learning_rate": 2.7489423890469968e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 948986.0,
      "step": 2176
    },
    {
      "epoch": 0.16744865779555418,
      "grad_norm": 3.125,
      "learning_rate": 2.7488270133066686e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 949806.0,
      "step": 2177
    },
    {
      "epoch": 0.16752557495577264,
      "grad_norm": 2.9375,
      "learning_rate": 2.748711637566341e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 950358.0,
      "step": 2178
    },
    {
      "epoch": 0.16760249211599107,
      "grad_norm": 5.03125,
      "learning_rate": 2.7485962618260135e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 950854.0,
      "step": 2179
    },
    {
      "epoch": 0.16767940927620953,
      "grad_norm": 7.15625,
      "learning_rate": 2.748480886085686e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 951188.0,
      "step": 2180
    },
    {
      "epoch": 0.16775632643642796,
      "grad_norm": 14.0625,
      "learning_rate": 2.748365510345358e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 951422.0,
      "step": 2181
    },
    {
      "epoch": 0.16783324359664642,
      "grad_norm": 3.28125,
      "learning_rate": 2.7482501346050307e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 951970.0,
      "step": 2182
    },
    {
      "epoch": 0.16791016075686485,
      "grad_norm": 6.65625,
      "learning_rate": 2.7481347588647026e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 952288.0,
      "step": 2183
    },
    {
      "epoch": 0.1679870779170833,
      "grad_norm": 4.65625,
      "learning_rate": 2.7480193831243752e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 952632.0,
      "step": 2184
    },
    {
      "epoch": 0.16806399507730174,
      "grad_norm": 7.375,
      "learning_rate": 2.7479040073840475e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 952936.0,
      "step": 2185
    },
    {
      "epoch": 0.1681409122375202,
      "grad_norm": 2.71875,
      "learning_rate": 2.7477886316437198e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 953760.0,
      "step": 2186
    },
    {
      "epoch": 0.16821782939773863,
      "grad_norm": 2.765625,
      "learning_rate": 2.7476732559033924e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 954334.0,
      "step": 2187
    },
    {
      "epoch": 0.1682947465579571,
      "grad_norm": 3.921875,
      "learning_rate": 2.7475578801630643e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 954908.0,
      "step": 2188
    },
    {
      "epoch": 0.16837166371817552,
      "grad_norm": 5.5,
      "learning_rate": 2.7474425044227365e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 955214.0,
      "step": 2189
    },
    {
      "epoch": 0.16844858087839398,
      "grad_norm": 4.75,
      "learning_rate": 2.7473271286824092e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 955526.0,
      "step": 2190
    },
    {
      "epoch": 0.16852549803861241,
      "grad_norm": 5.59375,
      "learning_rate": 2.7472117529420815e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 955878.0,
      "step": 2191
    },
    {
      "epoch": 0.16860241519883085,
      "grad_norm": 9.8125,
      "learning_rate": 2.7470963772017537e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9927007555961609,
      "num_tokens": 956154.0,
      "step": 2192
    },
    {
      "epoch": 0.1686793323590493,
      "grad_norm": 6.8125,
      "learning_rate": 2.7469810014614264e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 956444.0,
      "step": 2193
    },
    {
      "epoch": 0.16875624951926774,
      "grad_norm": 3.796875,
      "learning_rate": 2.7468656257210982e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.990439772605896,
      "num_tokens": 956968.0,
      "step": 2194
    },
    {
      "epoch": 0.1688331666794862,
      "grad_norm": 4.28125,
      "learning_rate": 2.746750249980771e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 957280.0,
      "step": 2195
    },
    {
      "epoch": 0.16891008383970463,
      "grad_norm": 5.3125,
      "learning_rate": 2.746634874240443e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 957574.0,
      "step": 2196
    },
    {
      "epoch": 0.1689870009999231,
      "grad_norm": 2.890625,
      "learning_rate": 2.7465194985001154e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 958142.0,
      "step": 2197
    },
    {
      "epoch": 0.16906391816014152,
      "grad_norm": 8.0625,
      "learning_rate": 2.7464041227597877e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 958474.0,
      "step": 2198
    },
    {
      "epoch": 0.16914083532035998,
      "grad_norm": 4.46875,
      "learning_rate": 2.74628874701946e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 958780.0,
      "step": 2199
    },
    {
      "epoch": 0.1692177524805784,
      "grad_norm": 3.34375,
      "learning_rate": 2.7461733712791322e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 959338.0,
      "step": 2200
    },
    {
      "epoch": 0.16929466964079687,
      "grad_norm": 10.375,
      "learning_rate": 2.746057995538805e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 959686.0,
      "step": 2201
    },
    {
      "epoch": 0.1693715868010153,
      "grad_norm": 3.0,
      "learning_rate": 2.745942619798477e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 960204.0,
      "step": 2202
    },
    {
      "epoch": 0.16944850396123376,
      "grad_norm": 3.875,
      "learning_rate": 2.7458272440581494e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 960538.0,
      "step": 2203
    },
    {
      "epoch": 0.1695254211214522,
      "grad_norm": 3.328125,
      "learning_rate": 2.7457118683178217e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 961072.0,
      "step": 2204
    },
    {
      "epoch": 0.16960233828167065,
      "grad_norm": 5.34375,
      "learning_rate": 2.745596492577494e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 961372.0,
      "step": 2205
    },
    {
      "epoch": 0.16967925544188908,
      "grad_norm": 7.25,
      "learning_rate": 2.745481116837166e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 961708.0,
      "step": 2206
    },
    {
      "epoch": 0.16975617260210754,
      "grad_norm": 2.453125,
      "learning_rate": 2.745365741096839e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 962258.0,
      "step": 2207
    },
    {
      "epoch": 0.16983308976232597,
      "grad_norm": 6.03125,
      "learning_rate": 2.745250365356511e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 962586.0,
      "step": 2208
    },
    {
      "epoch": 0.16991000692254443,
      "grad_norm": 2.96875,
      "learning_rate": 2.7451349896161834e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 963148.0,
      "step": 2209
    },
    {
      "epoch": 0.16998692408276286,
      "grad_norm": 3.296875,
      "learning_rate": 2.7450196138758557e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 963712.0,
      "step": 2210
    },
    {
      "epoch": 0.17006384124298132,
      "grad_norm": 2.734375,
      "learning_rate": 2.744904238135528e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 964230.0,
      "step": 2211
    },
    {
      "epoch": 0.17014075840319975,
      "grad_norm": 8.75,
      "learning_rate": 2.7447888623952006e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 964484.0,
      "step": 2212
    },
    {
      "epoch": 0.17021767556341819,
      "grad_norm": 6.46875,
      "learning_rate": 2.744673486654873e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 964850.0,
      "step": 2213
    },
    {
      "epoch": 0.17029459272363665,
      "grad_norm": 2.421875,
      "learning_rate": 2.744558110914545e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 965406.0,
      "step": 2214
    },
    {
      "epoch": 0.17037150988385508,
      "grad_norm": 4.03125,
      "learning_rate": 2.7444427351742174e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 965982.0,
      "step": 2215
    },
    {
      "epoch": 0.17044842704407354,
      "grad_norm": 2.625,
      "learning_rate": 2.7443273594338896e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 966532.0,
      "step": 2216
    },
    {
      "epoch": 0.17052534420429197,
      "grad_norm": 4.90625,
      "learning_rate": 2.744211983693562e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 966856.0,
      "step": 2217
    },
    {
      "epoch": 0.17060226136451043,
      "grad_norm": 5.0625,
      "learning_rate": 2.7440966079532346e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 967150.0,
      "step": 2218
    },
    {
      "epoch": 0.17067917852472886,
      "grad_norm": 2.171875,
      "learning_rate": 2.743981232212907e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 967974.0,
      "step": 2219
    },
    {
      "epoch": 0.17075609568494732,
      "grad_norm": 7.71875,
      "learning_rate": 2.743865856472579e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 968282.0,
      "step": 2220
    },
    {
      "epoch": 0.17083301284516575,
      "grad_norm": 4.71875,
      "learning_rate": 2.7437504807322513e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 968600.0,
      "step": 2221
    },
    {
      "epoch": 0.1709099300053842,
      "grad_norm": 9.4375,
      "learning_rate": 2.7436351049919236e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 968860.0,
      "step": 2222
    },
    {
      "epoch": 0.17098684716560264,
      "grad_norm": 4.125,
      "learning_rate": 2.7435197292515963e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 969424.0,
      "step": 2223
    },
    {
      "epoch": 0.1710637643258211,
      "grad_norm": 2.21875,
      "learning_rate": 2.7434043535112685e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 970208.0,
      "step": 2224
    },
    {
      "epoch": 0.17114068148603953,
      "grad_norm": 5.53125,
      "learning_rate": 2.743288977770941e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9945054650306702,
      "num_tokens": 970574.0,
      "step": 2225
    },
    {
      "epoch": 0.171217598646258,
      "grad_norm": 3.65625,
      "learning_rate": 2.743173602030613e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9889094233512878,
      "num_tokens": 971116.0,
      "step": 2226
    },
    {
      "epoch": 0.17129451580647642,
      "grad_norm": 7.59375,
      "learning_rate": 2.7430582262902853e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9941520690917969,
      "num_tokens": 971460.0,
      "step": 2227
    },
    {
      "epoch": 0.17137143296669488,
      "grad_norm": 7.28125,
      "learning_rate": 2.7429428505499576e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 971790.0,
      "step": 2228
    },
    {
      "epoch": 0.1714483501269133,
      "grad_norm": 3.171875,
      "learning_rate": 2.7428274748096302e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 972308.0,
      "step": 2229
    },
    {
      "epoch": 0.17152526728713177,
      "grad_norm": 4.40625,
      "learning_rate": 2.7427120990693025e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 972602.0,
      "step": 2230
    },
    {
      "epoch": 0.1716021844473502,
      "grad_norm": 8.0625,
      "learning_rate": 2.7425967233289748e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 972890.0,
      "step": 2231
    },
    {
      "epoch": 0.17167910160756863,
      "grad_norm": 5.125,
      "learning_rate": 2.742481347588647e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 973196.0,
      "step": 2232
    },
    {
      "epoch": 0.1717560187677871,
      "grad_norm": 7.0625,
      "learning_rate": 2.7423659718483193e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 973558.0,
      "step": 2233
    },
    {
      "epoch": 0.17183293592800553,
      "grad_norm": 2.5625,
      "learning_rate": 2.742250596107992e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 974602.0,
      "step": 2234
    },
    {
      "epoch": 0.17190985308822399,
      "grad_norm": 8.5625,
      "learning_rate": 2.7421352203676642e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 974914.0,
      "step": 2235
    },
    {
      "epoch": 0.17198677024844242,
      "grad_norm": 3.09375,
      "learning_rate": 2.7420198446273365e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 975718.0,
      "step": 2236
    },
    {
      "epoch": 0.17206368740866088,
      "grad_norm": 4.53125,
      "learning_rate": 2.7419044688870087e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 976028.0,
      "step": 2237
    },
    {
      "epoch": 0.1721406045688793,
      "grad_norm": 2.84375,
      "learning_rate": 2.741789093146681e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 976586.0,
      "step": 2238
    },
    {
      "epoch": 0.17221752172909777,
      "grad_norm": 2.78125,
      "learning_rate": 2.7416737174063533e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 977120.0,
      "step": 2239
    },
    {
      "epoch": 0.1722944388893162,
      "grad_norm": 3.84375,
      "learning_rate": 2.741558341666026e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 977652.0,
      "step": 2240
    },
    {
      "epoch": 0.17237135604953466,
      "grad_norm": 4.21875,
      "learning_rate": 2.741442965925698e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 977972.0,
      "step": 2241
    },
    {
      "epoch": 0.1724482732097531,
      "grad_norm": 2.90625,
      "learning_rate": 2.7413275901853705e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 978534.0,
      "step": 2242
    },
    {
      "epoch": 0.17252519036997155,
      "grad_norm": 4.5625,
      "learning_rate": 2.7412122144450427e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 978858.0,
      "step": 2243
    },
    {
      "epoch": 0.17260210753018998,
      "grad_norm": 3.359375,
      "learning_rate": 2.741096838704715e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.990328848361969,
      "num_tokens": 979376.0,
      "step": 2244
    },
    {
      "epoch": 0.17267902469040844,
      "grad_norm": 4.5625,
      "learning_rate": 2.7409814629643876e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 979870.0,
      "step": 2245
    },
    {
      "epoch": 0.17275594185062687,
      "grad_norm": 5.625,
      "learning_rate": 2.74086608722406e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 980176.0,
      "step": 2246
    },
    {
      "epoch": 0.17283285901084533,
      "grad_norm": 4.53125,
      "learning_rate": 2.740750711483732e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 980520.0,
      "step": 2247
    },
    {
      "epoch": 0.17290977617106376,
      "grad_norm": 4.625,
      "learning_rate": 2.7406353357434044e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9894179701805115,
      "num_tokens": 981088.0,
      "step": 2248
    },
    {
      "epoch": 0.17298669333128222,
      "grad_norm": 2.5625,
      "learning_rate": 2.7405199600030767e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 981868.0,
      "step": 2249
    },
    {
      "epoch": 0.17306361049150065,
      "grad_norm": 3.78125,
      "learning_rate": 2.740404584262749e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 982464.0,
      "step": 2250
    },
    {
      "epoch": 0.1731405276517191,
      "grad_norm": 4.03125,
      "learning_rate": 2.7402892085224216e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 983066.0,
      "step": 2251
    },
    {
      "epoch": 0.17321744481193754,
      "grad_norm": 4.0625,
      "learning_rate": 2.740173832782094e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 983386.0,
      "step": 2252
    },
    {
      "epoch": 0.17329436197215597,
      "grad_norm": 2.609375,
      "learning_rate": 2.740058457041766e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 984194.0,
      "step": 2253
    },
    {
      "epoch": 0.17337127913237443,
      "grad_norm": 2.296875,
      "learning_rate": 2.7399430813014384e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 984752.0,
      "step": 2254
    },
    {
      "epoch": 0.17344819629259287,
      "grad_norm": 10.25,
      "learning_rate": 2.7398277055611107e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 985006.0,
      "step": 2255
    },
    {
      "epoch": 0.17352511345281132,
      "grad_norm": 8.625,
      "learning_rate": 2.7397123298207833e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 985310.0,
      "step": 2256
    },
    {
      "epoch": 0.17360203061302976,
      "grad_norm": 7.0625,
      "learning_rate": 2.7395969540804556e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9972899556159973,
      "num_tokens": 985682.0,
      "step": 2257
    },
    {
      "epoch": 0.17367894777324822,
      "grad_norm": 4.75,
      "learning_rate": 2.739481578340128e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 986208.0,
      "step": 2258
    },
    {
      "epoch": 0.17375586493346665,
      "grad_norm": 3.359375,
      "learning_rate": 2.7393662025998e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 986782.0,
      "step": 2259
    },
    {
      "epoch": 0.1738327820936851,
      "grad_norm": 8.5625,
      "learning_rate": 2.7392508268594724e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9841772317886353,
      "num_tokens": 987100.0,
      "step": 2260
    },
    {
      "epoch": 0.17390969925390354,
      "grad_norm": 2.34375,
      "learning_rate": 2.7391354511191446e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 987686.0,
      "step": 2261
    },
    {
      "epoch": 0.173986616414122,
      "grad_norm": 10.625,
      "learning_rate": 2.7390200753788173e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 987978.0,
      "step": 2262
    },
    {
      "epoch": 0.17406353357434043,
      "grad_norm": 4.1875,
      "learning_rate": 2.7389046996384896e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 988304.0,
      "step": 2263
    },
    {
      "epoch": 0.1741404507345589,
      "grad_norm": 5.375,
      "learning_rate": 2.738789323898162e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 988622.0,
      "step": 2264
    },
    {
      "epoch": 0.17421736789477732,
      "grad_norm": 5.8125,
      "learning_rate": 2.738673948157834e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 988906.0,
      "step": 2265
    },
    {
      "epoch": 0.17429428505499578,
      "grad_norm": 5.40625,
      "learning_rate": 2.7385585724175063e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 989204.0,
      "step": 2266
    },
    {
      "epoch": 0.1743712022152142,
      "grad_norm": 7.5625,
      "learning_rate": 2.7384431966771786e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 989482.0,
      "step": 2267
    },
    {
      "epoch": 0.17444811937543267,
      "grad_norm": 3.375,
      "learning_rate": 2.7383278209368513e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 989806.0,
      "step": 2268
    },
    {
      "epoch": 0.1745250365356511,
      "grad_norm": 4.34375,
      "learning_rate": 2.738212445196523e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 990326.0,
      "step": 2269
    },
    {
      "epoch": 0.17460195369586956,
      "grad_norm": 3.03125,
      "learning_rate": 2.738097069456196e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 990910.0,
      "step": 2270
    },
    {
      "epoch": 0.174678870856088,
      "grad_norm": 4.09375,
      "learning_rate": 2.737981693715868e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 991236.0,
      "step": 2271
    },
    {
      "epoch": 0.17475578801630645,
      "grad_norm": 3.71875,
      "learning_rate": 2.7378663179755403e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 991780.0,
      "step": 2272
    },
    {
      "epoch": 0.17483270517652488,
      "grad_norm": 6.21875,
      "learning_rate": 2.737750942235213e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 992142.0,
      "step": 2273
    },
    {
      "epoch": 0.17490962233674331,
      "grad_norm": 4.375,
      "learning_rate": 2.7376355664948853e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 992744.0,
      "step": 2274
    },
    {
      "epoch": 0.17498653949696177,
      "grad_norm": 4.3125,
      "learning_rate": 2.737520190754557e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 993038.0,
      "step": 2275
    },
    {
      "epoch": 0.1750634566571802,
      "grad_norm": 3.890625,
      "learning_rate": 2.7374048150142298e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 993638.0,
      "step": 2276
    },
    {
      "epoch": 0.17514037381739866,
      "grad_norm": 4.46875,
      "learning_rate": 2.737289439273902e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 993982.0,
      "step": 2277
    },
    {
      "epoch": 0.1752172909776171,
      "grad_norm": 3.15625,
      "learning_rate": 2.7371740635335743e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 994538.0,
      "step": 2278
    },
    {
      "epoch": 0.17529420813783556,
      "grad_norm": 3.890625,
      "learning_rate": 2.737058687793247e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 995080.0,
      "step": 2279
    },
    {
      "epoch": 0.175371125298054,
      "grad_norm": 5.0625,
      "learning_rate": 2.736943312052919e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 995424.0,
      "step": 2280
    },
    {
      "epoch": 0.17544804245827245,
      "grad_norm": 7.78125,
      "learning_rate": 2.7368279363125915e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 995730.0,
      "step": 2281
    },
    {
      "epoch": 0.17552495961849088,
      "grad_norm": 9.9375,
      "learning_rate": 2.7367125605722637e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 996016.0,
      "step": 2282
    },
    {
      "epoch": 0.17560187677870934,
      "grad_norm": 4.84375,
      "learning_rate": 2.736597184831936e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 996350.0,
      "step": 2283
    },
    {
      "epoch": 0.17567879393892777,
      "grad_norm": 4.875,
      "learning_rate": 2.7364818090916087e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 996712.0,
      "step": 2284
    },
    {
      "epoch": 0.17575571109914623,
      "grad_norm": 7.625,
      "learning_rate": 2.736366433351281e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 997016.0,
      "step": 2285
    },
    {
      "epoch": 0.17583262825936466,
      "grad_norm": 5.25,
      "learning_rate": 2.7362510576109528e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 997542.0,
      "step": 2286
    },
    {
      "epoch": 0.17590954541958312,
      "grad_norm": 7.40625,
      "learning_rate": 2.7361356818706255e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 997824.0,
      "step": 2287
    },
    {
      "epoch": 0.17598646257980155,
      "grad_norm": 2.53125,
      "learning_rate": 2.7360203061302977e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 998386.0,
      "step": 2288
    },
    {
      "epoch": 0.17606337974002,
      "grad_norm": 3.90625,
      "learning_rate": 2.73590493038997e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 998708.0,
      "step": 2289
    },
    {
      "epoch": 0.17614029690023844,
      "grad_norm": 4.03125,
      "learning_rate": 2.7357895546496427e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 999002.0,
      "step": 2290
    },
    {
      "epoch": 0.1762172140604569,
      "grad_norm": 3.375,
      "learning_rate": 2.7356741789093145e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 999570.0,
      "step": 2291
    },
    {
      "epoch": 0.17629413122067533,
      "grad_norm": 4.375,
      "learning_rate": 2.735558803168987e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 999890.0,
      "step": 2292
    },
    {
      "epoch": 0.1763710483808938,
      "grad_norm": 11.5625,
      "learning_rate": 2.7354434274286594e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 1000154.0,
      "step": 2293
    },
    {
      "epoch": 0.17644796554111222,
      "grad_norm": 5.625,
      "learning_rate": 2.7353280516883317e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1000482.0,
      "step": 2294
    },
    {
      "epoch": 0.17652488270133065,
      "grad_norm": 6.53125,
      "learning_rate": 2.7352126759480044e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 1000792.0,
      "step": 2295
    },
    {
      "epoch": 0.1766017998615491,
      "grad_norm": 2.71875,
      "learning_rate": 2.7350973002076766e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9971292018890381,
      "num_tokens": 1001840.0,
      "step": 2296
    },
    {
      "epoch": 0.17667871702176755,
      "grad_norm": 4.40625,
      "learning_rate": 2.7349819244673485e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1002148.0,
      "step": 2297
    },
    {
      "epoch": 0.176755634181986,
      "grad_norm": 3.46875,
      "learning_rate": 2.734866548727021e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1002668.0,
      "step": 2298
    },
    {
      "epoch": 0.17683255134220444,
      "grad_norm": 3.65625,
      "learning_rate": 2.7347511729866934e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1003216.0,
      "step": 2299
    },
    {
      "epoch": 0.1769094685024229,
      "grad_norm": 9.5,
      "learning_rate": 2.7346357972463657e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 1003518.0,
      "step": 2300
    },
    {
      "epoch": 0.17698638566264133,
      "grad_norm": 4.53125,
      "learning_rate": 2.7345204215060383e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1003812.0,
      "step": 2301
    },
    {
      "epoch": 0.17706330282285979,
      "grad_norm": 7.8125,
      "learning_rate": 2.73440504576571e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1004150.0,
      "step": 2302
    },
    {
      "epoch": 0.17714021998307822,
      "grad_norm": 7.78125,
      "learning_rate": 2.734289670025383e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1004482.0,
      "step": 2303
    },
    {
      "epoch": 0.17721713714329668,
      "grad_norm": 3.984375,
      "learning_rate": 2.734174294285055e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1004794.0,
      "step": 2304
    },
    {
      "epoch": 0.1772940543035151,
      "grad_norm": 4.21875,
      "learning_rate": 2.7340589185447274e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 1005076.0,
      "step": 2305
    },
    {
      "epoch": 0.17737097146373357,
      "grad_norm": 7.75,
      "learning_rate": 2.7339435428044e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1005438.0,
      "step": 2306
    },
    {
      "epoch": 0.177447888623952,
      "grad_norm": 4.71875,
      "learning_rate": 2.733828167064072e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1005978.0,
      "step": 2307
    },
    {
      "epoch": 0.17752480578417046,
      "grad_norm": 5.46875,
      "learning_rate": 2.733712791323744e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1006300.0,
      "step": 2308
    },
    {
      "epoch": 0.1776017229443889,
      "grad_norm": 2.75,
      "learning_rate": 2.733597415583417e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1007086.0,
      "step": 2309
    },
    {
      "epoch": 0.17767864010460735,
      "grad_norm": 5.96875,
      "learning_rate": 2.733482039843089e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1007386.0,
      "step": 2310
    },
    {
      "epoch": 0.17775555726482578,
      "grad_norm": 2.546875,
      "learning_rate": 2.7333666641027613e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1008172.0,
      "step": 2311
    },
    {
      "epoch": 0.17783247442504424,
      "grad_norm": 6.46875,
      "learning_rate": 2.733251288362434e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1008468.0,
      "step": 2312
    },
    {
      "epoch": 0.17790939158526267,
      "grad_norm": 5.34375,
      "learning_rate": 2.733135912622106e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 1008782.0,
      "step": 2313
    },
    {
      "epoch": 0.1779863087454811,
      "grad_norm": 8.8125,
      "learning_rate": 2.7330205368817785e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9919571280479431,
      "num_tokens": 1009158.0,
      "step": 2314
    },
    {
      "epoch": 0.17806322590569956,
      "grad_norm": 6.5625,
      "learning_rate": 2.732905161141451e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9966996908187866,
      "num_tokens": 1009462.0,
      "step": 2315
    },
    {
      "epoch": 0.178140143065918,
      "grad_norm": 5.125,
      "learning_rate": 2.732789785401123e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 1009786.0,
      "step": 2316
    },
    {
      "epoch": 0.17821706022613645,
      "grad_norm": 6.53125,
      "learning_rate": 2.7326744096607953e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1010082.0,
      "step": 2317
    },
    {
      "epoch": 0.17829397738635488,
      "grad_norm": 9.0625,
      "learning_rate": 2.7325590339204676e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1010400.0,
      "step": 2318
    },
    {
      "epoch": 0.17837089454657334,
      "grad_norm": 2.421875,
      "learning_rate": 2.73244365818014e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1011186.0,
      "step": 2319
    },
    {
      "epoch": 0.17844781170679178,
      "grad_norm": 3.21875,
      "learning_rate": 2.7323282824398125e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1011710.0,
      "step": 2320
    },
    {
      "epoch": 0.17852472886701024,
      "grad_norm": 2.96875,
      "learning_rate": 2.7322129066994848e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 1012526.0,
      "step": 2321
    },
    {
      "epoch": 0.17860164602722867,
      "grad_norm": 3.015625,
      "learning_rate": 2.732097530959157e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 1013342.0,
      "step": 2322
    },
    {
      "epoch": 0.17867856318744713,
      "grad_norm": 3.796875,
      "learning_rate": 2.7319821552188297e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1013916.0,
      "step": 2323
    },
    {
      "epoch": 0.17875548034766556,
      "grad_norm": 5.09375,
      "learning_rate": 2.7318667794785015e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1014260.0,
      "step": 2324
    },
    {
      "epoch": 0.17883239750788402,
      "grad_norm": 6.6875,
      "learning_rate": 2.731751403738174e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 1014556.0,
      "step": 2325
    },
    {
      "epoch": 0.17890931466810245,
      "grad_norm": 2.453125,
      "learning_rate": 2.7316360279978465e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 1015576.0,
      "step": 2326
    },
    {
      "epoch": 0.1789862318283209,
      "grad_norm": 8.125,
      "learning_rate": 2.7315206522575187e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 1015904.0,
      "step": 2327
    },
    {
      "epoch": 0.17906314898853934,
      "grad_norm": 8.875,
      "learning_rate": 2.731405276517191e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1016190.0,
      "step": 2328
    },
    {
      "epoch": 0.1791400661487578,
      "grad_norm": 7.6875,
      "learning_rate": 2.7312899007768633e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 1016532.0,
      "step": 2329
    },
    {
      "epoch": 0.17921698330897623,
      "grad_norm": 2.359375,
      "learning_rate": 2.7311745250365355e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1017082.0,
      "step": 2330
    },
    {
      "epoch": 0.1792939004691947,
      "grad_norm": 4.78125,
      "learning_rate": 2.731059149296208e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9948365092277527,
      "num_tokens": 1017664.0,
      "step": 2331
    },
    {
      "epoch": 0.17937081762941312,
      "grad_norm": 3.140625,
      "learning_rate": 2.7309437735558805e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 1018208.0,
      "step": 2332
    },
    {
      "epoch": 0.17944773478963158,
      "grad_norm": 1.9375,
      "learning_rate": 2.7308283978155527e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9974554777145386,
      "num_tokens": 1018996.0,
      "step": 2333
    },
    {
      "epoch": 0.17952465194985,
      "grad_norm": 2.71875,
      "learning_rate": 2.730713022075225e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1019582.0,
      "step": 2334
    },
    {
      "epoch": 0.17960156911006844,
      "grad_norm": 8.125,
      "learning_rate": 2.7305976463348972e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1019894.0,
      "step": 2335
    },
    {
      "epoch": 0.1796784862702869,
      "grad_norm": 5.125,
      "learning_rate": 2.7304822705945695e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1020244.0,
      "step": 2336
    },
    {
      "epoch": 0.17975540343050533,
      "grad_norm": 8.125,
      "learning_rate": 2.730366894854242e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.98525071144104,
      "num_tokens": 1020584.0,
      "step": 2337
    },
    {
      "epoch": 0.1798323205907238,
      "grad_norm": 2.875,
      "learning_rate": 2.7302515191139144e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1021130.0,
      "step": 2338
    },
    {
      "epoch": 0.17990923775094222,
      "grad_norm": 2.3125,
      "learning_rate": 2.7301361433735867e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1021912.0,
      "step": 2339
    },
    {
      "epoch": 0.17998615491116068,
      "grad_norm": 5.53125,
      "learning_rate": 2.730020767633259e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 1022238.0,
      "step": 2340
    },
    {
      "epoch": 0.18006307207137912,
      "grad_norm": 2.34375,
      "learning_rate": 2.729905391892931e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1022784.0,
      "step": 2341
    },
    {
      "epoch": 0.18013998923159757,
      "grad_norm": 5.6875,
      "learning_rate": 2.729790016152604e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1023094.0,
      "step": 2342
    },
    {
      "epoch": 0.180216906391816,
      "grad_norm": 2.609375,
      "learning_rate": 2.729674640412276e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1023656.0,
      "step": 2343
    },
    {
      "epoch": 0.18029382355203447,
      "grad_norm": 2.1875,
      "learning_rate": 2.7295592646719484e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 1024444.0,
      "step": 2344
    },
    {
      "epoch": 0.1803707407122529,
      "grad_norm": 6.5625,
      "learning_rate": 2.7294438889316207e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1024750.0,
      "step": 2345
    },
    {
      "epoch": 0.18044765787247136,
      "grad_norm": 3.078125,
      "learning_rate": 2.729328513191293e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1025298.0,
      "step": 2346
    },
    {
      "epoch": 0.1805245750326898,
      "grad_norm": 3.625,
      "learning_rate": 2.729213137450965e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1025814.0,
      "step": 2347
    },
    {
      "epoch": 0.18060149219290825,
      "grad_norm": 2.140625,
      "learning_rate": 2.729097761710638e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1026572.0,
      "step": 2348
    },
    {
      "epoch": 0.18067840935312668,
      "grad_norm": 2.59375,
      "learning_rate": 2.72898238597031e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1027096.0,
      "step": 2349
    },
    {
      "epoch": 0.18075532651334514,
      "grad_norm": 3.140625,
      "learning_rate": 2.7288670102299824e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1027676.0,
      "step": 2350
    },
    {
      "epoch": 0.18083224367356357,
      "grad_norm": 3.078125,
      "learning_rate": 2.7287516344896546e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1028240.0,
      "step": 2351
    },
    {
      "epoch": 0.18090916083378203,
      "grad_norm": 6.4375,
      "learning_rate": 2.728636258749327e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1028550.0,
      "step": 2352
    },
    {
      "epoch": 0.18098607799400046,
      "grad_norm": 5.0,
      "learning_rate": 2.7285208830089996e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1028874.0,
      "step": 2353
    },
    {
      "epoch": 0.18106299515421892,
      "grad_norm": 5.6875,
      "learning_rate": 2.728405507268672e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1029222.0,
      "step": 2354
    },
    {
      "epoch": 0.18113991231443735,
      "grad_norm": 2.484375,
      "learning_rate": 2.728290131528344e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1029772.0,
      "step": 2355
    },
    {
      "epoch": 0.18121682947465578,
      "grad_norm": 5.34375,
      "learning_rate": 2.7281747557880163e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 1030110.0,
      "step": 2356
    },
    {
      "epoch": 0.18129374663487424,
      "grad_norm": 2.1875,
      "learning_rate": 2.7280593800476886e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9937655925750732,
      "num_tokens": 1030914.0,
      "step": 2357
    },
    {
      "epoch": 0.18137066379509267,
      "grad_norm": 7.375,
      "learning_rate": 2.727944004307361e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 1031186.0,
      "step": 2358
    },
    {
      "epoch": 0.18144758095531113,
      "grad_norm": 6.40625,
      "learning_rate": 2.7278286285670335e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1031534.0,
      "step": 2359
    },
    {
      "epoch": 0.18152449811552956,
      "grad_norm": 6.09375,
      "learning_rate": 2.727713252826706e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1031848.0,
      "step": 2360
    },
    {
      "epoch": 0.18160141527574802,
      "grad_norm": 6.125,
      "learning_rate": 2.727597877086378e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1032196.0,
      "step": 2361
    },
    {
      "epoch": 0.18167833243596646,
      "grad_norm": 6.0625,
      "learning_rate": 2.7274825013460503e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1032540.0,
      "step": 2362
    },
    {
      "epoch": 0.18175524959618491,
      "grad_norm": 3.890625,
      "learning_rate": 2.7273671256057226e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1032868.0,
      "step": 2363
    },
    {
      "epoch": 0.18183216675640335,
      "grad_norm": 2.6875,
      "learning_rate": 2.7272517498653953e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1033414.0,
      "step": 2364
    },
    {
      "epoch": 0.1819090839166218,
      "grad_norm": 3.046875,
      "learning_rate": 2.7271363741250675e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1033936.0,
      "step": 2365
    },
    {
      "epoch": 0.18198600107684024,
      "grad_norm": 2.25,
      "learning_rate": 2.7270209983847398e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1034488.0,
      "step": 2366
    },
    {
      "epoch": 0.1820629182370587,
      "grad_norm": 1.9609375,
      "learning_rate": 2.726905622644412e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1035062.0,
      "step": 2367
    },
    {
      "epoch": 0.18213983539727713,
      "grad_norm": 3.828125,
      "learning_rate": 2.7267902469040843e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1035632.0,
      "step": 2368
    },
    {
      "epoch": 0.1822167525574956,
      "grad_norm": 3.296875,
      "learning_rate": 2.7266748711637565e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 1036186.0,
      "step": 2369
    },
    {
      "epoch": 0.18229366971771402,
      "grad_norm": 3.78125,
      "learning_rate": 2.7265594954234292e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 1036762.0,
      "step": 2370
    },
    {
      "epoch": 0.18237058687793248,
      "grad_norm": 3.65625,
      "learning_rate": 2.7264441196831015e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1037112.0,
      "step": 2371
    },
    {
      "epoch": 0.1824475040381509,
      "grad_norm": 3.171875,
      "learning_rate": 2.7263287439427737e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1037916.0,
      "step": 2372
    },
    {
      "epoch": 0.18252442119836937,
      "grad_norm": 3.328125,
      "learning_rate": 2.726213368202446e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1038438.0,
      "step": 2373
    },
    {
      "epoch": 0.1826013383585878,
      "grad_norm": 2.9375,
      "learning_rate": 2.7260979924621183e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1038986.0,
      "step": 2374
    },
    {
      "epoch": 0.18267825551880623,
      "grad_norm": 8.4375,
      "learning_rate": 2.7259826167217905e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1039292.0,
      "step": 2375
    },
    {
      "epoch": 0.1827551726790247,
      "grad_norm": 10.625,
      "learning_rate": 2.725867240981463e-06,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 1039546.0,
      "step": 2376
    },
    {
      "epoch": 0.18283208983924312,
      "grad_norm": 7.0625,
      "learning_rate": 2.7257518652411355e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1039830.0,
      "step": 2377
    },
    {
      "epoch": 0.18290900699946158,
      "grad_norm": 5.25,
      "learning_rate": 2.7256364895008077e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1040140.0,
      "step": 2378
    },
    {
      "epoch": 0.18298592415968,
      "grad_norm": 7.53125,
      "learning_rate": 2.72552111376048e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1040442.0,
      "step": 2379
    },
    {
      "epoch": 0.18306284131989847,
      "grad_norm": 5.25,
      "learning_rate": 2.7254057380201522e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 1040766.0,
      "step": 2380
    },
    {
      "epoch": 0.1831397584801169,
      "grad_norm": 3.5625,
      "learning_rate": 2.725290362279825e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1041352.0,
      "step": 2381
    },
    {
      "epoch": 0.18321667564033536,
      "grad_norm": 5.4375,
      "learning_rate": 2.725174986539497e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 1041676.0,
      "step": 2382
    },
    {
      "epoch": 0.1832935928005538,
      "grad_norm": 2.515625,
      "learning_rate": 2.725059610799169e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 1042248.0,
      "step": 2383
    },
    {
      "epoch": 0.18337050996077225,
      "grad_norm": 2.90625,
      "learning_rate": 2.7249442350588417e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1042810.0,
      "step": 2384
    },
    {
      "epoch": 0.18344742712099069,
      "grad_norm": 2.421875,
      "learning_rate": 2.724828859318514e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1043378.0,
      "step": 2385
    },
    {
      "epoch": 0.18352434428120915,
      "grad_norm": 3.0,
      "learning_rate": 2.724713483578186e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1043952.0,
      "step": 2386
    },
    {
      "epoch": 0.18360126144142758,
      "grad_norm": 1.9765625,
      "learning_rate": 2.724598107837859e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1044496.0,
      "step": 2387
    },
    {
      "epoch": 0.18367817860164604,
      "grad_norm": 2.8125,
      "learning_rate": 2.724482732097531e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1045316.0,
      "step": 2388
    },
    {
      "epoch": 0.18375509576186447,
      "grad_norm": 3.75,
      "learning_rate": 2.7243673563572034e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1045832.0,
      "step": 2389
    },
    {
      "epoch": 0.18383201292208293,
      "grad_norm": 7.3125,
      "learning_rate": 2.7242519806168757e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1046150.0,
      "step": 2390
    },
    {
      "epoch": 0.18390893008230136,
      "grad_norm": 8.25,
      "learning_rate": 2.724136604876548e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1046420.0,
      "step": 2391
    },
    {
      "epoch": 0.18398584724251982,
      "grad_norm": 2.515625,
      "learning_rate": 2.7240212291362206e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 1047470.0,
      "step": 2392
    },
    {
      "epoch": 0.18406276440273825,
      "grad_norm": 8.8125,
      "learning_rate": 2.723905853395893e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 1047730.0,
      "step": 2393
    },
    {
      "epoch": 0.1841396815629567,
      "grad_norm": 2.90625,
      "learning_rate": 2.7237904776555647e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 1048320.0,
      "step": 2394
    },
    {
      "epoch": 0.18421659872317514,
      "grad_norm": 2.375,
      "learning_rate": 2.7236751019152374e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 1049118.0,
      "step": 2395
    },
    {
      "epoch": 0.18429351588339357,
      "grad_norm": 2.859375,
      "learning_rate": 2.7235597261749096e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1049664.0,
      "step": 2396
    },
    {
      "epoch": 0.18437043304361203,
      "grad_norm": 5.15625,
      "learning_rate": 2.723444350434582e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 1049984.0,
      "step": 2397
    },
    {
      "epoch": 0.18444735020383046,
      "grad_norm": 8.1875,
      "learning_rate": 2.7233289746942546e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 1050288.0,
      "step": 2398
    },
    {
      "epoch": 0.18452426736404892,
      "grad_norm": 2.96875,
      "learning_rate": 2.723213598953927e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1050836.0,
      "step": 2399
    },
    {
      "epoch": 0.18460118452426735,
      "grad_norm": 2.96875,
      "learning_rate": 2.723098223213599e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1051384.0,
      "step": 2400
    },
    {
      "epoch": 0.1846781016844858,
      "grad_norm": 8.1875,
      "learning_rate": 2.7229828474732713e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1051698.0,
      "step": 2401
    },
    {
      "epoch": 0.18475501884470424,
      "grad_norm": 12.25,
      "learning_rate": 2.7228674717329436e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 1051950.0,
      "step": 2402
    },
    {
      "epoch": 0.1848319360049227,
      "grad_norm": 6.875,
      "learning_rate": 2.7227520959926163e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1052280.0,
      "step": 2403
    },
    {
      "epoch": 0.18490885316514113,
      "grad_norm": 3.28125,
      "learning_rate": 2.7226367202522885e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.991482138633728,
      "num_tokens": 1052868.0,
      "step": 2404
    },
    {
      "epoch": 0.1849857703253596,
      "grad_norm": 9.0625,
      "learning_rate": 2.7225213445119604e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1053198.0,
      "step": 2405
    },
    {
      "epoch": 0.18506268748557803,
      "grad_norm": 1.9921875,
      "learning_rate": 2.722405968771633e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1053760.0,
      "step": 2406
    },
    {
      "epoch": 0.18513960464579649,
      "grad_norm": 3.765625,
      "learning_rate": 2.7222905930313053e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1054286.0,
      "step": 2407
    },
    {
      "epoch": 0.18521652180601492,
      "grad_norm": 8.5,
      "learning_rate": 2.7221752172909776e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 1054540.0,
      "step": 2408
    },
    {
      "epoch": 0.18529343896623338,
      "grad_norm": 2.234375,
      "learning_rate": 2.7220598415506503e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1055344.0,
      "step": 2409
    },
    {
      "epoch": 0.1853703561264518,
      "grad_norm": 4.03125,
      "learning_rate": 2.721944465810322e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1055840.0,
      "step": 2410
    },
    {
      "epoch": 0.18544727328667027,
      "grad_norm": 2.53125,
      "learning_rate": 2.7218290900699948e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1056396.0,
      "step": 2411
    },
    {
      "epoch": 0.1855241904468887,
      "grad_norm": 2.703125,
      "learning_rate": 2.721713714329667e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1056976.0,
      "step": 2412
    },
    {
      "epoch": 0.18560110760710716,
      "grad_norm": 4.59375,
      "learning_rate": 2.7215983385893393e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1057300.0,
      "step": 2413
    },
    {
      "epoch": 0.1856780247673256,
      "grad_norm": 5.78125,
      "learning_rate": 2.721482962849012e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1057608.0,
      "step": 2414
    },
    {
      "epoch": 0.18575494192754405,
      "grad_norm": 2.609375,
      "learning_rate": 2.7213675871086842e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1058166.0,
      "step": 2415
    },
    {
      "epoch": 0.18583185908776248,
      "grad_norm": 3.015625,
      "learning_rate": 2.721252211368356e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1058712.0,
      "step": 2416
    },
    {
      "epoch": 0.1859087762479809,
      "grad_norm": 2.484375,
      "learning_rate": 2.7211368356280287e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1059270.0,
      "step": 2417
    },
    {
      "epoch": 0.18598569340819937,
      "grad_norm": 3.828125,
      "learning_rate": 2.721021459887701e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 1059588.0,
      "step": 2418
    },
    {
      "epoch": 0.1860626105684178,
      "grad_norm": 7.21875,
      "learning_rate": 2.7209060841473733e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1059870.0,
      "step": 2419
    },
    {
      "epoch": 0.18613952772863626,
      "grad_norm": 6.1875,
      "learning_rate": 2.720790708407046e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 1060156.0,
      "step": 2420
    },
    {
      "epoch": 0.1862164448888547,
      "grad_norm": 3.765625,
      "learning_rate": 2.7206753326667178e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.994575023651123,
      "num_tokens": 1060710.0,
      "step": 2421
    },
    {
      "epoch": 0.18629336204907315,
      "grad_norm": 5.9375,
      "learning_rate": 2.7205599569263905e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 1061282.0,
      "step": 2422
    },
    {
      "epoch": 0.18637027920929158,
      "grad_norm": 11.5,
      "learning_rate": 2.7204445811860627e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 1061646.0,
      "step": 2423
    },
    {
      "epoch": 0.18644719636951004,
      "grad_norm": 10.75,
      "learning_rate": 2.720329205445735e-06,
      "loss": 0.0419,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 1061926.0,
      "step": 2424
    },
    {
      "epoch": 0.18652411352972847,
      "grad_norm": 2.0625,
      "learning_rate": 2.7202138297054072e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1062484.0,
      "step": 2425
    },
    {
      "epoch": 0.18660103068994693,
      "grad_norm": 7.78125,
      "learning_rate": 2.72009845396508e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1062794.0,
      "step": 2426
    },
    {
      "epoch": 0.18667794785016537,
      "grad_norm": 4.0625,
      "learning_rate": 2.7199830782247518e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1063364.0,
      "step": 2427
    },
    {
      "epoch": 0.18675486501038382,
      "grad_norm": 3.65625,
      "learning_rate": 2.7198677024844244e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1063688.0,
      "step": 2428
    },
    {
      "epoch": 0.18683178217060226,
      "grad_norm": 4.34375,
      "learning_rate": 2.7197523267440967e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 1064266.0,
      "step": 2429
    },
    {
      "epoch": 0.18690869933082072,
      "grad_norm": 2.875,
      "learning_rate": 2.719636951003769e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1064856.0,
      "step": 2430
    },
    {
      "epoch": 0.18698561649103915,
      "grad_norm": 7.0625,
      "learning_rate": 2.7195215752634416e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1065184.0,
      "step": 2431
    },
    {
      "epoch": 0.1870625336512576,
      "grad_norm": 4.25,
      "learning_rate": 2.7194061995231135e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 1065502.0,
      "step": 2432
    },
    {
      "epoch": 0.18713945081147604,
      "grad_norm": 4.4375,
      "learning_rate": 2.7192908237827857e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 1065848.0,
      "step": 2433
    },
    {
      "epoch": 0.1872163679716945,
      "grad_norm": 11.6875,
      "learning_rate": 2.7191754480424584e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1066110.0,
      "step": 2434
    },
    {
      "epoch": 0.18729328513191293,
      "grad_norm": 3.34375,
      "learning_rate": 2.7190600723021307e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1066432.0,
      "step": 2435
    },
    {
      "epoch": 0.1873702022921314,
      "grad_norm": 3.234375,
      "learning_rate": 2.718944696561803e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1066754.0,
      "step": 2436
    },
    {
      "epoch": 0.18744711945234982,
      "grad_norm": 2.375,
      "learning_rate": 2.7188293208214756e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1067300.0,
      "step": 2437
    },
    {
      "epoch": 0.18752403661256825,
      "grad_norm": 6.21875,
      "learning_rate": 2.7187139450811474e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 1067602.0,
      "step": 2438
    },
    {
      "epoch": 0.1876009537727867,
      "grad_norm": 6.71875,
      "learning_rate": 2.71859856934082e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 1068100.0,
      "step": 2439
    },
    {
      "epoch": 0.18767787093300514,
      "grad_norm": 5.8125,
      "learning_rate": 2.7184831936004924e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1068408.0,
      "step": 2440
    },
    {
      "epoch": 0.1877547880932236,
      "grad_norm": 6.75,
      "learning_rate": 2.7183678178601646e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 1068768.0,
      "step": 2441
    },
    {
      "epoch": 0.18783170525344203,
      "grad_norm": 4.34375,
      "learning_rate": 2.7182524421198373e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1069324.0,
      "step": 2442
    },
    {
      "epoch": 0.1879086224136605,
      "grad_norm": 4.3125,
      "learning_rate": 2.718137066379509e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1069874.0,
      "step": 2443
    },
    {
      "epoch": 0.18798553957387892,
      "grad_norm": 9.125,
      "learning_rate": 2.7180216906391814e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1070224.0,
      "step": 2444
    },
    {
      "epoch": 0.18806245673409738,
      "grad_norm": 3.171875,
      "learning_rate": 2.717906314898854e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1070768.0,
      "step": 2445
    },
    {
      "epoch": 0.18813937389431581,
      "grad_norm": 3.53125,
      "learning_rate": 2.7177909391585263e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 1071340.0,
      "step": 2446
    },
    {
      "epoch": 0.18821629105453427,
      "grad_norm": 5.6875,
      "learning_rate": 2.7176755634181986e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1071628.0,
      "step": 2447
    },
    {
      "epoch": 0.1882932082147527,
      "grad_norm": 3.59375,
      "learning_rate": 2.717560187677871e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1072172.0,
      "step": 2448
    },
    {
      "epoch": 0.18837012537497116,
      "grad_norm": 6.53125,
      "learning_rate": 2.717444811937543e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1072536.0,
      "step": 2449
    },
    {
      "epoch": 0.1884470425351896,
      "grad_norm": 6.625,
      "learning_rate": 2.717329436197216e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 1072892.0,
      "step": 2450
    },
    {
      "epoch": 0.18852395969540806,
      "grad_norm": 4.34375,
      "learning_rate": 2.717214060456888e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1073212.0,
      "step": 2451
    },
    {
      "epoch": 0.1886008768556265,
      "grad_norm": 3.21875,
      "learning_rate": 2.7170986847165603e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1073756.0,
      "step": 2452
    },
    {
      "epoch": 0.18867779401584495,
      "grad_norm": 5.09375,
      "learning_rate": 2.716983308976233e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1074080.0,
      "step": 2453
    },
    {
      "epoch": 0.18875471117606338,
      "grad_norm": 3.46875,
      "learning_rate": 2.716867933235905e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1074620.0,
      "step": 2454
    },
    {
      "epoch": 0.18883162833628184,
      "grad_norm": 7.78125,
      "learning_rate": 2.716752557495577e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1074976.0,
      "step": 2455
    },
    {
      "epoch": 0.18890854549650027,
      "grad_norm": 3.546875,
      "learning_rate": 2.7166371817552498e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1075326.0,
      "step": 2456
    },
    {
      "epoch": 0.1889854626567187,
      "grad_norm": 3.265625,
      "learning_rate": 2.716521806014922e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1075890.0,
      "step": 2457
    },
    {
      "epoch": 0.18906237981693716,
      "grad_norm": 5.40625,
      "learning_rate": 2.7164064302745943e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1076214.0,
      "step": 2458
    },
    {
      "epoch": 0.1891392969771556,
      "grad_norm": 6.65625,
      "learning_rate": 2.7162910545342666e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 1076550.0,
      "step": 2459
    },
    {
      "epoch": 0.18921621413737405,
      "grad_norm": 2.234375,
      "learning_rate": 2.716175678793939e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1077110.0,
      "step": 2460
    },
    {
      "epoch": 0.18929313129759248,
      "grad_norm": 1.9296875,
      "learning_rate": 2.7160603030536115e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1077696.0,
      "step": 2461
    },
    {
      "epoch": 0.18937004845781094,
      "grad_norm": 2.71875,
      "learning_rate": 2.7159449273132837e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1078282.0,
      "step": 2462
    },
    {
      "epoch": 0.18944696561802937,
      "grad_norm": 5.53125,
      "learning_rate": 2.715829551572956e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1078608.0,
      "step": 2463
    },
    {
      "epoch": 0.18952388277824783,
      "grad_norm": 7.53125,
      "learning_rate": 2.7157141758326287e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1078896.0,
      "step": 2464
    },
    {
      "epoch": 0.18960079993846626,
      "grad_norm": 2.40625,
      "learning_rate": 2.7155988000923005e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1079716.0,
      "step": 2465
    },
    {
      "epoch": 0.18967771709868472,
      "grad_norm": 3.59375,
      "learning_rate": 2.7154834243519728e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 1080284.0,
      "step": 2466
    },
    {
      "epoch": 0.18975463425890315,
      "grad_norm": 10.0625,
      "learning_rate": 2.7153680486116455e-06,
      "loss": 0.0391,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1080546.0,
      "step": 2467
    },
    {
      "epoch": 0.1898315514191216,
      "grad_norm": 4.75,
      "learning_rate": 2.7152526728713177e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 1080898.0,
      "step": 2468
    },
    {
      "epoch": 0.18990846857934005,
      "grad_norm": 1.9609375,
      "learning_rate": 2.71513729713099e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 1081712.0,
      "step": 2469
    },
    {
      "epoch": 0.1899853857395585,
      "grad_norm": 7.96875,
      "learning_rate": 2.7150219213906622e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1082030.0,
      "step": 2470
    },
    {
      "epoch": 0.19006230289977694,
      "grad_norm": 4.34375,
      "learning_rate": 2.7149065456503345e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 1082594.0,
      "step": 2471
    },
    {
      "epoch": 0.1901392200599954,
      "grad_norm": 10.9375,
      "learning_rate": 2.714791169910007e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1082872.0,
      "step": 2472
    },
    {
      "epoch": 0.19021613722021383,
      "grad_norm": 5.0,
      "learning_rate": 2.7146757941696794e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1083368.0,
      "step": 2473
    },
    {
      "epoch": 0.19029305438043229,
      "grad_norm": 5.15625,
      "learning_rate": 2.7145604184293517e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9943661689758301,
      "num_tokens": 1083724.0,
      "step": 2474
    },
    {
      "epoch": 0.19036997154065072,
      "grad_norm": 6.0625,
      "learning_rate": 2.714445042689024e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 1084050.0,
      "step": 2475
    },
    {
      "epoch": 0.19044688870086918,
      "grad_norm": 1.9296875,
      "learning_rate": 2.714329666948696e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9961832165718079,
      "num_tokens": 1084838.0,
      "step": 2476
    },
    {
      "epoch": 0.1905238058610876,
      "grad_norm": 8.1875,
      "learning_rate": 2.7142142912083685e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 1085098.0,
      "step": 2477
    },
    {
      "epoch": 0.19060072302130604,
      "grad_norm": 5.09375,
      "learning_rate": 2.714098915468041e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1085414.0,
      "step": 2478
    },
    {
      "epoch": 0.1906776401815245,
      "grad_norm": 7.46875,
      "learning_rate": 2.7139835397277134e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1085720.0,
      "step": 2479
    },
    {
      "epoch": 0.19075455734174293,
      "grad_norm": 3.0,
      "learning_rate": 2.7138681639873857e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9907235503196716,
      "num_tokens": 1086260.0,
      "step": 2480
    },
    {
      "epoch": 0.1908314745019614,
      "grad_norm": 7.8125,
      "learning_rate": 2.713752788247058e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 1086518.0,
      "step": 2481
    },
    {
      "epoch": 0.19090839166217982,
      "grad_norm": 2.296875,
      "learning_rate": 2.71363741250673e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1087064.0,
      "step": 2482
    },
    {
      "epoch": 0.19098530882239828,
      "grad_norm": 2.40625,
      "learning_rate": 2.7135220367664024e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1087634.0,
      "step": 2483
    },
    {
      "epoch": 0.1910622259826167,
      "grad_norm": 5.75,
      "learning_rate": 2.713406661026075e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1087952.0,
      "step": 2484
    },
    {
      "epoch": 0.19113914314283517,
      "grad_norm": 5.3125,
      "learning_rate": 2.7132912852857474e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1088548.0,
      "step": 2485
    },
    {
      "epoch": 0.1912160603030536,
      "grad_norm": 7.0,
      "learning_rate": 2.7131759095454196e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1088864.0,
      "step": 2486
    },
    {
      "epoch": 0.19129297746327206,
      "grad_norm": 3.578125,
      "learning_rate": 2.713060533805092e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1089396.0,
      "step": 2487
    },
    {
      "epoch": 0.1913698946234905,
      "grad_norm": 4.34375,
      "learning_rate": 2.712945158064764e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1089678.0,
      "step": 2488
    },
    {
      "epoch": 0.19144681178370895,
      "grad_norm": 5.53125,
      "learning_rate": 2.712829782324437e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1089964.0,
      "step": 2489
    },
    {
      "epoch": 0.19152372894392738,
      "grad_norm": 5.71875,
      "learning_rate": 2.712714406584109e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 1090286.0,
      "step": 2490
    },
    {
      "epoch": 0.19160064610414584,
      "grad_norm": 5.46875,
      "learning_rate": 2.7125990308437814e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 1090562.0,
      "step": 2491
    },
    {
      "epoch": 0.19167756326436428,
      "grad_norm": 7.34375,
      "learning_rate": 2.7124836551034536e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1090858.0,
      "step": 2492
    },
    {
      "epoch": 0.19175448042458274,
      "grad_norm": 3.140625,
      "learning_rate": 2.712368279363126e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1091404.0,
      "step": 2493
    },
    {
      "epoch": 0.19183139758480117,
      "grad_norm": 2.6875,
      "learning_rate": 2.712252903622798e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9961612224578857,
      "num_tokens": 1091926.0,
      "step": 2494
    },
    {
      "epoch": 0.19190831474501963,
      "grad_norm": 2.703125,
      "learning_rate": 2.712137527882471e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1092468.0,
      "step": 2495
    },
    {
      "epoch": 0.19198523190523806,
      "grad_norm": 4.1875,
      "learning_rate": 2.712022152142143e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1092774.0,
      "step": 2496
    },
    {
      "epoch": 0.19206214906545652,
      "grad_norm": 3.84375,
      "learning_rate": 2.7119067764018153e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9915110468864441,
      "num_tokens": 1093364.0,
      "step": 2497
    },
    {
      "epoch": 0.19213906622567495,
      "grad_norm": 4.125,
      "learning_rate": 2.7117914006614876e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1093670.0,
      "step": 2498
    },
    {
      "epoch": 0.19221598338589338,
      "grad_norm": 11.625,
      "learning_rate": 2.71167602492116e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 1093994.0,
      "step": 2499
    },
    {
      "epoch": 0.19229290054611184,
      "grad_norm": 4.34375,
      "learning_rate": 2.7115606491808325e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1094288.0,
      "step": 2500
    },
    {
      "epoch": 0.19236981770633027,
      "grad_norm": 1.9609375,
      "learning_rate": 2.7114452734405048e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 1094828.0,
      "step": 2501
    },
    {
      "epoch": 0.19244673486654873,
      "grad_norm": 5.71875,
      "learning_rate": 2.711329897700177e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1095128.0,
      "step": 2502
    },
    {
      "epoch": 0.19252365202676716,
      "grad_norm": 2.53125,
      "learning_rate": 2.7112145219598493e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 1095924.0,
      "step": 2503
    },
    {
      "epoch": 0.19260056918698562,
      "grad_norm": 3.828125,
      "learning_rate": 2.7110991462195216e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1096444.0,
      "step": 2504
    },
    {
      "epoch": 0.19267748634720405,
      "grad_norm": 9.3125,
      "learning_rate": 2.710983770479194e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 1096776.0,
      "step": 2505
    },
    {
      "epoch": 0.1927544035074225,
      "grad_norm": 5.90625,
      "learning_rate": 2.7108683947388665e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1097086.0,
      "step": 2506
    },
    {
      "epoch": 0.19283132066764094,
      "grad_norm": 3.515625,
      "learning_rate": 2.7107530189985388e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1097392.0,
      "step": 2507
    },
    {
      "epoch": 0.1929082378278594,
      "grad_norm": 2.921875,
      "learning_rate": 2.710637643258211e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1097938.0,
      "step": 2508
    },
    {
      "epoch": 0.19298515498807783,
      "grad_norm": 3.328125,
      "learning_rate": 2.7105222675178833e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1098508.0,
      "step": 2509
    },
    {
      "epoch": 0.1930620721482963,
      "grad_norm": 5.28125,
      "learning_rate": 2.7104068917775555e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1098830.0,
      "step": 2510
    },
    {
      "epoch": 0.19313898930851472,
      "grad_norm": 4.375,
      "learning_rate": 2.710291516037228e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1099164.0,
      "step": 2511
    },
    {
      "epoch": 0.19321590646873318,
      "grad_norm": 1.8125,
      "learning_rate": 2.7101761402969005e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 1099734.0,
      "step": 2512
    },
    {
      "epoch": 0.19329282362895162,
      "grad_norm": 6.53125,
      "learning_rate": 2.7100607645565723e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 1100066.0,
      "step": 2513
    },
    {
      "epoch": 0.19336974078917007,
      "grad_norm": 5.09375,
      "learning_rate": 2.709945388816245e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1100388.0,
      "step": 2514
    },
    {
      "epoch": 0.1934466579493885,
      "grad_norm": 7.4375,
      "learning_rate": 2.7098300130759172e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1100744.0,
      "step": 2515
    },
    {
      "epoch": 0.19352357510960697,
      "grad_norm": 7.625,
      "learning_rate": 2.7097146373355895e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1101030.0,
      "step": 2516
    },
    {
      "epoch": 0.1936004922698254,
      "grad_norm": 6.28125,
      "learning_rate": 2.709599261595262e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1101308.0,
      "step": 2517
    },
    {
      "epoch": 0.19367740943004386,
      "grad_norm": 7.53125,
      "learning_rate": 2.7094838858549344e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 1101616.0,
      "step": 2518
    },
    {
      "epoch": 0.1937543265902623,
      "grad_norm": 4.1875,
      "learning_rate": 2.7093685101146067e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 1101946.0,
      "step": 2519
    },
    {
      "epoch": 0.19383124375048072,
      "grad_norm": 6.21875,
      "learning_rate": 2.709253134374279e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1102278.0,
      "step": 2520
    },
    {
      "epoch": 0.19390816091069918,
      "grad_norm": 4.03125,
      "learning_rate": 2.7091377586339512e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1102600.0,
      "step": 2521
    },
    {
      "epoch": 0.1939850780709176,
      "grad_norm": 5.46875,
      "learning_rate": 2.709022382893624e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 1102916.0,
      "step": 2522
    },
    {
      "epoch": 0.19406199523113607,
      "grad_norm": 5.15625,
      "learning_rate": 2.708907007153296e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1103210.0,
      "step": 2523
    },
    {
      "epoch": 0.1941389123913545,
      "grad_norm": 5.125,
      "learning_rate": 2.708791631412968e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1103504.0,
      "step": 2524
    },
    {
      "epoch": 0.19421582955157296,
      "grad_norm": 4.75,
      "learning_rate": 2.7086762556726407e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1103844.0,
      "step": 2525
    },
    {
      "epoch": 0.1942927467117914,
      "grad_norm": 5.0,
      "learning_rate": 2.708560879932313e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1104154.0,
      "step": 2526
    },
    {
      "epoch": 0.19436966387200985,
      "grad_norm": 2.46875,
      "learning_rate": 2.708445504191985e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1104712.0,
      "step": 2527
    },
    {
      "epoch": 0.19444658103222828,
      "grad_norm": 2.84375,
      "learning_rate": 2.708330128451658e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1105298.0,
      "step": 2528
    },
    {
      "epoch": 0.19452349819244674,
      "grad_norm": 2.21875,
      "learning_rate": 2.70821475271133e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1105884.0,
      "step": 2529
    },
    {
      "epoch": 0.19460041535266517,
      "grad_norm": 3.28125,
      "learning_rate": 2.7080993769710024e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1106228.0,
      "step": 2530
    },
    {
      "epoch": 0.19467733251288363,
      "grad_norm": 5.90625,
      "learning_rate": 2.7079840012306746e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 1106566.0,
      "step": 2531
    },
    {
      "epoch": 0.19475424967310206,
      "grad_norm": 2.859375,
      "learning_rate": 2.707868625490347e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 1107162.0,
      "step": 2532
    },
    {
      "epoch": 0.19483116683332052,
      "grad_norm": 4.625,
      "learning_rate": 2.707753249750019e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1107502.0,
      "step": 2533
    },
    {
      "epoch": 0.19490808399353896,
      "grad_norm": 2.921875,
      "learning_rate": 2.707637874009692e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1108092.0,
      "step": 2534
    },
    {
      "epoch": 0.19498500115375741,
      "grad_norm": 3.171875,
      "learning_rate": 2.7075224982693637e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1108874.0,
      "step": 2535
    },
    {
      "epoch": 0.19506191831397585,
      "grad_norm": 4.40625,
      "learning_rate": 2.7074071225290364e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9909256100654602,
      "num_tokens": 1109426.0,
      "step": 2536
    },
    {
      "epoch": 0.1951388354741943,
      "grad_norm": 2.546875,
      "learning_rate": 2.7072917467887086e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 1110224.0,
      "step": 2537
    },
    {
      "epoch": 0.19521575263441274,
      "grad_norm": 3.90625,
      "learning_rate": 2.707176371048381e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1110574.0,
      "step": 2538
    },
    {
      "epoch": 0.19529266979463117,
      "grad_norm": 2.625,
      "learning_rate": 2.7070609953080536e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 1111618.0,
      "step": 2539
    },
    {
      "epoch": 0.19536958695484963,
      "grad_norm": 2.375,
      "learning_rate": 2.706945619567726e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 1112214.0,
      "step": 2540
    },
    {
      "epoch": 0.19544650411506806,
      "grad_norm": 6.53125,
      "learning_rate": 2.7068302438273976e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1112514.0,
      "step": 2541
    },
    {
      "epoch": 0.19552342127528652,
      "grad_norm": 2.453125,
      "learning_rate": 2.7067148680870703e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 1113564.0,
      "step": 2542
    },
    {
      "epoch": 0.19560033843550495,
      "grad_norm": 5.8125,
      "learning_rate": 2.7065994923467426e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1113872.0,
      "step": 2543
    },
    {
      "epoch": 0.1956772555957234,
      "grad_norm": 8.125,
      "learning_rate": 2.706484116606415e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1114178.0,
      "step": 2544
    },
    {
      "epoch": 0.19575417275594184,
      "grad_norm": 4.84375,
      "learning_rate": 2.7063687408660875e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1114500.0,
      "step": 2545
    },
    {
      "epoch": 0.1958310899161603,
      "grad_norm": 4.96875,
      "learning_rate": 2.7062533651257594e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1114816.0,
      "step": 2546
    },
    {
      "epoch": 0.19590800707637873,
      "grad_norm": 6.59375,
      "learning_rate": 2.706137989385432e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 1115098.0,
      "step": 2547
    },
    {
      "epoch": 0.1959849242365972,
      "grad_norm": 2.734375,
      "learning_rate": 2.7060226136451043e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1115648.0,
      "step": 2548
    },
    {
      "epoch": 0.19606184139681562,
      "grad_norm": 2.046875,
      "learning_rate": 2.7059072379047766e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1116194.0,
      "step": 2549
    },
    {
      "epoch": 0.19613875855703408,
      "grad_norm": 2.046875,
      "learning_rate": 2.7057918621644492e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9975369572639465,
      "num_tokens": 1117008.0,
      "step": 2550
    },
    {
      "epoch": 0.1962156757172525,
      "grad_norm": 4.1875,
      "learning_rate": 2.705676486424121e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1117544.0,
      "step": 2551
    },
    {
      "epoch": 0.19629259287747097,
      "grad_norm": 7.25,
      "learning_rate": 2.7055611106837933e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1117900.0,
      "step": 2552
    },
    {
      "epoch": 0.1963695100376894,
      "grad_norm": 1.703125,
      "learning_rate": 2.705445734943466e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9987593293190002,
      "num_tokens": 1118708.0,
      "step": 2553
    },
    {
      "epoch": 0.19644642719790786,
      "grad_norm": 3.1875,
      "learning_rate": 2.7053303592031383e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1119260.0,
      "step": 2554
    },
    {
      "epoch": 0.1965233443581263,
      "grad_norm": 3.21875,
      "learning_rate": 2.7052149834628105e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1119824.0,
      "step": 2555
    },
    {
      "epoch": 0.19660026151834475,
      "grad_norm": 2.625,
      "learning_rate": 2.705099607722483e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1120364.0,
      "step": 2556
    },
    {
      "epoch": 0.19667717867856319,
      "grad_norm": 7.875,
      "learning_rate": 2.704984231982155e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1120674.0,
      "step": 2557
    },
    {
      "epoch": 0.19675409583878165,
      "grad_norm": 2.515625,
      "learning_rate": 2.7048688562418277e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1121258.0,
      "step": 2558
    },
    {
      "epoch": 0.19683101299900008,
      "grad_norm": 1.953125,
      "learning_rate": 2.7047534805015e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1121808.0,
      "step": 2559
    },
    {
      "epoch": 0.1969079301592185,
      "grad_norm": 8.0625,
      "learning_rate": 2.7046381047611722e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.98525071144104,
      "num_tokens": 1122150.0,
      "step": 2560
    },
    {
      "epoch": 0.19698484731943697,
      "grad_norm": 7.3125,
      "learning_rate": 2.704522729020845e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1122456.0,
      "step": 2561
    },
    {
      "epoch": 0.1970617644796554,
      "grad_norm": 2.515625,
      "learning_rate": 2.7044073532805168e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1123000.0,
      "step": 2562
    },
    {
      "epoch": 0.19713868163987386,
      "grad_norm": 2.65625,
      "learning_rate": 2.704291977540189e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1123522.0,
      "step": 2563
    },
    {
      "epoch": 0.1972155988000923,
      "grad_norm": 7.03125,
      "learning_rate": 2.7041766017998617e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 1123830.0,
      "step": 2564
    },
    {
      "epoch": 0.19729251596031075,
      "grad_norm": 2.84375,
      "learning_rate": 2.704061226059534e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1124416.0,
      "step": 2565
    },
    {
      "epoch": 0.19736943312052918,
      "grad_norm": 4.15625,
      "learning_rate": 2.7039458503192062e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1124726.0,
      "step": 2566
    },
    {
      "epoch": 0.19744635028074764,
      "grad_norm": 6.40625,
      "learning_rate": 2.703830474578879e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 1125004.0,
      "step": 2567
    },
    {
      "epoch": 0.19752326744096607,
      "grad_norm": 4.78125,
      "learning_rate": 2.7037150988385507e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1125310.0,
      "step": 2568
    },
    {
      "epoch": 0.19760018460118453,
      "grad_norm": 2.5,
      "learning_rate": 2.7035997230982234e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 1126332.0,
      "step": 2569
    },
    {
      "epoch": 0.19767710176140296,
      "grad_norm": 5.0625,
      "learning_rate": 2.7034843473578957e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1126640.0,
      "step": 2570
    },
    {
      "epoch": 0.19775401892162142,
      "grad_norm": 7.65625,
      "learning_rate": 2.703368971617568e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1126926.0,
      "step": 2571
    },
    {
      "epoch": 0.19783093608183985,
      "grad_norm": 5.875,
      "learning_rate": 2.7032535958772406e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1127238.0,
      "step": 2572
    },
    {
      "epoch": 0.1979078532420583,
      "grad_norm": 4.78125,
      "learning_rate": 2.7031382201369124e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1127520.0,
      "step": 2573
    },
    {
      "epoch": 0.19798477040227674,
      "grad_norm": 6.375,
      "learning_rate": 2.7030228443965847e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1127864.0,
      "step": 2574
    },
    {
      "epoch": 0.1980616875624952,
      "grad_norm": 2.765625,
      "learning_rate": 2.7029074686562574e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 1128660.0,
      "step": 2575
    },
    {
      "epoch": 0.19813860472271363,
      "grad_norm": 3.171875,
      "learning_rate": 2.7027920929159296e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1129486.0,
      "step": 2576
    },
    {
      "epoch": 0.1982155218829321,
      "grad_norm": 4.40625,
      "learning_rate": 2.702676717175602e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1129768.0,
      "step": 2577
    },
    {
      "epoch": 0.19829243904315053,
      "grad_norm": 4.09375,
      "learning_rate": 2.702561341435274e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1130072.0,
      "step": 2578
    },
    {
      "epoch": 0.19836935620336899,
      "grad_norm": 4.59375,
      "learning_rate": 2.7024459656949464e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1130398.0,
      "step": 2579
    },
    {
      "epoch": 0.19844627336358742,
      "grad_norm": 3.8125,
      "learning_rate": 2.702330589954619e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 1130732.0,
      "step": 2580
    },
    {
      "epoch": 0.19852319052380585,
      "grad_norm": 6.625,
      "learning_rate": 2.7022152142142914e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1131060.0,
      "step": 2581
    },
    {
      "epoch": 0.1986001076840243,
      "grad_norm": 3.34375,
      "learning_rate": 2.7020998384739636e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 1131602.0,
      "step": 2582
    },
    {
      "epoch": 0.19867702484424274,
      "grad_norm": 2.234375,
      "learning_rate": 2.701984462733636e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1132146.0,
      "step": 2583
    },
    {
      "epoch": 0.1987539420044612,
      "grad_norm": 5.53125,
      "learning_rate": 2.701869086993308e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 1132448.0,
      "step": 2584
    },
    {
      "epoch": 0.19883085916467963,
      "grad_norm": 4.96875,
      "learning_rate": 2.7017537112529804e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 1132778.0,
      "step": 2585
    },
    {
      "epoch": 0.1989077763248981,
      "grad_norm": 7.0625,
      "learning_rate": 2.701638335512653e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 1133068.0,
      "step": 2586
    },
    {
      "epoch": 0.19898469348511652,
      "grad_norm": 6.75,
      "learning_rate": 2.7015229597723253e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1133352.0,
      "step": 2587
    },
    {
      "epoch": 0.19906161064533498,
      "grad_norm": 3.390625,
      "learning_rate": 2.7014075840319976e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1133916.0,
      "step": 2588
    },
    {
      "epoch": 0.1991385278055534,
      "grad_norm": 3.53125,
      "learning_rate": 2.70129220829167e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1134222.0,
      "step": 2589
    },
    {
      "epoch": 0.19921544496577187,
      "grad_norm": 2.921875,
      "learning_rate": 2.701176832551342e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1134806.0,
      "step": 2590
    },
    {
      "epoch": 0.1992923621259903,
      "grad_norm": 3.359375,
      "learning_rate": 2.7010614568110144e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 1135386.0,
      "step": 2591
    },
    {
      "epoch": 0.19936927928620876,
      "grad_norm": 6.71875,
      "learning_rate": 2.700946081070687e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 1135722.0,
      "step": 2592
    },
    {
      "epoch": 0.1994461964464272,
      "grad_norm": 2.75,
      "learning_rate": 2.7008307053303593e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1136272.0,
      "step": 2593
    },
    {
      "epoch": 0.19952311360664565,
      "grad_norm": 4.625,
      "learning_rate": 2.7007153295900316e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1136594.0,
      "step": 2594
    },
    {
      "epoch": 0.19960003076686408,
      "grad_norm": 3.125,
      "learning_rate": 2.700599953849704e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1137180.0,
      "step": 2595
    },
    {
      "epoch": 0.19967694792708254,
      "grad_norm": 6.03125,
      "learning_rate": 2.700484578109376e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 1137494.0,
      "step": 2596
    },
    {
      "epoch": 0.19975386508730097,
      "grad_norm": 7.84375,
      "learning_rate": 2.7003692023690488e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1137836.0,
      "step": 2597
    },
    {
      "epoch": 0.19983078224751943,
      "grad_norm": 3.09375,
      "learning_rate": 2.700253826628721e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 1138354.0,
      "step": 2598
    },
    {
      "epoch": 0.19990769940773787,
      "grad_norm": 6.625,
      "learning_rate": 2.7001384508883933e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1138640.0,
      "step": 2599
    },
    {
      "epoch": 0.19998461656795632,
      "grad_norm": 6.96875,
      "learning_rate": 2.7000230751480655e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1138984.0,
      "step": 2600
    },
    {
      "epoch": 0.20006153372817476,
      "grad_norm": 2.640625,
      "learning_rate": 2.699907699407738e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1139548.0,
      "step": 2601
    },
    {
      "epoch": 0.2001384508883932,
      "grad_norm": 5.71875,
      "learning_rate": 2.69979232366741e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 1139864.0,
      "step": 2602
    },
    {
      "epoch": 0.20021536804861165,
      "grad_norm": 8.375,
      "learning_rate": 2.6996769479270827e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1140166.0,
      "step": 2603
    },
    {
      "epoch": 0.20029228520883008,
      "grad_norm": 5.46875,
      "learning_rate": 2.699561572186755e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9930555820465088,
      "num_tokens": 1140456.0,
      "step": 2604
    },
    {
      "epoch": 0.20036920236904854,
      "grad_norm": 7.9375,
      "learning_rate": 2.6994461964464272e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1140736.0,
      "step": 2605
    },
    {
      "epoch": 0.20044611952926697,
      "grad_norm": 3.75,
      "learning_rate": 2.6993308207060995e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1141044.0,
      "step": 2606
    },
    {
      "epoch": 0.20052303668948543,
      "grad_norm": 12.375,
      "learning_rate": 2.6992154449657718e-06,
      "loss": 0.046,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 1141296.0,
      "step": 2607
    },
    {
      "epoch": 0.20059995384970386,
      "grad_norm": 3.234375,
      "learning_rate": 2.6991000692254444e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 1141858.0,
      "step": 2608
    },
    {
      "epoch": 0.20067687100992232,
      "grad_norm": 6.28125,
      "learning_rate": 2.6989846934851167e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 1142198.0,
      "step": 2609
    },
    {
      "epoch": 0.20075378817014075,
      "grad_norm": 6.6875,
      "learning_rate": 2.698869317744789e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 1142544.0,
      "step": 2610
    },
    {
      "epoch": 0.2008307053303592,
      "grad_norm": 5.125,
      "learning_rate": 2.6987539420044612e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.992337167263031,
      "num_tokens": 1142806.0,
      "step": 2611
    },
    {
      "epoch": 0.20090762249057764,
      "grad_norm": 2.359375,
      "learning_rate": 2.6986385662641335e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1143370.0,
      "step": 2612
    },
    {
      "epoch": 0.2009845396507961,
      "grad_norm": 5.625,
      "learning_rate": 2.6985231905238057e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1143720.0,
      "step": 2613
    },
    {
      "epoch": 0.20106145681101453,
      "grad_norm": 3.5,
      "learning_rate": 2.6984078147834784e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1144296.0,
      "step": 2614
    },
    {
      "epoch": 0.201138373971233,
      "grad_norm": 6.53125,
      "learning_rate": 2.6982924390431507e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 1144556.0,
      "step": 2615
    },
    {
      "epoch": 0.20121529113145142,
      "grad_norm": 5.78125,
      "learning_rate": 2.698177063302823e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1144862.0,
      "step": 2616
    },
    {
      "epoch": 0.20129220829166988,
      "grad_norm": 3.75,
      "learning_rate": 2.698061687562495e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9913644194602966,
      "num_tokens": 1145442.0,
      "step": 2617
    },
    {
      "epoch": 0.20136912545188831,
      "grad_norm": 3.34375,
      "learning_rate": 2.6979463118221674e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1145750.0,
      "step": 2618
    },
    {
      "epoch": 0.20144604261210677,
      "grad_norm": 3.265625,
      "learning_rate": 2.69783093608184e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1146326.0,
      "step": 2619
    },
    {
      "epoch": 0.2015229597723252,
      "grad_norm": 6.125,
      "learning_rate": 2.6977155603415124e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1146664.0,
      "step": 2620
    },
    {
      "epoch": 0.20159987693254364,
      "grad_norm": 6.59375,
      "learning_rate": 2.6976001846011846e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 1146924.0,
      "step": 2621
    },
    {
      "epoch": 0.2016767940927621,
      "grad_norm": 5.34375,
      "learning_rate": 2.697484808860857e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 1147264.0,
      "step": 2622
    },
    {
      "epoch": 0.20175371125298053,
      "grad_norm": 8.625,
      "learning_rate": 2.697369433120529e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 1147572.0,
      "step": 2623
    },
    {
      "epoch": 0.201830628413199,
      "grad_norm": 4.75,
      "learning_rate": 2.6972540573802014e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1147860.0,
      "step": 2624
    },
    {
      "epoch": 0.20190754557341742,
      "grad_norm": 4.25,
      "learning_rate": 2.697138681639874e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1148402.0,
      "step": 2625
    },
    {
      "epoch": 0.20198446273363588,
      "grad_norm": 3.734375,
      "learning_rate": 2.6970233058995464e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1148708.0,
      "step": 2626
    },
    {
      "epoch": 0.2020613798938543,
      "grad_norm": 5.3125,
      "learning_rate": 2.6969079301592186e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 1149052.0,
      "step": 2627
    },
    {
      "epoch": 0.20213829705407277,
      "grad_norm": 7.21875,
      "learning_rate": 2.696792554418891e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1149354.0,
      "step": 2628
    },
    {
      "epoch": 0.2022152142142912,
      "grad_norm": 4.71875,
      "learning_rate": 2.696677178678563e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1149950.0,
      "step": 2629
    },
    {
      "epoch": 0.20229213137450966,
      "grad_norm": 2.78125,
      "learning_rate": 2.696561802938236e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1150736.0,
      "step": 2630
    },
    {
      "epoch": 0.2023690485347281,
      "grad_norm": 5.5,
      "learning_rate": 2.696446427197908e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 1151070.0,
      "step": 2631
    },
    {
      "epoch": 0.20244596569494655,
      "grad_norm": 3.21875,
      "learning_rate": 2.6963310514575803e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1151656.0,
      "step": 2632
    },
    {
      "epoch": 0.20252288285516498,
      "grad_norm": 2.9375,
      "learning_rate": 2.6962156757172526e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1152212.0,
      "step": 2633
    },
    {
      "epoch": 0.20259980001538344,
      "grad_norm": 7.8125,
      "learning_rate": 2.696100299976925e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 1152546.0,
      "step": 2634
    },
    {
      "epoch": 0.20267671717560187,
      "grad_norm": 6.4375,
      "learning_rate": 2.695984924236597e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 1152902.0,
      "step": 2635
    },
    {
      "epoch": 0.20275363433582033,
      "grad_norm": 3.5,
      "learning_rate": 2.69586954849627e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1153218.0,
      "step": 2636
    },
    {
      "epoch": 0.20283055149603876,
      "grad_norm": 5.25,
      "learning_rate": 2.695754172755942e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1153546.0,
      "step": 2637
    },
    {
      "epoch": 0.20290746865625722,
      "grad_norm": 6.8125,
      "learning_rate": 2.6956387970156143e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1153854.0,
      "step": 2638
    },
    {
      "epoch": 0.20298438581647565,
      "grad_norm": 3.203125,
      "learning_rate": 2.6955234212752866e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1154402.0,
      "step": 2639
    },
    {
      "epoch": 0.2030613029766941,
      "grad_norm": 3.4375,
      "learning_rate": 2.695408045534959e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1154722.0,
      "step": 2640
    },
    {
      "epoch": 0.20313822013691255,
      "grad_norm": 2.59375,
      "learning_rate": 2.6952926697946315e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 1155318.0,
      "step": 2641
    },
    {
      "epoch": 0.20321513729713098,
      "grad_norm": 7.78125,
      "learning_rate": 2.6951772940543038e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1155626.0,
      "step": 2642
    },
    {
      "epoch": 0.20329205445734944,
      "grad_norm": 9.4375,
      "learning_rate": 2.695061918313976e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9794721603393555,
      "num_tokens": 1155968.0,
      "step": 2643
    },
    {
      "epoch": 0.20336897161756787,
      "grad_norm": 6.6875,
      "learning_rate": 2.6949465425736483e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1156278.0,
      "step": 2644
    },
    {
      "epoch": 0.20344588877778633,
      "grad_norm": 3.84375,
      "learning_rate": 2.6948311668333205e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 1156822.0,
      "step": 2645
    },
    {
      "epoch": 0.20352280593800476,
      "grad_norm": 1.6484375,
      "learning_rate": 2.694715791092993e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1157402.0,
      "step": 2646
    },
    {
      "epoch": 0.20359972309822322,
      "grad_norm": 7.9375,
      "learning_rate": 2.6946004153526655e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1157716.0,
      "step": 2647
    },
    {
      "epoch": 0.20367664025844165,
      "grad_norm": 6.78125,
      "learning_rate": 2.6944850396123377e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1158018.0,
      "step": 2648
    },
    {
      "epoch": 0.2037535574186601,
      "grad_norm": 6.25,
      "learning_rate": 2.69436966387201e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 1158352.0,
      "step": 2649
    },
    {
      "epoch": 0.20383047457887854,
      "grad_norm": 2.171875,
      "learning_rate": 2.6942542881316822e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1158908.0,
      "step": 2650
    },
    {
      "epoch": 0.203907391739097,
      "grad_norm": 2.734375,
      "learning_rate": 2.6941389123913545e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1159482.0,
      "step": 2651
    },
    {
      "epoch": 0.20398430889931543,
      "grad_norm": 2.359375,
      "learning_rate": 2.6940235366510268e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1160038.0,
      "step": 2652
    },
    {
      "epoch": 0.2040612260595339,
      "grad_norm": 4.96875,
      "learning_rate": 2.6939081609106994e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1160362.0,
      "step": 2653
    },
    {
      "epoch": 0.20413814321975232,
      "grad_norm": 3.28125,
      "learning_rate": 2.6937927851703713e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1160926.0,
      "step": 2654
    },
    {
      "epoch": 0.20421506037997078,
      "grad_norm": 2.515625,
      "learning_rate": 2.693677409430044e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1161484.0,
      "step": 2655
    },
    {
      "epoch": 0.2042919775401892,
      "grad_norm": 2.0625,
      "learning_rate": 2.6935620336897162e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1162052.0,
      "step": 2656
    },
    {
      "epoch": 0.20436889470040767,
      "grad_norm": 6.875,
      "learning_rate": 2.6934466579493885e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 1162366.0,
      "step": 2657
    },
    {
      "epoch": 0.2044458118606261,
      "grad_norm": 3.71875,
      "learning_rate": 2.693331282209061e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1162916.0,
      "step": 2658
    },
    {
      "epoch": 0.20452272902084456,
      "grad_norm": 2.796875,
      "learning_rate": 2.6932159064687334e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1163450.0,
      "step": 2659
    },
    {
      "epoch": 0.204599646181063,
      "grad_norm": 3.421875,
      "learning_rate": 2.6931005307284053e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1163970.0,
      "step": 2660
    },
    {
      "epoch": 0.20467656334128145,
      "grad_norm": 3.15625,
      "learning_rate": 2.692985154988078e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 1164548.0,
      "step": 2661
    },
    {
      "epoch": 0.20475348050149988,
      "grad_norm": 6.0625,
      "learning_rate": 2.69286977924775e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1164854.0,
      "step": 2662
    },
    {
      "epoch": 0.20483039766171832,
      "grad_norm": 5.6875,
      "learning_rate": 2.6927544035074225e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 1165170.0,
      "step": 2663
    },
    {
      "epoch": 0.20490731482193678,
      "grad_norm": 6.0,
      "learning_rate": 2.692639027767095e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1165454.0,
      "step": 2664
    },
    {
      "epoch": 0.2049842319821552,
      "grad_norm": 6.96875,
      "learning_rate": 2.692523652026767e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1165788.0,
      "step": 2665
    },
    {
      "epoch": 0.20506114914237367,
      "grad_norm": 5.75,
      "learning_rate": 2.6924082762864396e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1166092.0,
      "step": 2666
    },
    {
      "epoch": 0.2051380663025921,
      "grad_norm": 2.140625,
      "learning_rate": 2.692292900546112e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1166650.0,
      "step": 2667
    },
    {
      "epoch": 0.20521498346281056,
      "grad_norm": 5.8125,
      "learning_rate": 2.692177524805784e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1166966.0,
      "step": 2668
    },
    {
      "epoch": 0.205291900623029,
      "grad_norm": 6.59375,
      "learning_rate": 2.692062149065457e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 1167302.0,
      "step": 2669
    },
    {
      "epoch": 0.20536881778324745,
      "grad_norm": 8.6875,
      "learning_rate": 2.691946773325129e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 1167556.0,
      "step": 2670
    },
    {
      "epoch": 0.20544573494346588,
      "grad_norm": 6.625,
      "learning_rate": 2.691831397584801e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 1167858.0,
      "step": 2671
    },
    {
      "epoch": 0.20552265210368434,
      "grad_norm": 2.53125,
      "learning_rate": 2.6917160218444736e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 1168656.0,
      "step": 2672
    },
    {
      "epoch": 0.20559956926390277,
      "grad_norm": 3.796875,
      "learning_rate": 2.691600646104146e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1168962.0,
      "step": 2673
    },
    {
      "epoch": 0.20567648642412123,
      "grad_norm": 6.53125,
      "learning_rate": 2.691485270363818e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 1169274.0,
      "step": 2674
    },
    {
      "epoch": 0.20575340358433966,
      "grad_norm": 6.78125,
      "learning_rate": 2.691369894623491e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1169618.0,
      "step": 2675
    },
    {
      "epoch": 0.20583032074455812,
      "grad_norm": 2.796875,
      "learning_rate": 2.6912545188831627e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 1170158.0,
      "step": 2676
    },
    {
      "epoch": 0.20590723790477655,
      "grad_norm": 6.8125,
      "learning_rate": 2.6911391431428353e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1170488.0,
      "step": 2677
    },
    {
      "epoch": 0.205984155064995,
      "grad_norm": 1.8984375,
      "learning_rate": 2.6910237674025076e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1171030.0,
      "step": 2678
    },
    {
      "epoch": 0.20606107222521344,
      "grad_norm": 6.78125,
      "learning_rate": 2.69090839166218e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1171290.0,
      "step": 2679
    },
    {
      "epoch": 0.2061379893854319,
      "grad_norm": 6.3125,
      "learning_rate": 2.6907930159218525e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1171560.0,
      "step": 2680
    },
    {
      "epoch": 0.20621490654565033,
      "grad_norm": 3.671875,
      "learning_rate": 2.6906776401815244e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 1172136.0,
      "step": 2681
    },
    {
      "epoch": 0.2062918237058688,
      "grad_norm": 6.5,
      "learning_rate": 2.6905622644411966e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1172436.0,
      "step": 2682
    },
    {
      "epoch": 0.20636874086608722,
      "grad_norm": 5.3125,
      "learning_rate": 2.6904468887008693e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1172752.0,
      "step": 2683
    },
    {
      "epoch": 0.20644565802630566,
      "grad_norm": 2.1875,
      "learning_rate": 2.6903315129605416e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 1173322.0,
      "step": 2684
    },
    {
      "epoch": 0.20652257518652412,
      "grad_norm": 3.265625,
      "learning_rate": 2.690216137220214e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9904030561447144,
      "num_tokens": 1173844.0,
      "step": 2685
    },
    {
      "epoch": 0.20659949234674255,
      "grad_norm": 2.671875,
      "learning_rate": 2.6901007614798865e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1174368.0,
      "step": 2686
    },
    {
      "epoch": 0.206676409506961,
      "grad_norm": 5.15625,
      "learning_rate": 2.6899853857395583e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 1174714.0,
      "step": 2687
    },
    {
      "epoch": 0.20675332666717944,
      "grad_norm": 5.09375,
      "learning_rate": 2.689870009999231e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1175030.0,
      "step": 2688
    },
    {
      "epoch": 0.2068302438273979,
      "grad_norm": 7.90625,
      "learning_rate": 2.6897546342589033e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 1175340.0,
      "step": 2689
    },
    {
      "epoch": 0.20690716098761633,
      "grad_norm": 5.125,
      "learning_rate": 2.6896392585185755e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1175658.0,
      "step": 2690
    },
    {
      "epoch": 0.2069840781478348,
      "grad_norm": 3.46875,
      "learning_rate": 2.6895238827782482e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1176204.0,
      "step": 2691
    },
    {
      "epoch": 0.20706099530805322,
      "grad_norm": 5.53125,
      "learning_rate": 2.68940850703792e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1176486.0,
      "step": 2692
    },
    {
      "epoch": 0.20713791246827168,
      "grad_norm": 6.125,
      "learning_rate": 2.6892931312975923e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1176768.0,
      "step": 2693
    },
    {
      "epoch": 0.2072148296284901,
      "grad_norm": 2.3125,
      "learning_rate": 2.689177755557265e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1177324.0,
      "step": 2694
    },
    {
      "epoch": 0.20729174678870857,
      "grad_norm": 2.40625,
      "learning_rate": 2.6890623798169373e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 1178368.0,
      "step": 2695
    },
    {
      "epoch": 0.207368663948927,
      "grad_norm": 2.984375,
      "learning_rate": 2.6889470040766095e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1178920.0,
      "step": 2696
    },
    {
      "epoch": 0.20744558110914546,
      "grad_norm": 2.3125,
      "learning_rate": 2.688831628336282e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1179476.0,
      "step": 2697
    },
    {
      "epoch": 0.2075224982693639,
      "grad_norm": 5.65625,
      "learning_rate": 2.688716252595954e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 1179832.0,
      "step": 2698
    },
    {
      "epoch": 0.20759941542958235,
      "grad_norm": 6.71875,
      "learning_rate": 2.6886008768556267e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1180116.0,
      "step": 2699
    },
    {
      "epoch": 0.20767633258980078,
      "grad_norm": 2.28125,
      "learning_rate": 2.688485501115299e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1180650.0,
      "step": 2700
    },
    {
      "epoch": 0.20775324975001924,
      "grad_norm": 4.78125,
      "learning_rate": 2.6883701253749712e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1180956.0,
      "step": 2701
    },
    {
      "epoch": 0.20783016691023767,
      "grad_norm": 5.4375,
      "learning_rate": 2.6882547496346435e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 1181316.0,
      "step": 2702
    },
    {
      "epoch": 0.2079070840704561,
      "grad_norm": 5.53125,
      "learning_rate": 2.6881393738943157e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 1181678.0,
      "step": 2703
    },
    {
      "epoch": 0.20798400123067456,
      "grad_norm": 9.0625,
      "learning_rate": 2.688023998153988e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1182020.0,
      "step": 2704
    },
    {
      "epoch": 0.208060918390893,
      "grad_norm": 2.09375,
      "learning_rate": 2.6879086224136607e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1182582.0,
      "step": 2705
    },
    {
      "epoch": 0.20813783555111146,
      "grad_norm": 8.5625,
      "learning_rate": 2.687793246673333e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1182910.0,
      "step": 2706
    },
    {
      "epoch": 0.2082147527113299,
      "grad_norm": 1.875,
      "learning_rate": 2.687677870933005e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1183494.0,
      "step": 2707
    },
    {
      "epoch": 0.20829166987154835,
      "grad_norm": 2.96875,
      "learning_rate": 2.687562495192678e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1184040.0,
      "step": 2708
    },
    {
      "epoch": 0.20836858703176678,
      "grad_norm": 9.4375,
      "learning_rate": 2.6874471194523497e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 1184332.0,
      "step": 2709
    },
    {
      "epoch": 0.20844550419198524,
      "grad_norm": 2.921875,
      "learning_rate": 2.687331743712022e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9909256100654602,
      "num_tokens": 1184884.0,
      "step": 2710
    },
    {
      "epoch": 0.20852242135220367,
      "grad_norm": 1.8671875,
      "learning_rate": 2.6872163679716947e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1185446.0,
      "step": 2711
    },
    {
      "epoch": 0.20859933851242213,
      "grad_norm": 3.921875,
      "learning_rate": 2.687100992231367e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 1186014.0,
      "step": 2712
    },
    {
      "epoch": 0.20867625567264056,
      "grad_norm": 4.21875,
      "learning_rate": 2.686985616491039e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1186556.0,
      "step": 2713
    },
    {
      "epoch": 0.20875317283285902,
      "grad_norm": 2.234375,
      "learning_rate": 2.6868702407507114e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1187130.0,
      "step": 2714
    },
    {
      "epoch": 0.20883008999307745,
      "grad_norm": 6.0625,
      "learning_rate": 2.6867548650103837e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1187426.0,
      "step": 2715
    },
    {
      "epoch": 0.2089070071532959,
      "grad_norm": 2.125,
      "learning_rate": 2.6866394892700564e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1187986.0,
      "step": 2716
    },
    {
      "epoch": 0.20898392431351434,
      "grad_norm": 6.03125,
      "learning_rate": 2.6865241135297286e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1188270.0,
      "step": 2717
    },
    {
      "epoch": 0.2090608414737328,
      "grad_norm": 10.0625,
      "learning_rate": 2.686408737789401e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1188572.0,
      "step": 2718
    },
    {
      "epoch": 0.20913775863395123,
      "grad_norm": 5.75,
      "learning_rate": 2.686293362049073e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 1188890.0,
      "step": 2719
    },
    {
      "epoch": 0.2092146757941697,
      "grad_norm": 4.6875,
      "learning_rate": 2.6861779863087454e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1189224.0,
      "step": 2720
    },
    {
      "epoch": 0.20929159295438812,
      "grad_norm": 3.65625,
      "learning_rate": 2.6860626105684177e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1189536.0,
      "step": 2721
    },
    {
      "epoch": 0.20936851011460658,
      "grad_norm": 6.03125,
      "learning_rate": 2.6859472348280903e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 1189844.0,
      "step": 2722
    },
    {
      "epoch": 0.209445427274825,
      "grad_norm": 2.328125,
      "learning_rate": 2.6858318590877626e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1190394.0,
      "step": 2723
    },
    {
      "epoch": 0.20952234443504344,
      "grad_norm": 2.421875,
      "learning_rate": 2.685716483347435e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1190916.0,
      "step": 2724
    },
    {
      "epoch": 0.2095992615952619,
      "grad_norm": 4.6875,
      "learning_rate": 2.685601107607107e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1191490.0,
      "step": 2725
    },
    {
      "epoch": 0.20967617875548034,
      "grad_norm": 7.15625,
      "learning_rate": 2.6854857318667794e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1191846.0,
      "step": 2726
    },
    {
      "epoch": 0.2097530959156988,
      "grad_norm": 3.453125,
      "learning_rate": 2.685370356126452e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1192432.0,
      "step": 2727
    },
    {
      "epoch": 0.20983001307591723,
      "grad_norm": 6.84375,
      "learning_rate": 2.6852549803861243e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 1192768.0,
      "step": 2728
    },
    {
      "epoch": 0.20990693023613569,
      "grad_norm": 7.09375,
      "learning_rate": 2.6851396046457966e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1193048.0,
      "step": 2729
    },
    {
      "epoch": 0.20998384739635412,
      "grad_norm": 7.03125,
      "learning_rate": 2.685024228905469e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1193354.0,
      "step": 2730
    },
    {
      "epoch": 0.21006076455657258,
      "grad_norm": 4.09375,
      "learning_rate": 2.684908853165141e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1193666.0,
      "step": 2731
    },
    {
      "epoch": 0.210137681716791,
      "grad_norm": 4.6875,
      "learning_rate": 2.6847934774248133e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9929577708244324,
      "num_tokens": 1193952.0,
      "step": 2732
    },
    {
      "epoch": 0.21021459887700947,
      "grad_norm": 5.25,
      "learning_rate": 2.684678101684486e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1194270.0,
      "step": 2733
    },
    {
      "epoch": 0.2102915160372279,
      "grad_norm": 2.375,
      "learning_rate": 2.6845627259441583e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1194844.0,
      "step": 2734
    },
    {
      "epoch": 0.21036843319744636,
      "grad_norm": 3.96875,
      "learning_rate": 2.6844473502038305e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1195420.0,
      "step": 2735
    },
    {
      "epoch": 0.2104453503576648,
      "grad_norm": 5.25,
      "learning_rate": 2.684331974463503e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1195738.0,
      "step": 2736
    },
    {
      "epoch": 0.21052226751788325,
      "grad_norm": 8.125,
      "learning_rate": 2.684216598723175e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 1196040.0,
      "step": 2737
    },
    {
      "epoch": 0.21059918467810168,
      "grad_norm": 4.6875,
      "learning_rate": 2.6841012229828477e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1196338.0,
      "step": 2738
    },
    {
      "epoch": 0.21067610183832014,
      "grad_norm": 2.578125,
      "learning_rate": 2.68398584724252e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 1196888.0,
      "step": 2739
    },
    {
      "epoch": 0.21075301899853857,
      "grad_norm": 8.25,
      "learning_rate": 2.6838704715021923e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1197230.0,
      "step": 2740
    },
    {
      "epoch": 0.21082993615875703,
      "grad_norm": 2.890625,
      "learning_rate": 2.6837550957618645e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 1198028.0,
      "step": 2741
    },
    {
      "epoch": 0.21090685331897546,
      "grad_norm": 2.625,
      "learning_rate": 2.6836397200215368e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1198552.0,
      "step": 2742
    },
    {
      "epoch": 0.21098377047919392,
      "grad_norm": 3.40625,
      "learning_rate": 2.683524344281209e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1199132.0,
      "step": 2743
    },
    {
      "epoch": 0.21106068763941235,
      "grad_norm": 5.96875,
      "learning_rate": 2.6834089685408817e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1199430.0,
      "step": 2744
    },
    {
      "epoch": 0.21113760479963078,
      "grad_norm": 5.53125,
      "learning_rate": 2.683293592800554e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 1199720.0,
      "step": 2745
    },
    {
      "epoch": 0.21121452195984924,
      "grad_norm": 4.4375,
      "learning_rate": 2.6831782170602262e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1200300.0,
      "step": 2746
    },
    {
      "epoch": 0.21129143912006768,
      "grad_norm": 2.5,
      "learning_rate": 2.6830628413198985e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1200862.0,
      "step": 2747
    },
    {
      "epoch": 0.21136835628028613,
      "grad_norm": 3.21875,
      "learning_rate": 2.6829474655795707e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1201186.0,
      "step": 2748
    },
    {
      "epoch": 0.21144527344050457,
      "grad_norm": 2.921875,
      "learning_rate": 2.6828320898392434e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9948006868362427,
      "num_tokens": 1201764.0,
      "step": 2749
    },
    {
      "epoch": 0.21152219060072303,
      "grad_norm": 3.171875,
      "learning_rate": 2.6827167140989157e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1202310.0,
      "step": 2750
    },
    {
      "epoch": 0.21159910776094146,
      "grad_norm": 4.0,
      "learning_rate": 2.682601338358588e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 1202882.0,
      "step": 2751
    },
    {
      "epoch": 0.21167602492115992,
      "grad_norm": 12.25,
      "learning_rate": 2.68248596261826e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 1203114.0,
      "step": 2752
    },
    {
      "epoch": 0.21175294208137835,
      "grad_norm": 2.65625,
      "learning_rate": 2.6823705868779325e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1203442.0,
      "step": 2753
    },
    {
      "epoch": 0.2118298592415968,
      "grad_norm": 3.34375,
      "learning_rate": 2.6822552111376047e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1203766.0,
      "step": 2754
    },
    {
      "epoch": 0.21190677640181524,
      "grad_norm": 3.65625,
      "learning_rate": 2.6821398353972774e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1204362.0,
      "step": 2755
    },
    {
      "epoch": 0.2119836935620337,
      "grad_norm": 5.65625,
      "learning_rate": 2.6820244596569497e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 1204672.0,
      "step": 2756
    },
    {
      "epoch": 0.21206061072225213,
      "grad_norm": 3.328125,
      "learning_rate": 2.681909083916622e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1204996.0,
      "step": 2757
    },
    {
      "epoch": 0.2121375278824706,
      "grad_norm": 3.921875,
      "learning_rate": 2.681793708176294e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1205544.0,
      "step": 2758
    },
    {
      "epoch": 0.21221444504268902,
      "grad_norm": 2.328125,
      "learning_rate": 2.6816783324359664e-06,
      "loss": 0.0072,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 1206566.0,
      "step": 2759
    },
    {
      "epoch": 0.21229136220290748,
      "grad_norm": 1.7109375,
      "learning_rate": 2.6815629566956387e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1207130.0,
      "step": 2760
    },
    {
      "epoch": 0.2123682793631259,
      "grad_norm": 5.1875,
      "learning_rate": 2.6814475809553114e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1207450.0,
      "step": 2761
    },
    {
      "epoch": 0.21244519652334437,
      "grad_norm": 9.9375,
      "learning_rate": 2.6813322052149836e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1207768.0,
      "step": 2762
    },
    {
      "epoch": 0.2125221136835628,
      "grad_norm": 2.703125,
      "learning_rate": 2.681216829474656e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 1208364.0,
      "step": 2763
    },
    {
      "epoch": 0.21259903084378123,
      "grad_norm": 2.671875,
      "learning_rate": 2.681101453734328e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 1209148.0,
      "step": 2764
    },
    {
      "epoch": 0.2126759480039997,
      "grad_norm": 7.3125,
      "learning_rate": 2.6809860779940004e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 1209462.0,
      "step": 2765
    },
    {
      "epoch": 0.21275286516421812,
      "grad_norm": 6.21875,
      "learning_rate": 2.680870702253673e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 1209758.0,
      "step": 2766
    },
    {
      "epoch": 0.21282978232443658,
      "grad_norm": 6.9375,
      "learning_rate": 2.6807553265133453e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1210040.0,
      "step": 2767
    },
    {
      "epoch": 0.21290669948465502,
      "grad_norm": 5.75,
      "learning_rate": 2.680639950773017e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 1210336.0,
      "step": 2768
    },
    {
      "epoch": 0.21298361664487347,
      "grad_norm": 3.734375,
      "learning_rate": 2.68052457503269e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1210878.0,
      "step": 2769
    },
    {
      "epoch": 0.2130605338050919,
      "grad_norm": 3.46875,
      "learning_rate": 2.680409199292362e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1211372.0,
      "step": 2770
    },
    {
      "epoch": 0.21313745096531037,
      "grad_norm": 5.375,
      "learning_rate": 2.6802938235520344e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1211682.0,
      "step": 2771
    },
    {
      "epoch": 0.2132143681255288,
      "grad_norm": 7.0625,
      "learning_rate": 2.680178447811707e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1211994.0,
      "step": 2772
    },
    {
      "epoch": 0.21329128528574726,
      "grad_norm": 5.59375,
      "learning_rate": 2.6800630720713793e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1212322.0,
      "step": 2773
    },
    {
      "epoch": 0.2133682024459657,
      "grad_norm": 2.671875,
      "learning_rate": 2.6799476963310516e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1212862.0,
      "step": 2774
    },
    {
      "epoch": 0.21344511960618415,
      "grad_norm": 2.84375,
      "learning_rate": 2.679832320590724e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 1213660.0,
      "step": 2775
    },
    {
      "epoch": 0.21352203676640258,
      "grad_norm": 3.109375,
      "learning_rate": 2.679716944850396e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1214234.0,
      "step": 2776
    },
    {
      "epoch": 0.21359895392662104,
      "grad_norm": 2.59375,
      "learning_rate": 2.6796015691100688e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1215038.0,
      "step": 2777
    },
    {
      "epoch": 0.21367587108683947,
      "grad_norm": 5.5625,
      "learning_rate": 2.679486193369741e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1215388.0,
      "step": 2778
    },
    {
      "epoch": 0.21375278824705793,
      "grad_norm": 7.375,
      "learning_rate": 2.679370817629413e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1215696.0,
      "step": 2779
    },
    {
      "epoch": 0.21382970540727636,
      "grad_norm": 7.3125,
      "learning_rate": 2.6792554418890855e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1216058.0,
      "step": 2780
    },
    {
      "epoch": 0.21390662256749482,
      "grad_norm": 5.65625,
      "learning_rate": 2.679140066148758e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1216338.0,
      "step": 2781
    },
    {
      "epoch": 0.21398353972771325,
      "grad_norm": 7.5625,
      "learning_rate": 2.67902469040843e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1216624.0,
      "step": 2782
    },
    {
      "epoch": 0.2140604568879317,
      "grad_norm": 3.09375,
      "learning_rate": 2.6789093146681027e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1217202.0,
      "step": 2783
    },
    {
      "epoch": 0.21413737404815014,
      "grad_norm": 4.4375,
      "learning_rate": 2.6787939389277746e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1217528.0,
      "step": 2784
    },
    {
      "epoch": 0.21421429120836857,
      "grad_norm": 2.609375,
      "learning_rate": 2.6786785631874473e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1218084.0,
      "step": 2785
    },
    {
      "epoch": 0.21429120836858703,
      "grad_norm": 5.21875,
      "learning_rate": 2.6785631874471195e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1218368.0,
      "step": 2786
    },
    {
      "epoch": 0.21436812552880546,
      "grad_norm": 4.71875,
      "learning_rate": 2.6784478117067918e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1218690.0,
      "step": 2787
    },
    {
      "epoch": 0.21444504268902392,
      "grad_norm": 6.53125,
      "learning_rate": 2.6783324359664645e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1219046.0,
      "step": 2788
    },
    {
      "epoch": 0.21452195984924236,
      "grad_norm": 8.3125,
      "learning_rate": 2.6782170602261367e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 1219306.0,
      "step": 2789
    },
    {
      "epoch": 0.21459887700946081,
      "grad_norm": 3.09375,
      "learning_rate": 2.6781016844858085e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1219858.0,
      "step": 2790
    },
    {
      "epoch": 0.21467579416967925,
      "grad_norm": 6.6875,
      "learning_rate": 2.6779863087454812e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1220154.0,
      "step": 2791
    },
    {
      "epoch": 0.2147527113298977,
      "grad_norm": 5.53125,
      "learning_rate": 2.6778709330051535e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.993630588054657,
      "num_tokens": 1220470.0,
      "step": 2792
    },
    {
      "epoch": 0.21482962849011614,
      "grad_norm": 3.0,
      "learning_rate": 2.6777555572648257e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 1221022.0,
      "step": 2793
    },
    {
      "epoch": 0.2149065456503346,
      "grad_norm": 2.421875,
      "learning_rate": 2.6776401815244984e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1221574.0,
      "step": 2794
    },
    {
      "epoch": 0.21498346281055303,
      "grad_norm": 4.53125,
      "learning_rate": 2.6775248057841703e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1221898.0,
      "step": 2795
    },
    {
      "epoch": 0.2150603799707715,
      "grad_norm": 4.875,
      "learning_rate": 2.677409430043843e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.993630588054657,
      "num_tokens": 1222214.0,
      "step": 2796
    },
    {
      "epoch": 0.21513729713098992,
      "grad_norm": 4.3125,
      "learning_rate": 2.677294054303515e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1222794.0,
      "step": 2797
    },
    {
      "epoch": 0.21521421429120838,
      "grad_norm": 7.09375,
      "learning_rate": 2.6771786785631875e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1223100.0,
      "step": 2798
    },
    {
      "epoch": 0.2152911314514268,
      "grad_norm": 10.0625,
      "learning_rate": 2.67706330282286e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9850187301635742,
      "num_tokens": 1223368.0,
      "step": 2799
    },
    {
      "epoch": 0.21536804861164527,
      "grad_norm": 6.1875,
      "learning_rate": 2.6769479270825324e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1223716.0,
      "step": 2800
    },
    {
      "epoch": 0.2154449657718637,
      "grad_norm": 3.6875,
      "learning_rate": 2.6768325513422042e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1224040.0,
      "step": 2801
    },
    {
      "epoch": 0.21552188293208216,
      "grad_norm": 5.46875,
      "learning_rate": 2.676717175601877e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9894179701805115,
      "num_tokens": 1224608.0,
      "step": 2802
    },
    {
      "epoch": 0.2155988000923006,
      "grad_norm": 5.03125,
      "learning_rate": 2.676601799861549e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 1224944.0,
      "step": 2803
    },
    {
      "epoch": 0.21567571725251905,
      "grad_norm": 3.859375,
      "learning_rate": 2.6764864241212214e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1225480.0,
      "step": 2804
    },
    {
      "epoch": 0.21575263441273748,
      "grad_norm": 8.375,
      "learning_rate": 2.676371048380894e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.982300877571106,
      "num_tokens": 1225820.0,
      "step": 2805
    },
    {
      "epoch": 0.2158295515729559,
      "grad_norm": 6.9375,
      "learning_rate": 2.676255672640566e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1226112.0,
      "step": 2806
    },
    {
      "epoch": 0.21590646873317437,
      "grad_norm": 3.421875,
      "learning_rate": 2.6761402969002386e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 1226688.0,
      "step": 2807
    },
    {
      "epoch": 0.2159833858933928,
      "grad_norm": 6.84375,
      "learning_rate": 2.676024921159911e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1226970.0,
      "step": 2808
    },
    {
      "epoch": 0.21606030305361126,
      "grad_norm": 6.75,
      "learning_rate": 2.675909545419583e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1227316.0,
      "step": 2809
    },
    {
      "epoch": 0.2161372202138297,
      "grad_norm": 6.0,
      "learning_rate": 2.6757941696792554e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 1227634.0,
      "step": 2810
    },
    {
      "epoch": 0.21621413737404815,
      "grad_norm": 4.71875,
      "learning_rate": 2.675678793938928e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1227918.0,
      "step": 2811
    },
    {
      "epoch": 0.21629105453426659,
      "grad_norm": 2.890625,
      "learning_rate": 2.6755634181986e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1228458.0,
      "step": 2812
    },
    {
      "epoch": 0.21636797169448505,
      "grad_norm": 3.75,
      "learning_rate": 2.6754480424582726e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1228998.0,
      "step": 2813
    },
    {
      "epoch": 0.21644488885470348,
      "grad_norm": 7.40625,
      "learning_rate": 2.675332666717945e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1229304.0,
      "step": 2814
    },
    {
      "epoch": 0.21652180601492194,
      "grad_norm": 4.09375,
      "learning_rate": 2.675217290977617e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1229824.0,
      "step": 2815
    },
    {
      "epoch": 0.21659872317514037,
      "grad_norm": 2.015625,
      "learning_rate": 2.67510191523729e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1230370.0,
      "step": 2816
    },
    {
      "epoch": 0.21667564033535883,
      "grad_norm": 2.671875,
      "learning_rate": 2.6749865394969616e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1230894.0,
      "step": 2817
    },
    {
      "epoch": 0.21675255749557726,
      "grad_norm": 5.4375,
      "learning_rate": 2.674871163756634e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1231214.0,
      "step": 2818
    },
    {
      "epoch": 0.21682947465579572,
      "grad_norm": 4.46875,
      "learning_rate": 2.6747557880163066e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9931740760803223,
      "num_tokens": 1231508.0,
      "step": 2819
    },
    {
      "epoch": 0.21690639181601415,
      "grad_norm": 5.78125,
      "learning_rate": 2.674640412275979e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 1231854.0,
      "step": 2820
    },
    {
      "epoch": 0.2169833089762326,
      "grad_norm": 2.015625,
      "learning_rate": 2.674525036535651e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1232400.0,
      "step": 2821
    },
    {
      "epoch": 0.21706022613645104,
      "grad_norm": 2.671875,
      "learning_rate": 2.6744096607953233e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 1232916.0,
      "step": 2822
    },
    {
      "epoch": 0.2171371432966695,
      "grad_norm": 3.671875,
      "learning_rate": 2.6742942850549956e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 1233752.0,
      "step": 2823
    },
    {
      "epoch": 0.21721406045688793,
      "grad_norm": 5.8125,
      "learning_rate": 2.6741789093146683e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1234102.0,
      "step": 2824
    },
    {
      "epoch": 0.2172909776171064,
      "grad_norm": 5.21875,
      "learning_rate": 2.6740635335743405e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1234404.0,
      "step": 2825
    },
    {
      "epoch": 0.21736789477732482,
      "grad_norm": 2.734375,
      "learning_rate": 2.673948157834013e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1234962.0,
      "step": 2826
    },
    {
      "epoch": 0.21744481193754325,
      "grad_norm": 5.40625,
      "learning_rate": 2.6738327820936855e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1235272.0,
      "step": 2827
    },
    {
      "epoch": 0.2175217290977617,
      "grad_norm": 8.375,
      "learning_rate": 2.6737174063533573e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1235614.0,
      "step": 2828
    },
    {
      "epoch": 0.21759864625798014,
      "grad_norm": 5.65625,
      "learning_rate": 2.6736020306130296e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1235912.0,
      "step": 2829
    },
    {
      "epoch": 0.2176755634181986,
      "grad_norm": 2.40625,
      "learning_rate": 2.6734866548727023e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1236738.0,
      "step": 2830
    },
    {
      "epoch": 0.21775248057841703,
      "grad_norm": 8.5625,
      "learning_rate": 2.6733712791323745e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1237054.0,
      "step": 2831
    },
    {
      "epoch": 0.2178293977386355,
      "grad_norm": 5.40625,
      "learning_rate": 2.6732559033920468e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1237352.0,
      "step": 2832
    },
    {
      "epoch": 0.21790631489885393,
      "grad_norm": 2.25,
      "learning_rate": 2.673140527651719e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1237936.0,
      "step": 2833
    },
    {
      "epoch": 0.21798323205907238,
      "grad_norm": 5.65625,
      "learning_rate": 2.6730251519113913e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1238242.0,
      "step": 2834
    },
    {
      "epoch": 0.21806014921929082,
      "grad_norm": 4.71875,
      "learning_rate": 2.672909776171064e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1238562.0,
      "step": 2835
    },
    {
      "epoch": 0.21813706637950928,
      "grad_norm": 2.328125,
      "learning_rate": 2.6727944004307362e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1239344.0,
      "step": 2836
    },
    {
      "epoch": 0.2182139835397277,
      "grad_norm": 9.0625,
      "learning_rate": 2.6726790246904085e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1239646.0,
      "step": 2837
    },
    {
      "epoch": 0.21829090069994617,
      "grad_norm": 3.9375,
      "learning_rate": 2.672563648950081e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1240222.0,
      "step": 2838
    },
    {
      "epoch": 0.2183678178601646,
      "grad_norm": 7.34375,
      "learning_rate": 2.672448273209753e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9967845678329468,
      "num_tokens": 1240536.0,
      "step": 2839
    },
    {
      "epoch": 0.21844473502038306,
      "grad_norm": 1.9609375,
      "learning_rate": 2.6723328974694253e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 1241126.0,
      "step": 2840
    },
    {
      "epoch": 0.2185216521806015,
      "grad_norm": 2.390625,
      "learning_rate": 2.672217521729098e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1241712.0,
      "step": 2841
    },
    {
      "epoch": 0.21859856934081995,
      "grad_norm": 8.0625,
      "learning_rate": 2.67210214598877e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 1241984.0,
      "step": 2842
    },
    {
      "epoch": 0.21867548650103838,
      "grad_norm": 1.90625,
      "learning_rate": 2.6719867702484425e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 1242798.0,
      "step": 2843
    },
    {
      "epoch": 0.21875240366125684,
      "grad_norm": 2.265625,
      "learning_rate": 2.6718713945081147e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1243348.0,
      "step": 2844
    },
    {
      "epoch": 0.21882932082147527,
      "grad_norm": 5.9375,
      "learning_rate": 2.671756018767787e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1243656.0,
      "step": 2845
    },
    {
      "epoch": 0.2189062379816937,
      "grad_norm": 7.0,
      "learning_rate": 2.6716406430274597e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1243986.0,
      "step": 2846
    },
    {
      "epoch": 0.21898315514191216,
      "grad_norm": 10.625,
      "learning_rate": 2.671525267287132e-06,
      "loss": 0.04,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 1244244.0,
      "step": 2847
    },
    {
      "epoch": 0.2190600723021306,
      "grad_norm": 7.3125,
      "learning_rate": 2.671409891546804e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1244562.0,
      "step": 2848
    },
    {
      "epoch": 0.21913698946234905,
      "grad_norm": 2.5,
      "learning_rate": 2.671294515806477e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1245108.0,
      "step": 2849
    },
    {
      "epoch": 0.21921390662256748,
      "grad_norm": 7.5,
      "learning_rate": 2.6711791400661487e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1245414.0,
      "step": 2850
    },
    {
      "epoch": 0.21929082378278594,
      "grad_norm": 2.484375,
      "learning_rate": 2.671063764325821e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1245978.0,
      "step": 2851
    },
    {
      "epoch": 0.21936774094300437,
      "grad_norm": 3.34375,
      "learning_rate": 2.6709483885854936e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1246496.0,
      "step": 2852
    },
    {
      "epoch": 0.21944465810322283,
      "grad_norm": 3.78125,
      "learning_rate": 2.670833012845166e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1246846.0,
      "step": 2853
    },
    {
      "epoch": 0.21952157526344127,
      "grad_norm": 4.8125,
      "learning_rate": 2.670717637104838e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1247182.0,
      "step": 2854
    },
    {
      "epoch": 0.21959849242365972,
      "grad_norm": 4.1875,
      "learning_rate": 2.6706022613645104e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 1247492.0,
      "step": 2855
    },
    {
      "epoch": 0.21967540958387816,
      "grad_norm": 8.4375,
      "learning_rate": 2.6704868856241827e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9919999837875366,
      "num_tokens": 1247870.0,
      "step": 2856
    },
    {
      "epoch": 0.21975232674409662,
      "grad_norm": 4.1875,
      "learning_rate": 2.6703715098838553e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 1248158.0,
      "step": 2857
    },
    {
      "epoch": 0.21982924390431505,
      "grad_norm": 4.15625,
      "learning_rate": 2.6702561341435276e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1248480.0,
      "step": 2858
    },
    {
      "epoch": 0.2199061610645335,
      "grad_norm": 3.453125,
      "learning_rate": 2.6701407584032e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1249030.0,
      "step": 2859
    },
    {
      "epoch": 0.21998307822475194,
      "grad_norm": 3.90625,
      "learning_rate": 2.670025382662872e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1249546.0,
      "step": 2860
    },
    {
      "epoch": 0.2200599953849704,
      "grad_norm": 4.9375,
      "learning_rate": 2.6699100069225444e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 1249890.0,
      "step": 2861
    },
    {
      "epoch": 0.22013691254518883,
      "grad_norm": 6.4375,
      "learning_rate": 2.6697946311822166e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 1250218.0,
      "step": 2862
    },
    {
      "epoch": 0.2202138297054073,
      "grad_norm": 2.46875,
      "learning_rate": 2.6696792554418893e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1250780.0,
      "step": 2863
    },
    {
      "epoch": 0.22029074686562572,
      "grad_norm": 4.5625,
      "learning_rate": 2.6695638797015616e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1251078.0,
      "step": 2864
    },
    {
      "epoch": 0.22036766402584418,
      "grad_norm": 3.890625,
      "learning_rate": 2.669448503961234e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1251400.0,
      "step": 2865
    },
    {
      "epoch": 0.2204445811860626,
      "grad_norm": 5.375,
      "learning_rate": 2.669333128220906e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 1251746.0,
      "step": 2866
    },
    {
      "epoch": 0.22052149834628104,
      "grad_norm": 3.828125,
      "learning_rate": 2.6692177524805784e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1252092.0,
      "step": 2867
    },
    {
      "epoch": 0.2205984155064995,
      "grad_norm": 6.0625,
      "learning_rate": 2.6691023767402506e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 1252404.0,
      "step": 2868
    },
    {
      "epoch": 0.22067533266671793,
      "grad_norm": 7.625,
      "learning_rate": 2.6689870009999233e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 1252694.0,
      "step": 2869
    },
    {
      "epoch": 0.2207522498269364,
      "grad_norm": 4.34375,
      "learning_rate": 2.6688716252595955e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1253004.0,
      "step": 2870
    },
    {
      "epoch": 0.22082916698715482,
      "grad_norm": 2.953125,
      "learning_rate": 2.668756249519268e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1253538.0,
      "step": 2871
    },
    {
      "epoch": 0.22090608414737328,
      "grad_norm": 5.46875,
      "learning_rate": 2.66864087377894e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 1253864.0,
      "step": 2872
    },
    {
      "epoch": 0.22098300130759171,
      "grad_norm": 5.03125,
      "learning_rate": 2.6685254980386123e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1254188.0,
      "step": 2873
    },
    {
      "epoch": 0.22105991846781017,
      "grad_norm": 5.71875,
      "learning_rate": 2.668410122298285e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1254528.0,
      "step": 2874
    },
    {
      "epoch": 0.2211368356280286,
      "grad_norm": 5.78125,
      "learning_rate": 2.6682947465579573e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1254832.0,
      "step": 2875
    },
    {
      "epoch": 0.22121375278824706,
      "grad_norm": 2.515625,
      "learning_rate": 2.6681793708176295e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 1255628.0,
      "step": 2876
    },
    {
      "epoch": 0.2212906699484655,
      "grad_norm": 3.6875,
      "learning_rate": 2.6680639950773018e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1255974.0,
      "step": 2877
    },
    {
      "epoch": 0.22136758710868396,
      "grad_norm": 8.1875,
      "learning_rate": 2.667948619336974e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1256282.0,
      "step": 2878
    },
    {
      "epoch": 0.2214445042689024,
      "grad_norm": 8.625,
      "learning_rate": 2.6678332435966463e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1256624.0,
      "step": 2879
    },
    {
      "epoch": 0.22152142142912085,
      "grad_norm": 2.71875,
      "learning_rate": 2.667717867856319e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1257208.0,
      "step": 2880
    },
    {
      "epoch": 0.22159833858933928,
      "grad_norm": 6.03125,
      "learning_rate": 2.6676024921159912e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1257514.0,
      "step": 2881
    },
    {
      "epoch": 0.22167525574955774,
      "grad_norm": 4.46875,
      "learning_rate": 2.6674871163756635e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1257824.0,
      "step": 2882
    },
    {
      "epoch": 0.22175217290977617,
      "grad_norm": 4.4375,
      "learning_rate": 2.6673717406353358e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1258148.0,
      "step": 2883
    },
    {
      "epoch": 0.22182909006999463,
      "grad_norm": 3.9375,
      "learning_rate": 2.667256364895008e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1258728.0,
      "step": 2884
    },
    {
      "epoch": 0.22190600723021306,
      "grad_norm": 3.09375,
      "learning_rate": 2.6671409891546807e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1259250.0,
      "step": 2885
    },
    {
      "epoch": 0.22198292439043152,
      "grad_norm": 5.6875,
      "learning_rate": 2.667025613414353e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1259532.0,
      "step": 2886
    },
    {
      "epoch": 0.22205984155064995,
      "grad_norm": 3.0,
      "learning_rate": 2.666910237674025e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1260096.0,
      "step": 2887
    },
    {
      "epoch": 0.22213675871086838,
      "grad_norm": 2.546875,
      "learning_rate": 2.6667948619336975e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1260916.0,
      "step": 2888
    },
    {
      "epoch": 0.22221367587108684,
      "grad_norm": 2.734375,
      "learning_rate": 2.6666794861933697e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1261440.0,
      "step": 2889
    },
    {
      "epoch": 0.22229059303130527,
      "grad_norm": 5.65625,
      "learning_rate": 2.666564110453042e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1261732.0,
      "step": 2890
    },
    {
      "epoch": 0.22236751019152373,
      "grad_norm": 6.28125,
      "learning_rate": 2.6664487347127147e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1262060.0,
      "step": 2891
    },
    {
      "epoch": 0.22244442735174216,
      "grad_norm": 3.265625,
      "learning_rate": 2.666333358972387e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1262608.0,
      "step": 2892
    },
    {
      "epoch": 0.22252134451196062,
      "grad_norm": 4.59375,
      "learning_rate": 2.666217983232059e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 1262930.0,
      "step": 2893
    },
    {
      "epoch": 0.22259826167217905,
      "grad_norm": 5.84375,
      "learning_rate": 2.6661026074917314e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 1263262.0,
      "step": 2894
    },
    {
      "epoch": 0.2226751788323975,
      "grad_norm": 2.390625,
      "learning_rate": 2.6659872317514037e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1263824.0,
      "step": 2895
    },
    {
      "epoch": 0.22275209599261594,
      "grad_norm": 2.46875,
      "learning_rate": 2.6658718560110764e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1264404.0,
      "step": 2896
    },
    {
      "epoch": 0.2228290131528344,
      "grad_norm": 6.78125,
      "learning_rate": 2.6657564802707486e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1264690.0,
      "step": 2897
    },
    {
      "epoch": 0.22290593031305284,
      "grad_norm": 1.96875,
      "learning_rate": 2.6656411045304205e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 1265252.0,
      "step": 2898
    },
    {
      "epoch": 0.2229828474732713,
      "grad_norm": 3.765625,
      "learning_rate": 2.665525728790093e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 1265844.0,
      "step": 2899
    },
    {
      "epoch": 0.22305976463348973,
      "grad_norm": 3.71875,
      "learning_rate": 2.6654103530497654e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 1266408.0,
      "step": 2900
    },
    {
      "epoch": 0.22313668179370819,
      "grad_norm": 6.28125,
      "learning_rate": 2.6652949773094377e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 1266722.0,
      "step": 2901
    },
    {
      "epoch": 0.22321359895392662,
      "grad_norm": 4.09375,
      "learning_rate": 2.6651796015691103e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1267044.0,
      "step": 2902
    },
    {
      "epoch": 0.22329051611414508,
      "grad_norm": 2.9375,
      "learning_rate": 2.6650642258287826e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 1267618.0,
      "step": 2903
    },
    {
      "epoch": 0.2233674332743635,
      "grad_norm": 3.5625,
      "learning_rate": 2.664948850088455e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1268144.0,
      "step": 2904
    },
    {
      "epoch": 0.22344435043458197,
      "grad_norm": 4.28125,
      "learning_rate": 2.664833474348127e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 1268660.0,
      "step": 2905
    },
    {
      "epoch": 0.2235212675948004,
      "grad_norm": 1.84375,
      "learning_rate": 2.6647180986077994e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9975728392601013,
      "num_tokens": 1269486.0,
      "step": 2906
    },
    {
      "epoch": 0.22359818475501886,
      "grad_norm": 4.5,
      "learning_rate": 2.664602722867472e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1269830.0,
      "step": 2907
    },
    {
      "epoch": 0.2236751019152373,
      "grad_norm": 2.921875,
      "learning_rate": 2.6644873471271443e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1270374.0,
      "step": 2908
    },
    {
      "epoch": 0.22375201907545572,
      "grad_norm": 7.625,
      "learning_rate": 2.664371971386816e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1270658.0,
      "step": 2909
    },
    {
      "epoch": 0.22382893623567418,
      "grad_norm": 2.3125,
      "learning_rate": 2.664256595646489e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1271232.0,
      "step": 2910
    },
    {
      "epoch": 0.2239058533958926,
      "grad_norm": 4.0,
      "learning_rate": 2.664141219906161e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1271582.0,
      "step": 2911
    },
    {
      "epoch": 0.22398277055611107,
      "grad_norm": 2.515625,
      "learning_rate": 2.6640258441658334e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1272160.0,
      "step": 2912
    },
    {
      "epoch": 0.2240596877163295,
      "grad_norm": 1.828125,
      "learning_rate": 2.663910468425506e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1272744.0,
      "step": 2913
    },
    {
      "epoch": 0.22413660487654796,
      "grad_norm": 8.125,
      "learning_rate": 2.6637950926851783e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1273028.0,
      "step": 2914
    },
    {
      "epoch": 0.2242135220367664,
      "grad_norm": 9.125,
      "learning_rate": 2.6636797169448506e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 1273288.0,
      "step": 2915
    },
    {
      "epoch": 0.22429043919698485,
      "grad_norm": 7.8125,
      "learning_rate": 2.663564341204523e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1273618.0,
      "step": 2916
    },
    {
      "epoch": 0.22436735635720328,
      "grad_norm": 7.0625,
      "learning_rate": 2.663448965464195e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1273900.0,
      "step": 2917
    },
    {
      "epoch": 0.22444427351742174,
      "grad_norm": 2.40625,
      "learning_rate": 2.6633335897238673e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1274448.0,
      "step": 2918
    },
    {
      "epoch": 0.22452119067764018,
      "grad_norm": 5.75,
      "learning_rate": 2.66321821398354e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1274758.0,
      "step": 2919
    },
    {
      "epoch": 0.22459810783785863,
      "grad_norm": 3.859375,
      "learning_rate": 2.663102838243212e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1275326.0,
      "step": 2920
    },
    {
      "epoch": 0.22467502499807707,
      "grad_norm": 4.5625,
      "learning_rate": 2.6629874625028845e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1275670.0,
      "step": 2921
    },
    {
      "epoch": 0.22475194215829553,
      "grad_norm": 2.390625,
      "learning_rate": 2.6628720867625568e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 1276474.0,
      "step": 2922
    },
    {
      "epoch": 0.22482885931851396,
      "grad_norm": 3.8125,
      "learning_rate": 2.662756711022229e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1276992.0,
      "step": 2923
    },
    {
      "epoch": 0.22490577647873242,
      "grad_norm": 6.6875,
      "learning_rate": 2.6626413352819017e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1277270.0,
      "step": 2924
    },
    {
      "epoch": 0.22498269363895085,
      "grad_norm": 5.4375,
      "learning_rate": 2.6625259595415736e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1277592.0,
      "step": 2925
    },
    {
      "epoch": 0.2250596107991693,
      "grad_norm": 3.34375,
      "learning_rate": 2.662410583801246e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1278182.0,
      "step": 2926
    },
    {
      "epoch": 0.22513652795938774,
      "grad_norm": 5.3125,
      "learning_rate": 2.6622952080609185e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1278470.0,
      "step": 2927
    },
    {
      "epoch": 0.22521344511960617,
      "grad_norm": 6.65625,
      "learning_rate": 2.6621798323205908e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 1278826.0,
      "step": 2928
    },
    {
      "epoch": 0.22529036227982463,
      "grad_norm": 6.0625,
      "learning_rate": 2.662064456580263e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 1279126.0,
      "step": 2929
    },
    {
      "epoch": 0.22536727944004306,
      "grad_norm": 7.03125,
      "learning_rate": 2.6619490808399357e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 1279482.0,
      "step": 2930
    },
    {
      "epoch": 0.22544419660026152,
      "grad_norm": 7.40625,
      "learning_rate": 2.6618337050996075e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1279744.0,
      "step": 2931
    },
    {
      "epoch": 0.22552111376047995,
      "grad_norm": 6.375,
      "learning_rate": 2.66171832935928e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1280084.0,
      "step": 2932
    },
    {
      "epoch": 0.2255980309206984,
      "grad_norm": 2.140625,
      "learning_rate": 2.6616029536189525e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1280630.0,
      "step": 2933
    },
    {
      "epoch": 0.22567494808091684,
      "grad_norm": 3.453125,
      "learning_rate": 2.6614875778786247e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 1281152.0,
      "step": 2934
    },
    {
      "epoch": 0.2257518652411353,
      "grad_norm": 4.3125,
      "learning_rate": 2.6613722021382974e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1281476.0,
      "step": 2935
    },
    {
      "epoch": 0.22582878240135373,
      "grad_norm": 4.3125,
      "learning_rate": 2.6612568263979692e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1281788.0,
      "step": 2936
    },
    {
      "epoch": 0.2259056995615722,
      "grad_norm": 4.375,
      "learning_rate": 2.6611414506576415e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 1282114.0,
      "step": 2937
    },
    {
      "epoch": 0.22598261672179062,
      "grad_norm": 2.734375,
      "learning_rate": 2.661026074917314e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1282694.0,
      "step": 2938
    },
    {
      "epoch": 0.22605953388200908,
      "grad_norm": 5.875,
      "learning_rate": 2.6609106991769864e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1283040.0,
      "step": 2939
    },
    {
      "epoch": 0.22613645104222752,
      "grad_norm": 2.375,
      "learning_rate": 2.6607953234366587e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1283592.0,
      "step": 2940
    },
    {
      "epoch": 0.22621336820244597,
      "grad_norm": 9.375,
      "learning_rate": 2.6606799476963314e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 1283846.0,
      "step": 2941
    },
    {
      "epoch": 0.2262902853626644,
      "grad_norm": 5.3125,
      "learning_rate": 2.6605645719560032e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1284180.0,
      "step": 2942
    },
    {
      "epoch": 0.22636720252288287,
      "grad_norm": 6.78125,
      "learning_rate": 2.660449196215676e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9786477088928223,
      "num_tokens": 1284462.0,
      "step": 2943
    },
    {
      "epoch": 0.2264441196831013,
      "grad_norm": 2.9375,
      "learning_rate": 2.660333820475348e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1285046.0,
      "step": 2944
    },
    {
      "epoch": 0.22652103684331976,
      "grad_norm": 3.578125,
      "learning_rate": 2.6602184447350204e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1285370.0,
      "step": 2945
    },
    {
      "epoch": 0.2265979540035382,
      "grad_norm": 2.765625,
      "learning_rate": 2.660103068994693e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1285950.0,
      "step": 2946
    },
    {
      "epoch": 0.22667487116375665,
      "grad_norm": 2.421875,
      "learning_rate": 2.659987693254365e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 1286746.0,
      "step": 2947
    },
    {
      "epoch": 0.22675178832397508,
      "grad_norm": 12.4375,
      "learning_rate": 2.659872317514037e-06,
      "loss": 0.0415,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 1286978.0,
      "step": 2948
    },
    {
      "epoch": 0.2268287054841935,
      "grad_norm": 4.25,
      "learning_rate": 2.65975694177371e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1287290.0,
      "step": 2949
    },
    {
      "epoch": 0.22690562264441197,
      "grad_norm": 3.328125,
      "learning_rate": 2.659641566033382e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 1287866.0,
      "step": 2950
    },
    {
      "epoch": 0.2269825398046304,
      "grad_norm": 3.921875,
      "learning_rate": 2.6595261902930544e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1288188.0,
      "step": 2951
    },
    {
      "epoch": 0.22705945696484886,
      "grad_norm": 3.5,
      "learning_rate": 2.659410814552727e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1288708.0,
      "step": 2952
    },
    {
      "epoch": 0.2271363741250673,
      "grad_norm": 2.234375,
      "learning_rate": 2.659295438812399e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1289264.0,
      "step": 2953
    },
    {
      "epoch": 0.22721329128528575,
      "grad_norm": 5.625,
      "learning_rate": 2.6591800630720716e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1289594.0,
      "step": 2954
    },
    {
      "epoch": 0.22729020844550418,
      "grad_norm": 3.109375,
      "learning_rate": 2.659064687331744e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 1290162.0,
      "step": 2955
    },
    {
      "epoch": 0.22736712560572264,
      "grad_norm": 6.28125,
      "learning_rate": 2.658949311591416e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1290458.0,
      "step": 2956
    },
    {
      "epoch": 0.22744404276594107,
      "grad_norm": 3.6875,
      "learning_rate": 2.6588339358510888e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1291022.0,
      "step": 2957
    },
    {
      "epoch": 0.22752095992615953,
      "grad_norm": 3.140625,
      "learning_rate": 2.6587185601107606e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1291540.0,
      "step": 2958
    },
    {
      "epoch": 0.22759787708637796,
      "grad_norm": 2.5,
      "learning_rate": 2.658603184370433e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1292104.0,
      "step": 2959
    },
    {
      "epoch": 0.22767479424659642,
      "grad_norm": 5.53125,
      "learning_rate": 2.6584878086301056e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 1292456.0,
      "step": 2960
    },
    {
      "epoch": 0.22775171140681486,
      "grad_norm": 3.421875,
      "learning_rate": 2.658372432889778e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1293032.0,
      "step": 2961
    },
    {
      "epoch": 0.22782862856703331,
      "grad_norm": 3.5625,
      "learning_rate": 2.65825705714945e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 1293634.0,
      "step": 2962
    },
    {
      "epoch": 0.22790554572725175,
      "grad_norm": 4.65625,
      "learning_rate": 2.6581416814091223e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1293940.0,
      "step": 2963
    },
    {
      "epoch": 0.2279824628874702,
      "grad_norm": 5.25,
      "learning_rate": 2.6580263056687946e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1294246.0,
      "step": 2964
    },
    {
      "epoch": 0.22805938004768864,
      "grad_norm": 9.625,
      "learning_rate": 2.6579109299284673e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1294524.0,
      "step": 2965
    },
    {
      "epoch": 0.2281362972079071,
      "grad_norm": 2.53125,
      "learning_rate": 2.6577955541881395e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 1295348.0,
      "step": 2966
    },
    {
      "epoch": 0.22821321436812553,
      "grad_norm": 4.21875,
      "learning_rate": 2.6576801784478118e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1295868.0,
      "step": 2967
    },
    {
      "epoch": 0.228290131528344,
      "grad_norm": 9.875,
      "learning_rate": 2.657564802707484e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 1296176.0,
      "step": 2968
    },
    {
      "epoch": 0.22836704868856242,
      "grad_norm": 7.5625,
      "learning_rate": 2.6574494269671563e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 1296530.0,
      "step": 2969
    },
    {
      "epoch": 0.22844396584878085,
      "grad_norm": 6.53125,
      "learning_rate": 2.6573340512268286e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 1296832.0,
      "step": 2970
    },
    {
      "epoch": 0.2285208830089993,
      "grad_norm": 2.046875,
      "learning_rate": 2.6572186754865012e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1297394.0,
      "step": 2971
    },
    {
      "epoch": 0.22859780016921774,
      "grad_norm": 6.0,
      "learning_rate": 2.6571032997461735e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1297688.0,
      "step": 2972
    },
    {
      "epoch": 0.2286747173294362,
      "grad_norm": 5.125,
      "learning_rate": 2.6569879240058458e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1298008.0,
      "step": 2973
    },
    {
      "epoch": 0.22875163448965463,
      "grad_norm": 6.34375,
      "learning_rate": 2.656872548265518e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1298338.0,
      "step": 2974
    },
    {
      "epoch": 0.2288285516498731,
      "grad_norm": 5.6875,
      "learning_rate": 2.6567571725251903e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 1298694.0,
      "step": 2975
    },
    {
      "epoch": 0.22890546881009152,
      "grad_norm": 2.421875,
      "learning_rate": 2.656641796784863e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1299228.0,
      "step": 2976
    },
    {
      "epoch": 0.22898238597030998,
      "grad_norm": 2.140625,
      "learning_rate": 2.656526421044535e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 1300032.0,
      "step": 2977
    },
    {
      "epoch": 0.2290593031305284,
      "grad_norm": 7.78125,
      "learning_rate": 2.6564110453042075e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9788732528686523,
      "num_tokens": 1300318.0,
      "step": 2978
    },
    {
      "epoch": 0.22913622029074687,
      "grad_norm": 2.828125,
      "learning_rate": 2.6562956695638797e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1300892.0,
      "step": 2979
    },
    {
      "epoch": 0.2292131374509653,
      "grad_norm": 2.734375,
      "learning_rate": 2.656180293823552e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1301436.0,
      "step": 2980
    },
    {
      "epoch": 0.22929005461118376,
      "grad_norm": 6.46875,
      "learning_rate": 2.6560649180832242e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 1301748.0,
      "step": 2981
    },
    {
      "epoch": 0.2293669717714022,
      "grad_norm": 6.5,
      "learning_rate": 2.655949542342897e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 1302064.0,
      "step": 2982
    },
    {
      "epoch": 0.22944388893162065,
      "grad_norm": 7.28125,
      "learning_rate": 2.655834166602569e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9941520690917969,
      "num_tokens": 1302408.0,
      "step": 2983
    },
    {
      "epoch": 0.22952080609183909,
      "grad_norm": 6.0625,
      "learning_rate": 2.6557187908622414e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 1302764.0,
      "step": 2984
    },
    {
      "epoch": 0.22959772325205755,
      "grad_norm": 5.15625,
      "learning_rate": 2.6556034151219137e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1303048.0,
      "step": 2985
    },
    {
      "epoch": 0.22967464041227598,
      "grad_norm": 2.953125,
      "learning_rate": 2.655488039381586e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1303374.0,
      "step": 2986
    },
    {
      "epoch": 0.22975155757249444,
      "grad_norm": 5.9375,
      "learning_rate": 2.6553726636412582e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1303708.0,
      "step": 2987
    },
    {
      "epoch": 0.22982847473271287,
      "grad_norm": 4.1875,
      "learning_rate": 2.655257287900931e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1304244.0,
      "step": 2988
    },
    {
      "epoch": 0.22990539189293133,
      "grad_norm": 5.3125,
      "learning_rate": 2.655141912160603e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1304542.0,
      "step": 2989
    },
    {
      "epoch": 0.22998230905314976,
      "grad_norm": 3.5,
      "learning_rate": 2.6550265364202754e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1304870.0,
      "step": 2990
    },
    {
      "epoch": 0.2300592262133682,
      "grad_norm": 3.28125,
      "learning_rate": 2.6549111606799477e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 1305438.0,
      "step": 2991
    },
    {
      "epoch": 0.23013614337358665,
      "grad_norm": 2.03125,
      "learning_rate": 2.65479578493962e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1306012.0,
      "step": 2992
    },
    {
      "epoch": 0.23021306053380508,
      "grad_norm": 2.890625,
      "learning_rate": 2.6546804091992926e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 1306600.0,
      "step": 2993
    },
    {
      "epoch": 0.23028997769402354,
      "grad_norm": 8.4375,
      "learning_rate": 2.654565033458965e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1306928.0,
      "step": 2994
    },
    {
      "epoch": 0.23036689485424197,
      "grad_norm": 4.125,
      "learning_rate": 2.654449657718637e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 1307244.0,
      "step": 2995
    },
    {
      "epoch": 0.23044381201446043,
      "grad_norm": 4.3125,
      "learning_rate": 2.6543342819783094e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1307566.0,
      "step": 2996
    },
    {
      "epoch": 0.23052072917467886,
      "grad_norm": 2.5625,
      "learning_rate": 2.6542189062379816e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1308128.0,
      "step": 2997
    },
    {
      "epoch": 0.23059764633489732,
      "grad_norm": 7.28125,
      "learning_rate": 2.654103530497654e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1308414.0,
      "step": 2998
    },
    {
      "epoch": 0.23067456349511575,
      "grad_norm": 6.0625,
      "learning_rate": 2.6539881547573266e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1308776.0,
      "step": 2999
    },
    {
      "epoch": 0.2307514806553342,
      "grad_norm": 9.9375,
      "learning_rate": 2.653872779016999e-06,
      "loss": 0.0394,
      "step": 3000
    },
    {
      "epoch": 0.2307514806553342,
      "eval_loss": 0.02130208909511566,
      "eval_mean_token_accuracy": 0.9912069249153137,
      "eval_num_tokens": 1309032.0,
      "eval_runtime": 21.0582,
      "eval_samples_per_second": 9.497,
      "eval_steps_per_second": 1.187,
      "step": 3000
    },
    {
      "epoch": 0.23082839781555264,
      "grad_norm": 3.390625,
      "learning_rate": 2.653757403276671e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9884600043296814,
      "num_tokens": 1309574.0,
      "step": 3001
    },
    {
      "epoch": 0.2309053149757711,
      "grad_norm": 4.09375,
      "learning_rate": 2.6536420275363434e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1310090.0,
      "step": 3002
    },
    {
      "epoch": 0.23098223213598953,
      "grad_norm": 6.75,
      "learning_rate": 2.6535266517960156e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1310372.0,
      "step": 3003
    },
    {
      "epoch": 0.231059149296208,
      "grad_norm": 3.3125,
      "learning_rate": 2.6534112760556883e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1310948.0,
      "step": 3004
    },
    {
      "epoch": 0.23113606645642643,
      "grad_norm": 3.96875,
      "learning_rate": 2.6532959003153606e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1311442.0,
      "step": 3005
    },
    {
      "epoch": 0.23121298361664488,
      "grad_norm": 3.25,
      "learning_rate": 2.653180524575033e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1312022.0,
      "step": 3006
    },
    {
      "epoch": 0.23128990077686332,
      "grad_norm": 4.1875,
      "learning_rate": 2.653065148834705e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1312372.0,
      "step": 3007
    },
    {
      "epoch": 0.23136681793708178,
      "grad_norm": 2.390625,
      "learning_rate": 2.6529497730943773e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 1313422.0,
      "step": 3008
    },
    {
      "epoch": 0.2314437350973002,
      "grad_norm": 3.984375,
      "learning_rate": 2.6528343973540496e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1313996.0,
      "step": 3009
    },
    {
      "epoch": 0.23152065225751864,
      "grad_norm": 6.625,
      "learning_rate": 2.6527190216137223e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 1314268.0,
      "step": 3010
    },
    {
      "epoch": 0.2315975694177371,
      "grad_norm": 4.0,
      "learning_rate": 2.6526036458733945e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1314618.0,
      "step": 3011
    },
    {
      "epoch": 0.23167448657795553,
      "grad_norm": 2.03125,
      "learning_rate": 2.6524882701330668e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1315404.0,
      "step": 3012
    },
    {
      "epoch": 0.231751403738174,
      "grad_norm": 2.109375,
      "learning_rate": 2.652372894392739e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1315954.0,
      "step": 3013
    },
    {
      "epoch": 0.23182832089839242,
      "grad_norm": 7.1875,
      "learning_rate": 2.6522575186524113e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1316238.0,
      "step": 3014
    },
    {
      "epoch": 0.23190523805861088,
      "grad_norm": 2.484375,
      "learning_rate": 2.652142142912084e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1316824.0,
      "step": 3015
    },
    {
      "epoch": 0.2319821552188293,
      "grad_norm": 3.0,
      "learning_rate": 2.6520267671717562e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9907235503196716,
      "num_tokens": 1317364.0,
      "step": 3016
    },
    {
      "epoch": 0.23205907237904777,
      "grad_norm": 3.34375,
      "learning_rate": 2.6519113914314285e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 1317888.0,
      "step": 3017
    },
    {
      "epoch": 0.2321359895392662,
      "grad_norm": 4.90625,
      "learning_rate": 2.6517960156911008e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1318182.0,
      "step": 3018
    },
    {
      "epoch": 0.23221290669948466,
      "grad_norm": 5.875,
      "learning_rate": 2.651680639950773e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1318494.0,
      "step": 3019
    },
    {
      "epoch": 0.2322898238597031,
      "grad_norm": 9.9375,
      "learning_rate": 2.6515652642104453e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1318796.0,
      "step": 3020
    },
    {
      "epoch": 0.23236674101992155,
      "grad_norm": 6.28125,
      "learning_rate": 2.651449888470118e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9817629456520081,
      "num_tokens": 1319126.0,
      "step": 3021
    },
    {
      "epoch": 0.23244365818013998,
      "grad_norm": 7.0,
      "learning_rate": 2.6513345127297902e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 1319434.0,
      "step": 3022
    },
    {
      "epoch": 0.23252057534035844,
      "grad_norm": 3.6875,
      "learning_rate": 2.6512191369894625e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 1320030.0,
      "step": 3023
    },
    {
      "epoch": 0.23259749250057687,
      "grad_norm": 2.703125,
      "learning_rate": 2.6511037612491347e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1320594.0,
      "step": 3024
    },
    {
      "epoch": 0.23267440966079533,
      "grad_norm": 6.40625,
      "learning_rate": 2.650988385508807e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1320864.0,
      "step": 3025
    },
    {
      "epoch": 0.23275132682101377,
      "grad_norm": 2.234375,
      "learning_rate": 2.6508730097684797e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 1321420.0,
      "step": 3026
    },
    {
      "epoch": 0.23282824398123222,
      "grad_norm": 6.875,
      "learning_rate": 2.650757634028152e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 1321724.0,
      "step": 3027
    },
    {
      "epoch": 0.23290516114145066,
      "grad_norm": 4.25,
      "learning_rate": 2.6506422582878238e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1322034.0,
      "step": 3028
    },
    {
      "epoch": 0.23298207830166912,
      "grad_norm": 5.0625,
      "learning_rate": 2.6505268825474964e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1322362.0,
      "step": 3029
    },
    {
      "epoch": 0.23305899546188755,
      "grad_norm": 5.40625,
      "learning_rate": 2.6504115068071687e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 1322674.0,
      "step": 3030
    },
    {
      "epoch": 0.23313591262210598,
      "grad_norm": 7.125,
      "learning_rate": 2.650296131066841e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 1323018.0,
      "step": 3031
    },
    {
      "epoch": 0.23321282978232444,
      "grad_norm": 2.484375,
      "learning_rate": 2.6501807553265136e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1323574.0,
      "step": 3032
    },
    {
      "epoch": 0.23328974694254287,
      "grad_norm": 2.46875,
      "learning_rate": 2.650065379586186e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1324142.0,
      "step": 3033
    },
    {
      "epoch": 0.23336666410276133,
      "grad_norm": 5.5625,
      "learning_rate": 2.649950003845858e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 1324426.0,
      "step": 3034
    },
    {
      "epoch": 0.23344358126297976,
      "grad_norm": 1.953125,
      "learning_rate": 2.6498346281055304e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1324972.0,
      "step": 3035
    },
    {
      "epoch": 0.23352049842319822,
      "grad_norm": 4.65625,
      "learning_rate": 2.6497192523652027e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1325282.0,
      "step": 3036
    },
    {
      "epoch": 0.23359741558341665,
      "grad_norm": 5.90625,
      "learning_rate": 2.649603876624875e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1325622.0,
      "step": 3037
    },
    {
      "epoch": 0.2336743327436351,
      "grad_norm": 5.375,
      "learning_rate": 2.6494885008845476e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 1325954.0,
      "step": 3038
    },
    {
      "epoch": 0.23375124990385354,
      "grad_norm": 2.46875,
      "learning_rate": 2.6493731251442194e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1326712.0,
      "step": 3039
    },
    {
      "epoch": 0.233828167064072,
      "grad_norm": 6.40625,
      "learning_rate": 2.649257749403892e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 1327026.0,
      "step": 3040
    },
    {
      "epoch": 0.23390508422429043,
      "grad_norm": 4.0,
      "learning_rate": 2.6491423736635644e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1327338.0,
      "step": 3041
    },
    {
      "epoch": 0.2339820013845089,
      "grad_norm": 2.53125,
      "learning_rate": 2.6490269979232366e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1327882.0,
      "step": 3042
    },
    {
      "epoch": 0.23405891854472732,
      "grad_norm": 3.359375,
      "learning_rate": 2.6489116221829093e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1328468.0,
      "step": 3043
    },
    {
      "epoch": 0.23413583570494578,
      "grad_norm": 7.09375,
      "learning_rate": 2.6487962464425816e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 1328748.0,
      "step": 3044
    },
    {
      "epoch": 0.23421275286516421,
      "grad_norm": 5.28125,
      "learning_rate": 2.6486808707022534e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1329060.0,
      "step": 3045
    },
    {
      "epoch": 0.23428967002538267,
      "grad_norm": 4.53125,
      "learning_rate": 2.648565494961926e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1329364.0,
      "step": 3046
    },
    {
      "epoch": 0.2343665871856011,
      "grad_norm": 3.625,
      "learning_rate": 2.6484501192215984e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1329676.0,
      "step": 3047
    },
    {
      "epoch": 0.23444350434581956,
      "grad_norm": 4.75,
      "learning_rate": 2.6483347434812706e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 1330258.0,
      "step": 3048
    },
    {
      "epoch": 0.234520421506038,
      "grad_norm": 5.46875,
      "learning_rate": 2.6482193677409433e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1330586.0,
      "step": 3049
    },
    {
      "epoch": 0.23459733866625646,
      "grad_norm": 4.71875,
      "learning_rate": 2.648103992000615e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1330932.0,
      "step": 3050
    },
    {
      "epoch": 0.2346742558264749,
      "grad_norm": 3.734375,
      "learning_rate": 2.647988616260288e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1331464.0,
      "step": 3051
    },
    {
      "epoch": 0.23475117298669332,
      "grad_norm": 2.5625,
      "learning_rate": 2.64787324051996e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 1332246.0,
      "step": 3052
    },
    {
      "epoch": 0.23482809014691178,
      "grad_norm": 8.75,
      "learning_rate": 2.6477578647796323e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 1332518.0,
      "step": 3053
    },
    {
      "epoch": 0.2349050073071302,
      "grad_norm": 2.9375,
      "learning_rate": 2.647642489039305e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1333060.0,
      "step": 3054
    },
    {
      "epoch": 0.23498192446734867,
      "grad_norm": 8.5,
      "learning_rate": 2.6475271132989773e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 1333318.0,
      "step": 3055
    },
    {
      "epoch": 0.2350588416275671,
      "grad_norm": 10.0,
      "learning_rate": 2.647411737558649e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1333620.0,
      "step": 3056
    },
    {
      "epoch": 0.23513575878778556,
      "grad_norm": 2.390625,
      "learning_rate": 2.647296361818322e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1334144.0,
      "step": 3057
    },
    {
      "epoch": 0.235212675948004,
      "grad_norm": 7.0625,
      "learning_rate": 2.647180986077994e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 1334504.0,
      "step": 3058
    },
    {
      "epoch": 0.23528959310822245,
      "grad_norm": 3.125,
      "learning_rate": 2.6470656103376663e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1335030.0,
      "step": 3059
    },
    {
      "epoch": 0.23536651026844088,
      "grad_norm": 2.4375,
      "learning_rate": 2.646950234597339e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1335576.0,
      "step": 3060
    },
    {
      "epoch": 0.23544342742865934,
      "grad_norm": 3.28125,
      "learning_rate": 2.646834858857011e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1336162.0,
      "step": 3061
    },
    {
      "epoch": 0.23552034458887777,
      "grad_norm": 5.8125,
      "learning_rate": 2.6467194831166835e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 1336480.0,
      "step": 3062
    },
    {
      "epoch": 0.23559726174909623,
      "grad_norm": 5.90625,
      "learning_rate": 2.6466041073763558e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9928057789802551,
      "num_tokens": 1336760.0,
      "step": 3063
    },
    {
      "epoch": 0.23567417890931466,
      "grad_norm": 4.25,
      "learning_rate": 2.646488731636028e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1337058.0,
      "step": 3064
    },
    {
      "epoch": 0.23575109606953312,
      "grad_norm": 5.46875,
      "learning_rate": 2.6463733558957007e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 1337384.0,
      "step": 3065
    },
    {
      "epoch": 0.23582801322975155,
      "grad_norm": 5.90625,
      "learning_rate": 2.6462579801553725e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1337692.0,
      "step": 3066
    },
    {
      "epoch": 0.23590493038997,
      "grad_norm": 4.34375,
      "learning_rate": 2.646142604415045e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 1337980.0,
      "step": 3067
    },
    {
      "epoch": 0.23598184755018844,
      "grad_norm": 8.5,
      "learning_rate": 2.6460272286747175e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 1338280.0,
      "step": 3068
    },
    {
      "epoch": 0.2360587647104069,
      "grad_norm": 7.625,
      "learning_rate": 2.6459118529343897e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 1338568.0,
      "step": 3069
    },
    {
      "epoch": 0.23613568187062534,
      "grad_norm": 5.125,
      "learning_rate": 2.645796477194062e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1338850.0,
      "step": 3070
    },
    {
      "epoch": 0.2362125990308438,
      "grad_norm": 4.5625,
      "learning_rate": 2.6456811014537347e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 1339450.0,
      "step": 3071
    },
    {
      "epoch": 0.23628951619106223,
      "grad_norm": 5.78125,
      "learning_rate": 2.6455657257134065e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1339800.0,
      "step": 3072
    },
    {
      "epoch": 0.23636643335128066,
      "grad_norm": 2.828125,
      "learning_rate": 2.645450349973079e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1340318.0,
      "step": 3073
    },
    {
      "epoch": 0.23644335051149912,
      "grad_norm": 5.21875,
      "learning_rate": 2.6453349742327514e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 1340656.0,
      "step": 3074
    },
    {
      "epoch": 0.23652026767171755,
      "grad_norm": 2.359375,
      "learning_rate": 2.6452195984924237e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1341178.0,
      "step": 3075
    },
    {
      "epoch": 0.236597184831936,
      "grad_norm": 3.140625,
      "learning_rate": 2.6451042227520964e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 1341994.0,
      "step": 3076
    },
    {
      "epoch": 0.23667410199215444,
      "grad_norm": 7.34375,
      "learning_rate": 2.6449888470117682e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1342302.0,
      "step": 3077
    },
    {
      "epoch": 0.2367510191523729,
      "grad_norm": 3.296875,
      "learning_rate": 2.6448734712714405e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 1342874.0,
      "step": 3078
    },
    {
      "epoch": 0.23682793631259133,
      "grad_norm": 3.546875,
      "learning_rate": 2.644758095531113e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1343450.0,
      "step": 3079
    },
    {
      "epoch": 0.2369048534728098,
      "grad_norm": 6.40625,
      "learning_rate": 2.6446427197907854e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1343728.0,
      "step": 3080
    },
    {
      "epoch": 0.23698177063302822,
      "grad_norm": 3.453125,
      "learning_rate": 2.6445273440504577e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1344296.0,
      "step": 3081
    },
    {
      "epoch": 0.23705868779324668,
      "grad_norm": 4.71875,
      "learning_rate": 2.6444119683101304e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 1344612.0,
      "step": 3082
    },
    {
      "epoch": 0.2371356049534651,
      "grad_norm": 4.5,
      "learning_rate": 2.644296592569802e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1344962.0,
      "step": 3083
    },
    {
      "epoch": 0.23721252211368357,
      "grad_norm": 4.125,
      "learning_rate": 2.644181216829475e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1345282.0,
      "step": 3084
    },
    {
      "epoch": 0.237289439273902,
      "grad_norm": 3.46875,
      "learning_rate": 2.644065841089147e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1345850.0,
      "step": 3085
    },
    {
      "epoch": 0.23736635643412046,
      "grad_norm": 3.4375,
      "learning_rate": 2.6439504653488194e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1346402.0,
      "step": 3086
    },
    {
      "epoch": 0.2374432735943389,
      "grad_norm": 4.46875,
      "learning_rate": 2.6438350896084916e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 1346716.0,
      "step": 3087
    },
    {
      "epoch": 0.23752019075455735,
      "grad_norm": 2.53125,
      "learning_rate": 2.643719713868164e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1347276.0,
      "step": 3088
    },
    {
      "epoch": 0.23759710791477578,
      "grad_norm": 3.6875,
      "learning_rate": 2.643604338127836e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1347808.0,
      "step": 3089
    },
    {
      "epoch": 0.23767402507499424,
      "grad_norm": 4.375,
      "learning_rate": 2.643488962387509e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1348130.0,
      "step": 3090
    },
    {
      "epoch": 0.23775094223521268,
      "grad_norm": 7.25,
      "learning_rate": 2.643373586647181e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1348432.0,
      "step": 3091
    },
    {
      "epoch": 0.2378278593954311,
      "grad_norm": 3.609375,
      "learning_rate": 2.6432582109068534e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1348974.0,
      "step": 3092
    },
    {
      "epoch": 0.23790477655564957,
      "grad_norm": 6.4375,
      "learning_rate": 2.643142835166526e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1349288.0,
      "step": 3093
    },
    {
      "epoch": 0.237981693715868,
      "grad_norm": 5.71875,
      "learning_rate": 2.643027459426198e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1349600.0,
      "step": 3094
    },
    {
      "epoch": 0.23805861087608646,
      "grad_norm": 4.21875,
      "learning_rate": 2.64291208368587e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1350172.0,
      "step": 3095
    },
    {
      "epoch": 0.2381355280363049,
      "grad_norm": 3.96875,
      "learning_rate": 2.642796707945543e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1350498.0,
      "step": 3096
    },
    {
      "epoch": 0.23821244519652335,
      "grad_norm": 2.3125,
      "learning_rate": 2.642681332205215e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9971292018890381,
      "num_tokens": 1351546.0,
      "step": 3097
    },
    {
      "epoch": 0.23828936235674178,
      "grad_norm": 10.3125,
      "learning_rate": 2.6425659564648873e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1351824.0,
      "step": 3098
    },
    {
      "epoch": 0.23836627951696024,
      "grad_norm": 6.34375,
      "learning_rate": 2.6424505807245596e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1352108.0,
      "step": 3099
    },
    {
      "epoch": 0.23844319667717867,
      "grad_norm": 6.0625,
      "learning_rate": 2.642335204984232e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 1352422.0,
      "step": 3100
    },
    {
      "epoch": 0.23852011383739713,
      "grad_norm": 2.296875,
      "learning_rate": 2.6422198292439045e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1353242.0,
      "step": 3101
    },
    {
      "epoch": 0.23859703099761556,
      "grad_norm": 4.4375,
      "learning_rate": 2.642104453503577e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1353548.0,
      "step": 3102
    },
    {
      "epoch": 0.23867394815783402,
      "grad_norm": 4.03125,
      "learning_rate": 2.641989077763249e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1354068.0,
      "step": 3103
    },
    {
      "epoch": 0.23875086531805245,
      "grad_norm": 6.25,
      "learning_rate": 2.6418737020229213e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 1354428.0,
      "step": 3104
    },
    {
      "epoch": 0.2388277824782709,
      "grad_norm": 3.140625,
      "learning_rate": 2.6417583262825936e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 1354996.0,
      "step": 3105
    },
    {
      "epoch": 0.23890469963848934,
      "grad_norm": 6.9375,
      "learning_rate": 2.641642950542266e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 1355336.0,
      "step": 3106
    },
    {
      "epoch": 0.2389816167987078,
      "grad_norm": 5.375,
      "learning_rate": 2.6415275748019385e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1355624.0,
      "step": 3107
    },
    {
      "epoch": 0.23905853395892623,
      "grad_norm": 6.84375,
      "learning_rate": 2.6414121990616108e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1355910.0,
      "step": 3108
    },
    {
      "epoch": 0.2391354511191447,
      "grad_norm": 4.9375,
      "learning_rate": 2.641296823321283e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1356222.0,
      "step": 3109
    },
    {
      "epoch": 0.23921236827936312,
      "grad_norm": 5.34375,
      "learning_rate": 2.6411814475809553e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 1356514.0,
      "step": 3110
    },
    {
      "epoch": 0.23928928543958158,
      "grad_norm": 2.328125,
      "learning_rate": 2.6410660718406275e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9961722493171692,
      "num_tokens": 1357562.0,
      "step": 3111
    },
    {
      "epoch": 0.23936620259980002,
      "grad_norm": 14.375,
      "learning_rate": 2.6409506961003002e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9821428656578064,
      "num_tokens": 1357790.0,
      "step": 3112
    },
    {
      "epoch": 0.23944311976001845,
      "grad_norm": 2.53125,
      "learning_rate": 2.6408353203599725e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 1358312.0,
      "step": 3113
    },
    {
      "epoch": 0.2395200369202369,
      "grad_norm": 6.4375,
      "learning_rate": 2.6407199446196447e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 1358642.0,
      "step": 3114
    },
    {
      "epoch": 0.23959695408045534,
      "grad_norm": 2.59375,
      "learning_rate": 2.640604568879317e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1359210.0,
      "step": 3115
    },
    {
      "epoch": 0.2396738712406738,
      "grad_norm": 6.21875,
      "learning_rate": 2.6404891931389893e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1359522.0,
      "step": 3116
    },
    {
      "epoch": 0.23975078840089223,
      "grad_norm": 4.28125,
      "learning_rate": 2.6403738173986615e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1359820.0,
      "step": 3117
    },
    {
      "epoch": 0.2398277055611107,
      "grad_norm": 6.15625,
      "learning_rate": 2.640258441658334e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1360132.0,
      "step": 3118
    },
    {
      "epoch": 0.23990462272132912,
      "grad_norm": 5.46875,
      "learning_rate": 2.6401430659180064e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1360438.0,
      "step": 3119
    },
    {
      "epoch": 0.23998153988154758,
      "grad_norm": 4.9375,
      "learning_rate": 2.6400276901776787e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 1361030.0,
      "step": 3120
    },
    {
      "epoch": 0.240058457041766,
      "grad_norm": 5.84375,
      "learning_rate": 2.639912314437351e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 1361368.0,
      "step": 3121
    },
    {
      "epoch": 0.24013537420198447,
      "grad_norm": 2.296875,
      "learning_rate": 2.6397969386970232e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1361914.0,
      "step": 3122
    },
    {
      "epoch": 0.2402122913622029,
      "grad_norm": 6.3125,
      "learning_rate": 2.639681562956696e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1362198.0,
      "step": 3123
    },
    {
      "epoch": 0.24028920852242136,
      "grad_norm": 2.671875,
      "learning_rate": 2.639566187216368e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 1362758.0,
      "step": 3124
    },
    {
      "epoch": 0.2403661256826398,
      "grad_norm": 5.6875,
      "learning_rate": 2.6394508114760404e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 1363066.0,
      "step": 3125
    },
    {
      "epoch": 0.24044304284285825,
      "grad_norm": 5.65625,
      "learning_rate": 2.6393354357357127e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 1363380.0,
      "step": 3126
    },
    {
      "epoch": 0.24051996000307668,
      "grad_norm": 2.609375,
      "learning_rate": 2.639220059995385e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1363924.0,
      "step": 3127
    },
    {
      "epoch": 0.24059687716329514,
      "grad_norm": 3.984375,
      "learning_rate": 2.639104684255057e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1364420.0,
      "step": 3128
    },
    {
      "epoch": 0.24067379432351357,
      "grad_norm": 8.75,
      "learning_rate": 2.63898930851473e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 1364778.0,
      "step": 3129
    },
    {
      "epoch": 0.24075071148373203,
      "grad_norm": 7.9375,
      "learning_rate": 2.638873932774402e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1365108.0,
      "step": 3130
    },
    {
      "epoch": 0.24082762864395046,
      "grad_norm": 2.234375,
      "learning_rate": 2.6387585570340744e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 1365918.0,
      "step": 3131
    },
    {
      "epoch": 0.24090454580416892,
      "grad_norm": 2.875,
      "learning_rate": 2.6386431812937467e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1366462.0,
      "step": 3132
    },
    {
      "epoch": 0.24098146296438736,
      "grad_norm": 4.8125,
      "learning_rate": 2.638527805553419e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1366806.0,
      "step": 3133
    },
    {
      "epoch": 0.2410583801246058,
      "grad_norm": 7.25,
      "learning_rate": 2.6384124298130916e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1367168.0,
      "step": 3134
    },
    {
      "epoch": 0.24113529728482425,
      "grad_norm": 5.40625,
      "learning_rate": 2.638297054072764e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 1367516.0,
      "step": 3135
    },
    {
      "epoch": 0.24121221444504268,
      "grad_norm": 6.1875,
      "learning_rate": 2.638181678332436e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1367824.0,
      "step": 3136
    },
    {
      "epoch": 0.24128913160526114,
      "grad_norm": 1.859375,
      "learning_rate": 2.6380663025921084e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1368398.0,
      "step": 3137
    },
    {
      "epoch": 0.24136604876547957,
      "grad_norm": 3.78125,
      "learning_rate": 2.6379509268517806e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 1368998.0,
      "step": 3138
    },
    {
      "epoch": 0.24144296592569803,
      "grad_norm": 4.34375,
      "learning_rate": 2.637835551111453e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1369326.0,
      "step": 3139
    },
    {
      "epoch": 0.24151988308591646,
      "grad_norm": 2.953125,
      "learning_rate": 2.6377201753711256e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1369872.0,
      "step": 3140
    },
    {
      "epoch": 0.24159680024613492,
      "grad_norm": 6.40625,
      "learning_rate": 2.637604799630798e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1370202.0,
      "step": 3141
    },
    {
      "epoch": 0.24167371740635335,
      "grad_norm": 8.5625,
      "learning_rate": 2.63748942389047e-06,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1370472.0,
      "step": 3142
    },
    {
      "epoch": 0.2417506345665718,
      "grad_norm": 9.1875,
      "learning_rate": 2.6373740481501423e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1370754.0,
      "step": 3143
    },
    {
      "epoch": 0.24182755172679024,
      "grad_norm": 2.484375,
      "learning_rate": 2.6372586724098146e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 1371776.0,
      "step": 3144
    },
    {
      "epoch": 0.2419044688870087,
      "grad_norm": 5.3125,
      "learning_rate": 2.637143296669487e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 1372094.0,
      "step": 3145
    },
    {
      "epoch": 0.24198138604722713,
      "grad_norm": 6.65625,
      "learning_rate": 2.6370279209291595e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1372434.0,
      "step": 3146
    },
    {
      "epoch": 0.2420583032074456,
      "grad_norm": 2.15625,
      "learning_rate": 2.636912545188832e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1372996.0,
      "step": 3147
    },
    {
      "epoch": 0.24213522036766402,
      "grad_norm": 10.75,
      "learning_rate": 2.636797169448504e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.984375,
      "num_tokens": 1373320.0,
      "step": 3148
    },
    {
      "epoch": 0.24221213752788248,
      "grad_norm": 5.375,
      "learning_rate": 2.6366817937081763e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 1373902.0,
      "step": 3149
    },
    {
      "epoch": 0.2422890546881009,
      "grad_norm": 5.03125,
      "learning_rate": 2.6365664179678486e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 1374238.0,
      "step": 3150
    },
    {
      "epoch": 0.24236597184831937,
      "grad_norm": 2.125,
      "learning_rate": 2.6364510422275213e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 1375062.0,
      "step": 3151
    },
    {
      "epoch": 0.2424428890085378,
      "grad_norm": 7.90625,
      "learning_rate": 2.6363356664871935e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 1375352.0,
      "step": 3152
    },
    {
      "epoch": 0.24251980616875624,
      "grad_norm": 4.90625,
      "learning_rate": 2.6362202907468653e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1375664.0,
      "step": 3153
    },
    {
      "epoch": 0.2425967233289747,
      "grad_norm": 6.6875,
      "learning_rate": 2.636104915006538e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 1376002.0,
      "step": 3154
    },
    {
      "epoch": 0.24267364048919313,
      "grad_norm": 2.109375,
      "learning_rate": 2.6359895392662103e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 1376810.0,
      "step": 3155
    },
    {
      "epoch": 0.24275055764941159,
      "grad_norm": 10.375,
      "learning_rate": 2.6358741635258825e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 1377064.0,
      "step": 3156
    },
    {
      "epoch": 0.24282747480963002,
      "grad_norm": 7.625,
      "learning_rate": 2.6357587877855552e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1377372.0,
      "step": 3157
    },
    {
      "epoch": 0.24290439196984848,
      "grad_norm": 2.21875,
      "learning_rate": 2.6356434120452275e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 1378186.0,
      "step": 3158
    },
    {
      "epoch": 0.2429813091300669,
      "grad_norm": 3.75,
      "learning_rate": 2.6355280363048997e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1378750.0,
      "step": 3159
    },
    {
      "epoch": 0.24305822629028537,
      "grad_norm": 3.875,
      "learning_rate": 2.635412660564572e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1379320.0,
      "step": 3160
    },
    {
      "epoch": 0.2431351434505038,
      "grad_norm": 7.5,
      "learning_rate": 2.6352972848242443e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1379654.0,
      "step": 3161
    },
    {
      "epoch": 0.24321206061072226,
      "grad_norm": 2.546875,
      "learning_rate": 2.635181909083917e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 1380714.0,
      "step": 3162
    },
    {
      "epoch": 0.2432889777709407,
      "grad_norm": 2.546875,
      "learning_rate": 2.635066533343589e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1381274.0,
      "step": 3163
    },
    {
      "epoch": 0.24336589493115915,
      "grad_norm": 10.4375,
      "learning_rate": 2.634951157603261e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 1381622.0,
      "step": 3164
    },
    {
      "epoch": 0.24344281209137758,
      "grad_norm": 4.875,
      "learning_rate": 2.6348357818629337e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1381942.0,
      "step": 3165
    },
    {
      "epoch": 0.24351972925159604,
      "grad_norm": 7.46875,
      "learning_rate": 2.634720406122606e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1382304.0,
      "step": 3166
    },
    {
      "epoch": 0.24359664641181447,
      "grad_norm": 2.71875,
      "learning_rate": 2.6346050303822782e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1382848.0,
      "step": 3167
    },
    {
      "epoch": 0.24367356357203293,
      "grad_norm": 9.3125,
      "learning_rate": 2.634489654641951e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 1383206.0,
      "step": 3168
    },
    {
      "epoch": 0.24375048073225136,
      "grad_norm": 8.875,
      "learning_rate": 2.6343742789016227e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 1383486.0,
      "step": 3169
    },
    {
      "epoch": 0.24382739789246982,
      "grad_norm": 4.09375,
      "learning_rate": 2.6342589031612954e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1383808.0,
      "step": 3170
    },
    {
      "epoch": 0.24390431505268825,
      "grad_norm": 8.75,
      "learning_rate": 2.6341435274209677e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1384108.0,
      "step": 3171
    },
    {
      "epoch": 0.2439812322129067,
      "grad_norm": 4.5,
      "learning_rate": 2.63402815168064e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1384442.0,
      "step": 3172
    },
    {
      "epoch": 0.24405814937312514,
      "grad_norm": 5.125,
      "learning_rate": 2.6339127759403126e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1384754.0,
      "step": 3173
    },
    {
      "epoch": 0.24413506653334358,
      "grad_norm": 2.640625,
      "learning_rate": 2.633797400199985e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 1385300.0,
      "step": 3174
    },
    {
      "epoch": 0.24421198369356203,
      "grad_norm": 3.96875,
      "learning_rate": 2.6336820244596567e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1385820.0,
      "step": 3175
    },
    {
      "epoch": 0.24428890085378047,
      "grad_norm": 2.671875,
      "learning_rate": 2.6335666487193294e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 1386410.0,
      "step": 3176
    },
    {
      "epoch": 0.24436581801399893,
      "grad_norm": 2.265625,
      "learning_rate": 2.6334512729790017e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1387190.0,
      "step": 3177
    },
    {
      "epoch": 0.24444273517421736,
      "grad_norm": 3.484375,
      "learning_rate": 2.633335897238674e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1387766.0,
      "step": 3178
    },
    {
      "epoch": 0.24451965233443582,
      "grad_norm": 3.84375,
      "learning_rate": 2.6332205214983466e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 1388088.0,
      "step": 3179
    },
    {
      "epoch": 0.24459656949465425,
      "grad_norm": 8.0,
      "learning_rate": 2.6331051457580184e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1388348.0,
      "step": 3180
    },
    {
      "epoch": 0.2446734866548727,
      "grad_norm": 4.78125,
      "learning_rate": 2.632989770017691e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1388654.0,
      "step": 3181
    },
    {
      "epoch": 0.24475040381509114,
      "grad_norm": 4.96875,
      "learning_rate": 2.6328743942773634e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1388954.0,
      "step": 3182
    },
    {
      "epoch": 0.2448273209753096,
      "grad_norm": 2.296875,
      "learning_rate": 2.6327590185370356e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 1389500.0,
      "step": 3183
    },
    {
      "epoch": 0.24490423813552803,
      "grad_norm": 6.375,
      "learning_rate": 2.6326436427967083e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 1389834.0,
      "step": 3184
    },
    {
      "epoch": 0.2449811552957465,
      "grad_norm": 2.8125,
      "learning_rate": 2.6325282670563806e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1390420.0,
      "step": 3185
    },
    {
      "epoch": 0.24505807245596492,
      "grad_norm": 8.25,
      "learning_rate": 2.6324128913160524e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1390720.0,
      "step": 3186
    },
    {
      "epoch": 0.24513498961618338,
      "grad_norm": 4.15625,
      "learning_rate": 2.632297515575725e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1391044.0,
      "step": 3187
    },
    {
      "epoch": 0.2452119067764018,
      "grad_norm": 2.96875,
      "learning_rate": 2.6321821398353973e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1391590.0,
      "step": 3188
    },
    {
      "epoch": 0.24528882393662027,
      "grad_norm": 5.6875,
      "learning_rate": 2.6320667640950696e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 1391942.0,
      "step": 3189
    },
    {
      "epoch": 0.2453657410968387,
      "grad_norm": 4.0625,
      "learning_rate": 2.6319513883547423e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1392288.0,
      "step": 3190
    },
    {
      "epoch": 0.24544265825705716,
      "grad_norm": 6.53125,
      "learning_rate": 2.631836012614414e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1392566.0,
      "step": 3191
    },
    {
      "epoch": 0.2455195754172756,
      "grad_norm": 2.40625,
      "learning_rate": 2.631720636874087e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1393348.0,
      "step": 3192
    },
    {
      "epoch": 0.24559649257749405,
      "grad_norm": 5.5625,
      "learning_rate": 2.631605261133759e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1393896.0,
      "step": 3193
    },
    {
      "epoch": 0.24567340973771248,
      "grad_norm": 6.84375,
      "learning_rate": 2.6314898853934313e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 1394186.0,
      "step": 3194
    },
    {
      "epoch": 0.24575032689793092,
      "grad_norm": 3.953125,
      "learning_rate": 2.6313745096531036e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 1394998.0,
      "step": 3195
    },
    {
      "epoch": 0.24582724405814937,
      "grad_norm": 3.421875,
      "learning_rate": 2.6312591339127763e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1395542.0,
      "step": 3196
    },
    {
      "epoch": 0.2459041612183678,
      "grad_norm": 5.1875,
      "learning_rate": 2.631143758172448e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1396090.0,
      "step": 3197
    },
    {
      "epoch": 0.24598107837858627,
      "grad_norm": 4.8125,
      "learning_rate": 2.6310283824321208e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1396374.0,
      "step": 3198
    },
    {
      "epoch": 0.2460579955388047,
      "grad_norm": 4.625,
      "learning_rate": 2.630913006691793e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1396718.0,
      "step": 3199
    },
    {
      "epoch": 0.24613491269902316,
      "grad_norm": 6.3125,
      "learning_rate": 2.6307976309514653e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1397014.0,
      "step": 3200
    },
    {
      "epoch": 0.2462118298592416,
      "grad_norm": 3.765625,
      "learning_rate": 2.630682255211138e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1397336.0,
      "step": 3201
    },
    {
      "epoch": 0.24628874701946005,
      "grad_norm": 5.8125,
      "learning_rate": 2.63056687947081e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1397620.0,
      "step": 3202
    },
    {
      "epoch": 0.24636566417967848,
      "grad_norm": 5.03125,
      "learning_rate": 2.630451503730482e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1397930.0,
      "step": 3203
    },
    {
      "epoch": 0.24644258133989694,
      "grad_norm": 3.609375,
      "learning_rate": 2.6303361279901547e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1398494.0,
      "step": 3204
    },
    {
      "epoch": 0.24651949850011537,
      "grad_norm": 4.875,
      "learning_rate": 2.630220752249827e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1398798.0,
      "step": 3205
    },
    {
      "epoch": 0.24659641566033383,
      "grad_norm": 4.46875,
      "learning_rate": 2.6301053765094993e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1399394.0,
      "step": 3206
    },
    {
      "epoch": 0.24667333282055226,
      "grad_norm": 2.0625,
      "learning_rate": 2.6299900007691715e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1399940.0,
      "step": 3207
    },
    {
      "epoch": 0.24675024998077072,
      "grad_norm": 5.0,
      "learning_rate": 2.6298746250288438e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1400246.0,
      "step": 3208
    },
    {
      "epoch": 0.24682716714098915,
      "grad_norm": 4.84375,
      "learning_rate": 2.6297592492885165e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 1400554.0,
      "step": 3209
    },
    {
      "epoch": 0.2469040843012076,
      "grad_norm": 2.3125,
      "learning_rate": 2.6296438735481887e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9975308775901794,
      "num_tokens": 1401366.0,
      "step": 3210
    },
    {
      "epoch": 0.24698100146142604,
      "grad_norm": 3.0625,
      "learning_rate": 2.629528497807861e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1401914.0,
      "step": 3211
    },
    {
      "epoch": 0.2470579186216445,
      "grad_norm": 15.0,
      "learning_rate": 2.6294131220675337e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.9821428656578064,
      "num_tokens": 1402142.0,
      "step": 3212
    },
    {
      "epoch": 0.24713483578186293,
      "grad_norm": 8.0625,
      "learning_rate": 2.6292977463272055e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 1402432.0,
      "step": 3213
    },
    {
      "epoch": 0.2472117529420814,
      "grad_norm": 3.34375,
      "learning_rate": 2.6291823705868777e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1402990.0,
      "step": 3214
    },
    {
      "epoch": 0.24728867010229982,
      "grad_norm": 4.1875,
      "learning_rate": 2.6290669948465504e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1403334.0,
      "step": 3215
    },
    {
      "epoch": 0.24736558726251826,
      "grad_norm": 4.25,
      "learning_rate": 2.6289516191062227e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1403886.0,
      "step": 3216
    },
    {
      "epoch": 0.24744250442273671,
      "grad_norm": 4.84375,
      "learning_rate": 2.628836243365895e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1404180.0,
      "step": 3217
    },
    {
      "epoch": 0.24751942158295515,
      "grad_norm": 6.59375,
      "learning_rate": 2.628720867625567e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1404510.0,
      "step": 3218
    },
    {
      "epoch": 0.2475963387431736,
      "grad_norm": 2.734375,
      "learning_rate": 2.6286054918852395e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1405066.0,
      "step": 3219
    },
    {
      "epoch": 0.24767325590339204,
      "grad_norm": 8.5625,
      "learning_rate": 2.628490116144912e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1405406.0,
      "step": 3220
    },
    {
      "epoch": 0.2477501730636105,
      "grad_norm": 8.625,
      "learning_rate": 2.6283747404045844e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9919571280479431,
      "num_tokens": 1405782.0,
      "step": 3221
    },
    {
      "epoch": 0.24782709022382893,
      "grad_norm": 6.15625,
      "learning_rate": 2.6282593646642567e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1406116.0,
      "step": 3222
    },
    {
      "epoch": 0.2479040073840474,
      "grad_norm": 12.25,
      "learning_rate": 2.6281439889239293e-06,
      "loss": 0.0399,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 1406348.0,
      "step": 3223
    },
    {
      "epoch": 0.24798092454426582,
      "grad_norm": 2.46875,
      "learning_rate": 2.628028613183601e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1406910.0,
      "step": 3224
    },
    {
      "epoch": 0.24805784170448428,
      "grad_norm": 5.34375,
      "learning_rate": 2.6279132374432734e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 1407220.0,
      "step": 3225
    },
    {
      "epoch": 0.2481347588647027,
      "grad_norm": 7.28125,
      "learning_rate": 2.627797861702946e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 1407562.0,
      "step": 3226
    },
    {
      "epoch": 0.24821167602492117,
      "grad_norm": 3.25,
      "learning_rate": 2.6276824859626184e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1408084.0,
      "step": 3227
    },
    {
      "epoch": 0.2482885931851396,
      "grad_norm": 5.84375,
      "learning_rate": 2.6275671102222906e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 1408372.0,
      "step": 3228
    },
    {
      "epoch": 0.24836551034535806,
      "grad_norm": 4.125,
      "learning_rate": 2.627451734481963e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1408706.0,
      "step": 3229
    },
    {
      "epoch": 0.2484424275055765,
      "grad_norm": 2.078125,
      "learning_rate": 2.627336358741635e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1409262.0,
      "step": 3230
    },
    {
      "epoch": 0.24851934466579495,
      "grad_norm": 3.546875,
      "learning_rate": 2.627220983001308e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1409612.0,
      "step": 3231
    },
    {
      "epoch": 0.24859626182601338,
      "grad_norm": 7.46875,
      "learning_rate": 2.62710560726098e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1409918.0,
      "step": 3232
    },
    {
      "epoch": 0.24867317898623184,
      "grad_norm": 2.34375,
      "learning_rate": 2.6269902315206523e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 1410502.0,
      "step": 3233
    },
    {
      "epoch": 0.24875009614645027,
      "grad_norm": 4.8125,
      "learning_rate": 2.6268748557803246e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1410800.0,
      "step": 3234
    },
    {
      "epoch": 0.2488270133066687,
      "grad_norm": 2.0625,
      "learning_rate": 2.626759480039997e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1411346.0,
      "step": 3235
    },
    {
      "epoch": 0.24890393046688716,
      "grad_norm": 2.28125,
      "learning_rate": 2.626644104299669e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 1412142.0,
      "step": 3236
    },
    {
      "epoch": 0.2489808476271056,
      "grad_norm": 3.875,
      "learning_rate": 2.626528728559342e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 1412488.0,
      "step": 3237
    },
    {
      "epoch": 0.24905776478732405,
      "grad_norm": 4.5,
      "learning_rate": 2.626413352819014e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1412800.0,
      "step": 3238
    },
    {
      "epoch": 0.24913468194754249,
      "grad_norm": 2.203125,
      "learning_rate": 2.6262979770786863e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1413360.0,
      "step": 3239
    },
    {
      "epoch": 0.24921159910776094,
      "grad_norm": 5.6875,
      "learning_rate": 2.6261826013383586e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1413698.0,
      "step": 3240
    },
    {
      "epoch": 0.24928851626797938,
      "grad_norm": 2.46875,
      "learning_rate": 2.626067225598031e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1414284.0,
      "step": 3241
    },
    {
      "epoch": 0.24936543342819784,
      "grad_norm": 6.8125,
      "learning_rate": 2.6259518498577035e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 1414586.0,
      "step": 3242
    },
    {
      "epoch": 0.24944235058841627,
      "grad_norm": 5.03125,
      "learning_rate": 2.6258364741173758e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1414902.0,
      "step": 3243
    },
    {
      "epoch": 0.24951926774863473,
      "grad_norm": 2.9375,
      "learning_rate": 2.625721098377048e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1415228.0,
      "step": 3244
    },
    {
      "epoch": 0.24959618490885316,
      "grad_norm": 5.53125,
      "learning_rate": 2.6256057226367203e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.99303138256073,
      "num_tokens": 1415516.0,
      "step": 3245
    },
    {
      "epoch": 0.24967310206907162,
      "grad_norm": 4.96875,
      "learning_rate": 2.6254903468963925e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 1415826.0,
      "step": 3246
    },
    {
      "epoch": 0.24975001922929005,
      "grad_norm": 13.25,
      "learning_rate": 2.625374971156065e-06,
      "loss": 0.0404,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 1416058.0,
      "step": 3247
    },
    {
      "epoch": 0.2498269363895085,
      "grad_norm": 3.3125,
      "learning_rate": 2.6252595954157375e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 1416580.0,
      "step": 3248
    },
    {
      "epoch": 0.24990385354972694,
      "grad_norm": 5.40625,
      "learning_rate": 2.6251442196754097e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1416864.0,
      "step": 3249
    },
    {
      "epoch": 0.2499807707099454,
      "grad_norm": 3.375,
      "learning_rate": 2.625028843935082e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 1417434.0,
      "step": 3250
    },
    {
      "epoch": 0.25005768787016386,
      "grad_norm": 7.0,
      "learning_rate": 2.6249134681947543e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 1417762.0,
      "step": 3251
    },
    {
      "epoch": 0.25013460503038226,
      "grad_norm": 6.59375,
      "learning_rate": 2.6247980924544265e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 1418106.0,
      "step": 3252
    },
    {
      "epoch": 0.2502115221906007,
      "grad_norm": 3.265625,
      "learning_rate": 2.6246827167140988e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1418674.0,
      "step": 3253
    },
    {
      "epoch": 0.2502884393508192,
      "grad_norm": 3.53125,
      "learning_rate": 2.6245673409737715e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1419232.0,
      "step": 3254
    },
    {
      "epoch": 0.25036535651103764,
      "grad_norm": 2.96875,
      "learning_rate": 2.6244519652334437e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 1419816.0,
      "step": 3255
    },
    {
      "epoch": 0.25044227367125604,
      "grad_norm": 6.875,
      "learning_rate": 2.624336589493116e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1420116.0,
      "step": 3256
    },
    {
      "epoch": 0.2505191908314745,
      "grad_norm": 6.375,
      "learning_rate": 2.6242212137527882e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 1420472.0,
      "step": 3257
    },
    {
      "epoch": 0.25059610799169296,
      "grad_norm": 2.375,
      "learning_rate": 2.6241058380124605e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 1421286.0,
      "step": 3258
    },
    {
      "epoch": 0.25067302515191137,
      "grad_norm": 3.0625,
      "learning_rate": 2.623990462272133e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1421838.0,
      "step": 3259
    },
    {
      "epoch": 0.2507499423121298,
      "grad_norm": 7.03125,
      "learning_rate": 2.6238750865318054e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1422178.0,
      "step": 3260
    },
    {
      "epoch": 0.2508268594723483,
      "grad_norm": 7.6875,
      "learning_rate": 2.6237597107914777e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1422520.0,
      "step": 3261
    },
    {
      "epoch": 0.25090377663256674,
      "grad_norm": 3.140625,
      "learning_rate": 2.62364433505115e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1423054.0,
      "step": 3262
    },
    {
      "epoch": 0.25098069379278515,
      "grad_norm": 3.109375,
      "learning_rate": 2.623528959310822e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 1423404.0,
      "step": 3263
    },
    {
      "epoch": 0.2510576109530036,
      "grad_norm": 10.25,
      "learning_rate": 2.6234135835704945e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 1423684.0,
      "step": 3264
    },
    {
      "epoch": 0.25113452811322207,
      "grad_norm": 5.4375,
      "learning_rate": 2.623298207830167e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1423992.0,
      "step": 3265
    },
    {
      "epoch": 0.2512114452734405,
      "grad_norm": 2.15625,
      "learning_rate": 2.6231828320898394e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 1424802.0,
      "step": 3266
    },
    {
      "epoch": 0.25128836243365893,
      "grad_norm": 6.75,
      "learning_rate": 2.6230674563495117e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1425086.0,
      "step": 3267
    },
    {
      "epoch": 0.2513652795938774,
      "grad_norm": 2.421875,
      "learning_rate": 2.622952080609184e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1425868.0,
      "step": 3268
    },
    {
      "epoch": 0.25144219675409585,
      "grad_norm": 9.4375,
      "learning_rate": 2.622836704868856e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 1426148.0,
      "step": 3269
    },
    {
      "epoch": 0.2515191139143143,
      "grad_norm": 9.1875,
      "learning_rate": 2.622721329128529e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1426428.0,
      "step": 3270
    },
    {
      "epoch": 0.2515960310745327,
      "grad_norm": 9.125,
      "learning_rate": 2.622605953388201e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 1426752.0,
      "step": 3271
    },
    {
      "epoch": 0.25167294823475117,
      "grad_norm": 3.140625,
      "learning_rate": 2.622490577647873e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1427328.0,
      "step": 3272
    },
    {
      "epoch": 0.25174986539496963,
      "grad_norm": 3.25,
      "learning_rate": 2.6223752019075456e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1427852.0,
      "step": 3273
    },
    {
      "epoch": 0.2518267825551881,
      "grad_norm": 4.9375,
      "learning_rate": 2.622259826167218e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1428158.0,
      "step": 3274
    },
    {
      "epoch": 0.2519036997154065,
      "grad_norm": 2.96875,
      "learning_rate": 2.62214445042689e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1428742.0,
      "step": 3275
    },
    {
      "epoch": 0.25198061687562495,
      "grad_norm": 5.1875,
      "learning_rate": 2.622029074686563e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1429064.0,
      "step": 3276
    },
    {
      "epoch": 0.2520575340358434,
      "grad_norm": 2.4375,
      "learning_rate": 2.621913698946235e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1429582.0,
      "step": 3277
    },
    {
      "epoch": 0.2521344511960618,
      "grad_norm": 5.375,
      "learning_rate": 2.6217983232059073e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 1429864.0,
      "step": 3278
    },
    {
      "epoch": 0.2522113683562803,
      "grad_norm": 6.375,
      "learning_rate": 2.6216829474655796e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1430192.0,
      "step": 3279
    },
    {
      "epoch": 0.25228828551649873,
      "grad_norm": 5.46875,
      "learning_rate": 2.621567571725252e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1430532.0,
      "step": 3280
    },
    {
      "epoch": 0.2523652026767172,
      "grad_norm": 6.28125,
      "learning_rate": 2.6214521959849245e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1430882.0,
      "step": 3281
    },
    {
      "epoch": 0.2524421198369356,
      "grad_norm": 6.1875,
      "learning_rate": 2.621336820244597e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1431160.0,
      "step": 3282
    },
    {
      "epoch": 0.25251903699715406,
      "grad_norm": 4.59375,
      "learning_rate": 2.6212214445042686e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1431756.0,
      "step": 3283
    },
    {
      "epoch": 0.2525959541573725,
      "grad_norm": 1.9609375,
      "learning_rate": 2.6211060687639413e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 1432560.0,
      "step": 3284
    },
    {
      "epoch": 0.252672871317591,
      "grad_norm": 2.453125,
      "learning_rate": 2.6209906930236136e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1433140.0,
      "step": 3285
    },
    {
      "epoch": 0.2527497884778094,
      "grad_norm": 3.25,
      "learning_rate": 2.620875317283286e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 1433700.0,
      "step": 3286
    },
    {
      "epoch": 0.25282670563802784,
      "grad_norm": 6.40625,
      "learning_rate": 2.6207599415429585e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1434010.0,
      "step": 3287
    },
    {
      "epoch": 0.2529036227982463,
      "grad_norm": 6.28125,
      "learning_rate": 2.6206445658026308e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1434320.0,
      "step": 3288
    },
    {
      "epoch": 0.25298053995846476,
      "grad_norm": 7.46875,
      "learning_rate": 2.620529190062303e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1434658.0,
      "step": 3289
    },
    {
      "epoch": 0.25305745711868316,
      "grad_norm": 2.0625,
      "learning_rate": 2.6204138143219753e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1435438.0,
      "step": 3290
    },
    {
      "epoch": 0.2531343742789016,
      "grad_norm": 9.875,
      "learning_rate": 2.6202984385816475e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 1435796.0,
      "step": 3291
    },
    {
      "epoch": 0.2532112914391201,
      "grad_norm": 1.8671875,
      "learning_rate": 2.6201830628413202e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1436352.0,
      "step": 3292
    },
    {
      "epoch": 0.25328820859933854,
      "grad_norm": 6.625,
      "learning_rate": 2.6200676871009925e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1436692.0,
      "step": 3293
    },
    {
      "epoch": 0.25336512575955694,
      "grad_norm": 3.796875,
      "learning_rate": 2.6199523113606643e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1437042.0,
      "step": 3294
    },
    {
      "epoch": 0.2534420429197754,
      "grad_norm": 2.703125,
      "learning_rate": 2.619836935620337e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1437586.0,
      "step": 3295
    },
    {
      "epoch": 0.25351896007999386,
      "grad_norm": 8.0,
      "learning_rate": 2.6197215598800093e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 1437870.0,
      "step": 3296
    },
    {
      "epoch": 0.25359587724021226,
      "grad_norm": 9.6875,
      "learning_rate": 2.6196061841396815e-06,
      "loss": 0.0418,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1438130.0,
      "step": 3297
    },
    {
      "epoch": 0.2536727944004307,
      "grad_norm": 7.53125,
      "learning_rate": 2.619490808399354e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 1438480.0,
      "step": 3298
    },
    {
      "epoch": 0.2537497115606492,
      "grad_norm": 2.5,
      "learning_rate": 2.6193754326590265e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1439038.0,
      "step": 3299
    },
    {
      "epoch": 0.25382662872086764,
      "grad_norm": 5.4375,
      "learning_rate": 2.6192600569186987e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1439348.0,
      "step": 3300
    },
    {
      "epoch": 0.25390354588108605,
      "grad_norm": 3.453125,
      "learning_rate": 2.619144681178371e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1439924.0,
      "step": 3301
    },
    {
      "epoch": 0.2539804630413045,
      "grad_norm": 5.25,
      "learning_rate": 2.6190293054380432e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1440218.0,
      "step": 3302
    },
    {
      "epoch": 0.25405738020152296,
      "grad_norm": 7.9375,
      "learning_rate": 2.6189139296977155e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 1440496.0,
      "step": 3303
    },
    {
      "epoch": 0.2541342973617414,
      "grad_norm": 4.625,
      "learning_rate": 2.618798553957388e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1440822.0,
      "step": 3304
    },
    {
      "epoch": 0.2542112145219598,
      "grad_norm": 5.15625,
      "learning_rate": 2.61868317821706e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9837133288383484,
      "num_tokens": 1441130.0,
      "step": 3305
    },
    {
      "epoch": 0.2542881316821783,
      "grad_norm": 3.125,
      "learning_rate": 2.6185678024767327e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 1441674.0,
      "step": 3306
    },
    {
      "epoch": 0.25436504884239675,
      "grad_norm": 2.859375,
      "learning_rate": 2.618452426736405e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 1442206.0,
      "step": 3307
    },
    {
      "epoch": 0.2544419660026152,
      "grad_norm": 2.453125,
      "learning_rate": 2.618337050996077e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1442758.0,
      "step": 3308
    },
    {
      "epoch": 0.2545188831628336,
      "grad_norm": 4.6875,
      "learning_rate": 2.61822167525575e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1443066.0,
      "step": 3309
    },
    {
      "epoch": 0.25459580032305207,
      "grad_norm": 2.671875,
      "learning_rate": 2.6181062995154217e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1443640.0,
      "step": 3310
    },
    {
      "epoch": 0.2546727174832705,
      "grad_norm": 8.1875,
      "learning_rate": 2.617990923775094e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1443918.0,
      "step": 3311
    },
    {
      "epoch": 0.254749634643489,
      "grad_norm": 2.234375,
      "learning_rate": 2.6178755480347667e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1444704.0,
      "step": 3312
    },
    {
      "epoch": 0.2548265518037074,
      "grad_norm": 4.03125,
      "learning_rate": 2.617760172294439e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1445276.0,
      "step": 3313
    },
    {
      "epoch": 0.25490346896392585,
      "grad_norm": 3.078125,
      "learning_rate": 2.617644796554111e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9959595799446106,
      "num_tokens": 1445772.0,
      "step": 3314
    },
    {
      "epoch": 0.2549803861241443,
      "grad_norm": 2.6875,
      "learning_rate": 2.617529420813784e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 1446346.0,
      "step": 3315
    },
    {
      "epoch": 0.25505730328436277,
      "grad_norm": 2.59375,
      "learning_rate": 2.6174140450734557e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1446904.0,
      "step": 3316
    },
    {
      "epoch": 0.2551342204445812,
      "grad_norm": 6.96875,
      "learning_rate": 2.6172986693331284e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 1447212.0,
      "step": 3317
    },
    {
      "epoch": 0.25521113760479963,
      "grad_norm": 10.0625,
      "learning_rate": 2.6171832935928006e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1447540.0,
      "step": 3318
    },
    {
      "epoch": 0.2552880547650181,
      "grad_norm": 6.40625,
      "learning_rate": 2.617067917852473e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1447826.0,
      "step": 3319
    },
    {
      "epoch": 0.2553649719252365,
      "grad_norm": 6.78125,
      "learning_rate": 2.6169525421121456e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1448162.0,
      "step": 3320
    },
    {
      "epoch": 0.25544188908545495,
      "grad_norm": 2.984375,
      "learning_rate": 2.6168371663718174e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1448718.0,
      "step": 3321
    },
    {
      "epoch": 0.2555188062456734,
      "grad_norm": 4.96875,
      "learning_rate": 2.6167217906314897e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1449046.0,
      "step": 3322
    },
    {
      "epoch": 0.2555957234058919,
      "grad_norm": 8.25,
      "learning_rate": 2.6166064148911623e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 1449376.0,
      "step": 3323
    },
    {
      "epoch": 0.2556726405661103,
      "grad_norm": 2.734375,
      "learning_rate": 2.6164910391508346e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1449938.0,
      "step": 3324
    },
    {
      "epoch": 0.25574955772632874,
      "grad_norm": 4.25,
      "learning_rate": 2.616375663410507e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1450260.0,
      "step": 3325
    },
    {
      "epoch": 0.2558264748865472,
      "grad_norm": 3.609375,
      "learning_rate": 2.6162602876701795e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 1450814.0,
      "step": 3326
    },
    {
      "epoch": 0.25590339204676565,
      "grad_norm": 4.59375,
      "learning_rate": 2.6161449119298514e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 1451148.0,
      "step": 3327
    },
    {
      "epoch": 0.25598030920698406,
      "grad_norm": 5.28125,
      "learning_rate": 2.616029536189524e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1451456.0,
      "step": 3328
    },
    {
      "epoch": 0.2560572263672025,
      "grad_norm": 3.59375,
      "learning_rate": 2.6159141604491963e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1451992.0,
      "step": 3329
    },
    {
      "epoch": 0.256134143527421,
      "grad_norm": 3.4375,
      "learning_rate": 2.6157987847088686e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1452524.0,
      "step": 3330
    },
    {
      "epoch": 0.25621106068763944,
      "grad_norm": 17.25,
      "learning_rate": 2.6156834089685413e-06,
      "loss": 0.0391,
      "mean_token_accuracy": 0.9821428656578064,
      "num_tokens": 1452752.0,
      "step": 3331
    },
    {
      "epoch": 0.25628797784785784,
      "grad_norm": 4.125,
      "learning_rate": 2.615568033228213e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1453316.0,
      "step": 3332
    },
    {
      "epoch": 0.2563648950080763,
      "grad_norm": 8.4375,
      "learning_rate": 2.6154526574878854e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1453594.0,
      "step": 3333
    },
    {
      "epoch": 0.25644181216829476,
      "grad_norm": 3.921875,
      "learning_rate": 2.615337281747558e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1453916.0,
      "step": 3334
    },
    {
      "epoch": 0.2565187293285132,
      "grad_norm": 5.0,
      "learning_rate": 2.6152219060072303e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 1454200.0,
      "step": 3335
    },
    {
      "epoch": 0.2565956464887316,
      "grad_norm": 4.40625,
      "learning_rate": 2.6151065302669026e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 1454518.0,
      "step": 3336
    },
    {
      "epoch": 0.2566725636489501,
      "grad_norm": 7.125,
      "learning_rate": 2.614991154526575e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1454820.0,
      "step": 3337
    },
    {
      "epoch": 0.25674948080916854,
      "grad_norm": 10.4375,
      "learning_rate": 2.614875778786247e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9960784316062927,
      "num_tokens": 1455076.0,
      "step": 3338
    },
    {
      "epoch": 0.25682639796938694,
      "grad_norm": 2.8125,
      "learning_rate": 2.6147604030459197e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 1455610.0,
      "step": 3339
    },
    {
      "epoch": 0.2569033151296054,
      "grad_norm": 3.4375,
      "learning_rate": 2.614645027305592e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1456156.0,
      "step": 3340
    },
    {
      "epoch": 0.25698023228982386,
      "grad_norm": 3.203125,
      "learning_rate": 2.6145296515652643e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1456698.0,
      "step": 3341
    },
    {
      "epoch": 0.2570571494500423,
      "grad_norm": 5.4375,
      "learning_rate": 2.614414275824937e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1456998.0,
      "step": 3342
    },
    {
      "epoch": 0.2571340666102607,
      "grad_norm": 8.375,
      "learning_rate": 2.6142989000846088e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1457230.0,
      "step": 3343
    },
    {
      "epoch": 0.2572109837704792,
      "grad_norm": 5.0625,
      "learning_rate": 2.614183524344281e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1457552.0,
      "step": 3344
    },
    {
      "epoch": 0.25728790093069764,
      "grad_norm": 7.5625,
      "learning_rate": 2.6140681486039537e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1457866.0,
      "step": 3345
    },
    {
      "epoch": 0.2573648180909161,
      "grad_norm": 5.8125,
      "learning_rate": 2.613952772863626e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1458186.0,
      "step": 3346
    },
    {
      "epoch": 0.2574417352511345,
      "grad_norm": 8.125,
      "learning_rate": 2.6138373971232982e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1458514.0,
      "step": 3347
    },
    {
      "epoch": 0.25751865241135297,
      "grad_norm": 3.625,
      "learning_rate": 2.6137220213829705e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 1459074.0,
      "step": 3348
    },
    {
      "epoch": 0.2575955695715714,
      "grad_norm": 5.96875,
      "learning_rate": 2.6136066456426428e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1459362.0,
      "step": 3349
    },
    {
      "epoch": 0.2576724867317899,
      "grad_norm": 9.375,
      "learning_rate": 2.6134912699023154e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1459704.0,
      "step": 3350
    },
    {
      "epoch": 0.2577494038920083,
      "grad_norm": 6.21875,
      "learning_rate": 2.6133758941619877e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 1460036.0,
      "step": 3351
    },
    {
      "epoch": 0.25782632105222675,
      "grad_norm": 4.375,
      "learning_rate": 2.61326051842166e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1460358.0,
      "step": 3352
    },
    {
      "epoch": 0.2579032382124452,
      "grad_norm": 2.40625,
      "learning_rate": 2.6131451426813326e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1461162.0,
      "step": 3353
    },
    {
      "epoch": 0.25798015537266367,
      "grad_norm": 6.40625,
      "learning_rate": 2.6130297669410045e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 1461446.0,
      "step": 3354
    },
    {
      "epoch": 0.25805707253288207,
      "grad_norm": 3.34375,
      "learning_rate": 2.6129143912006767e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1461988.0,
      "step": 3355
    },
    {
      "epoch": 0.25813398969310053,
      "grad_norm": 2.421875,
      "learning_rate": 2.6127990154603494e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 1462796.0,
      "step": 3356
    },
    {
      "epoch": 0.258210906853319,
      "grad_norm": 2.609375,
      "learning_rate": 2.6126836397200217e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9936061501502991,
      "num_tokens": 1463580.0,
      "step": 3357
    },
    {
      "epoch": 0.2582878240135374,
      "grad_norm": 2.640625,
      "learning_rate": 2.612568263979694e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1464130.0,
      "step": 3358
    },
    {
      "epoch": 0.25836474117375585,
      "grad_norm": 4.90625,
      "learning_rate": 2.612452888239366e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 1464490.0,
      "step": 3359
    },
    {
      "epoch": 0.2584416583339743,
      "grad_norm": 6.03125,
      "learning_rate": 2.6123375124990384e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1464782.0,
      "step": 3360
    },
    {
      "epoch": 0.25851857549419277,
      "grad_norm": 2.359375,
      "learning_rate": 2.612222136758711e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1465356.0,
      "step": 3361
    },
    {
      "epoch": 0.2585954926544112,
      "grad_norm": 8.125,
      "learning_rate": 2.6121067610183834e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1465698.0,
      "step": 3362
    },
    {
      "epoch": 0.25867240981462963,
      "grad_norm": 3.640625,
      "learning_rate": 2.6119913852780556e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 1466020.0,
      "step": 3363
    },
    {
      "epoch": 0.2587493269748481,
      "grad_norm": 6.9375,
      "learning_rate": 2.611876009537728e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 1466310.0,
      "step": 3364
    },
    {
      "epoch": 0.25882624413506655,
      "grad_norm": 2.15625,
      "learning_rate": 2.6117606337974e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1466872.0,
      "step": 3365
    },
    {
      "epoch": 0.25890316129528496,
      "grad_norm": 7.3125,
      "learning_rate": 2.6116452580570724e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 1467204.0,
      "step": 3366
    },
    {
      "epoch": 0.2589800784555034,
      "grad_norm": 2.984375,
      "learning_rate": 2.611529882316745e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1467744.0,
      "step": 3367
    },
    {
      "epoch": 0.2590569956157219,
      "grad_norm": 6.65625,
      "learning_rate": 2.6114145065764174e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 1468034.0,
      "step": 3368
    },
    {
      "epoch": 0.25913391277594033,
      "grad_norm": 5.84375,
      "learning_rate": 2.6112991308360896e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1468318.0,
      "step": 3369
    },
    {
      "epoch": 0.25921082993615874,
      "grad_norm": 2.984375,
      "learning_rate": 2.611183755095762e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1468858.0,
      "step": 3370
    },
    {
      "epoch": 0.2592877470963772,
      "grad_norm": 2.265625,
      "learning_rate": 2.611068379355434e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1469638.0,
      "step": 3371
    },
    {
      "epoch": 0.25936466425659566,
      "grad_norm": 3.484375,
      "learning_rate": 2.6109530036151064e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1470218.0,
      "step": 3372
    },
    {
      "epoch": 0.2594415814168141,
      "grad_norm": 5.25,
      "learning_rate": 2.610837627874779e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1470554.0,
      "step": 3373
    },
    {
      "epoch": 0.2595184985770325,
      "grad_norm": 6.0,
      "learning_rate": 2.6107222521344513e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 1470920.0,
      "step": 3374
    },
    {
      "epoch": 0.259595415737251,
      "grad_norm": 7.4375,
      "learning_rate": 2.6106068763941236e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1471220.0,
      "step": 3375
    },
    {
      "epoch": 0.25967233289746944,
      "grad_norm": 5.78125,
      "learning_rate": 2.610491500653796e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1471502.0,
      "step": 3376
    },
    {
      "epoch": 0.2597492500576879,
      "grad_norm": 3.375,
      "learning_rate": 2.610376124913468e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1472018.0,
      "step": 3377
    },
    {
      "epoch": 0.2598261672179063,
      "grad_norm": 5.28125,
      "learning_rate": 2.6102607491731408e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 1472334.0,
      "step": 3378
    },
    {
      "epoch": 0.25990308437812476,
      "grad_norm": 6.40625,
      "learning_rate": 2.610145373432813e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1472636.0,
      "step": 3379
    },
    {
      "epoch": 0.2599800015383432,
      "grad_norm": 2.40625,
      "learning_rate": 2.6100299976924853e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1473178.0,
      "step": 3380
    },
    {
      "epoch": 0.2600569186985616,
      "grad_norm": 4.125,
      "learning_rate": 2.6099146219521576e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1473702.0,
      "step": 3381
    },
    {
      "epoch": 0.2601338358587801,
      "grad_norm": 4.625,
      "learning_rate": 2.60979924621183e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 1473980.0,
      "step": 3382
    },
    {
      "epoch": 0.26021075301899854,
      "grad_norm": 3.625,
      "learning_rate": 2.609683870471502e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1474500.0,
      "step": 3383
    },
    {
      "epoch": 0.260287670179217,
      "grad_norm": 3.296875,
      "learning_rate": 2.6095684947311748e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1475086.0,
      "step": 3384
    },
    {
      "epoch": 0.2603645873394354,
      "grad_norm": 3.15625,
      "learning_rate": 2.609453118990847e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1475912.0,
      "step": 3385
    },
    {
      "epoch": 0.26044150449965386,
      "grad_norm": 4.5625,
      "learning_rate": 2.6093377432505193e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 1476230.0,
      "step": 3386
    },
    {
      "epoch": 0.2605184216598723,
      "grad_norm": 3.21875,
      "learning_rate": 2.6092223675101915e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 1476790.0,
      "step": 3387
    },
    {
      "epoch": 0.2605953388200908,
      "grad_norm": 2.15625,
      "learning_rate": 2.6091069917698638e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1477338.0,
      "step": 3388
    },
    {
      "epoch": 0.2606722559803092,
      "grad_norm": 6.59375,
      "learning_rate": 2.6089916160295365e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1477616.0,
      "step": 3389
    },
    {
      "epoch": 0.26074917314052765,
      "grad_norm": 5.71875,
      "learning_rate": 2.6088762402892087e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1477944.0,
      "step": 3390
    },
    {
      "epoch": 0.2608260903007461,
      "grad_norm": 7.0625,
      "learning_rate": 2.608760864548881e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 1478284.0,
      "step": 3391
    },
    {
      "epoch": 0.26090300746096456,
      "grad_norm": 3.375,
      "learning_rate": 2.6086454888085532e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 1478868.0,
      "step": 3392
    },
    {
      "epoch": 0.26097992462118297,
      "grad_norm": 3.890625,
      "learning_rate": 2.6085301130682255e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1479184.0,
      "step": 3393
    },
    {
      "epoch": 0.2610568417814014,
      "grad_norm": 5.40625,
      "learning_rate": 2.6084147373278978e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1479518.0,
      "step": 3394
    },
    {
      "epoch": 0.2611337589416199,
      "grad_norm": 4.09375,
      "learning_rate": 2.6082993615875704e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1480092.0,
      "step": 3395
    },
    {
      "epoch": 0.26121067610183835,
      "grad_norm": 7.0,
      "learning_rate": 2.6081839858472427e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 1480398.0,
      "step": 3396
    },
    {
      "epoch": 0.26128759326205675,
      "grad_norm": 8.625,
      "learning_rate": 2.608068610106915e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1480658.0,
      "step": 3397
    },
    {
      "epoch": 0.2613645104222752,
      "grad_norm": 9.5625,
      "learning_rate": 2.6079532343665872e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 1480912.0,
      "step": 3398
    },
    {
      "epoch": 0.26144142758249367,
      "grad_norm": 2.4375,
      "learning_rate": 2.6078378586262595e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 1481932.0,
      "step": 3399
    },
    {
      "epoch": 0.26151834474271207,
      "grad_norm": 4.3125,
      "learning_rate": 2.607722482885932e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1482282.0,
      "step": 3400
    },
    {
      "epoch": 0.26159526190293053,
      "grad_norm": 2.40625,
      "learning_rate": 2.6076071071456044e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 1483090.0,
      "step": 3401
    },
    {
      "epoch": 0.261672179063149,
      "grad_norm": 7.15625,
      "learning_rate": 2.6074917314052767e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1483402.0,
      "step": 3402
    },
    {
      "epoch": 0.26174909622336745,
      "grad_norm": 4.6875,
      "learning_rate": 2.607376355664949e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1483700.0,
      "step": 3403
    },
    {
      "epoch": 0.26182601338358585,
      "grad_norm": 7.25,
      "learning_rate": 2.607260979924621e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 1483990.0,
      "step": 3404
    },
    {
      "epoch": 0.2619029305438043,
      "grad_norm": 7.28125,
      "learning_rate": 2.6071456041842934e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1484300.0,
      "step": 3405
    },
    {
      "epoch": 0.2619798477040228,
      "grad_norm": 3.296875,
      "learning_rate": 2.607030228443966e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1484878.0,
      "step": 3406
    },
    {
      "epoch": 0.26205676486424123,
      "grad_norm": 2.28125,
      "learning_rate": 2.6069148527036384e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1485698.0,
      "step": 3407
    },
    {
      "epoch": 0.26213368202445964,
      "grad_norm": 6.84375,
      "learning_rate": 2.6067994769633106e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 1486006.0,
      "step": 3408
    },
    {
      "epoch": 0.2622105991846781,
      "grad_norm": 6.78125,
      "learning_rate": 2.606684101222983e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1486292.0,
      "step": 3409
    },
    {
      "epoch": 0.26228751634489655,
      "grad_norm": 6.15625,
      "learning_rate": 2.606568725482655e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 1486600.0,
      "step": 3410
    },
    {
      "epoch": 0.262364433505115,
      "grad_norm": 4.875,
      "learning_rate": 2.606453349742328e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1486924.0,
      "step": 3411
    },
    {
      "epoch": 0.2624413506653334,
      "grad_norm": 7.0,
      "learning_rate": 2.606337974002e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1487242.0,
      "step": 3412
    },
    {
      "epoch": 0.2625182678255519,
      "grad_norm": 2.796875,
      "learning_rate": 2.606222598261672e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1487786.0,
      "step": 3413
    },
    {
      "epoch": 0.26259518498577034,
      "grad_norm": 6.9375,
      "learning_rate": 2.6061072225213446e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1488072.0,
      "step": 3414
    },
    {
      "epoch": 0.2626721021459888,
      "grad_norm": 3.671875,
      "learning_rate": 2.605991846781017e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1488394.0,
      "step": 3415
    },
    {
      "epoch": 0.2627490193062072,
      "grad_norm": 3.859375,
      "learning_rate": 2.605876471040689e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1488888.0,
      "step": 3416
    },
    {
      "epoch": 0.26282593646642566,
      "grad_norm": 1.875,
      "learning_rate": 2.605761095300362e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1489450.0,
      "step": 3417
    },
    {
      "epoch": 0.2629028536266441,
      "grad_norm": 5.78125,
      "learning_rate": 2.605645719560034e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 1489810.0,
      "step": 3418
    },
    {
      "epoch": 0.2629797707868626,
      "grad_norm": 6.0,
      "learning_rate": 2.6055303438197063e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1490096.0,
      "step": 3419
    },
    {
      "epoch": 0.263056687947081,
      "grad_norm": 2.546875,
      "learning_rate": 2.6054149680793786e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1490900.0,
      "step": 3420
    },
    {
      "epoch": 0.26313360510729944,
      "grad_norm": 2.265625,
      "learning_rate": 2.605299592339051e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1491450.0,
      "step": 3421
    },
    {
      "epoch": 0.2632105222675179,
      "grad_norm": 2.46875,
      "learning_rate": 2.605184216598723e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1492006.0,
      "step": 3422
    },
    {
      "epoch": 0.2632874394277363,
      "grad_norm": 6.96875,
      "learning_rate": 2.6050688408583958e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1492286.0,
      "step": 3423
    },
    {
      "epoch": 0.26336435658795476,
      "grad_norm": 2.671875,
      "learning_rate": 2.6049534651180676e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 1493096.0,
      "step": 3424
    },
    {
      "epoch": 0.2634412737481732,
      "grad_norm": 2.71875,
      "learning_rate": 2.6048380893777403e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1493674.0,
      "step": 3425
    },
    {
      "epoch": 0.2635181909083917,
      "grad_norm": 5.15625,
      "learning_rate": 2.6047227136374126e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1493980.0,
      "step": 3426
    },
    {
      "epoch": 0.2635951080686101,
      "grad_norm": 2.140625,
      "learning_rate": 2.604607337897085e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1494766.0,
      "step": 3427
    },
    {
      "epoch": 0.26367202522882854,
      "grad_norm": 2.625,
      "learning_rate": 2.6044919621567575e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 1495316.0,
      "step": 3428
    },
    {
      "epoch": 0.263748942389047,
      "grad_norm": 3.359375,
      "learning_rate": 2.6043765864164298e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 1495840.0,
      "step": 3429
    },
    {
      "epoch": 0.26382585954926546,
      "grad_norm": 5.71875,
      "learning_rate": 2.6042612106761016e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1496132.0,
      "step": 3430
    },
    {
      "epoch": 0.26390277670948387,
      "grad_norm": 1.8671875,
      "learning_rate": 2.6041458349357743e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 1496946.0,
      "step": 3431
    },
    {
      "epoch": 0.2639796938697023,
      "grad_norm": 1.734375,
      "learning_rate": 2.6040304591954465e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1497732.0,
      "step": 3432
    },
    {
      "epoch": 0.2640566110299208,
      "grad_norm": 6.5,
      "learning_rate": 2.6039150834551188e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1498066.0,
      "step": 3433
    },
    {
      "epoch": 0.26413352819013924,
      "grad_norm": 5.21875,
      "learning_rate": 2.6037997077147915e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1498410.0,
      "step": 3434
    },
    {
      "epoch": 0.26421044535035765,
      "grad_norm": 6.9375,
      "learning_rate": 2.6036843319744633e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 1498700.0,
      "step": 3435
    },
    {
      "epoch": 0.2642873625105761,
      "grad_norm": 7.28125,
      "learning_rate": 2.603568956234136e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1499040.0,
      "step": 3436
    },
    {
      "epoch": 0.26436427967079457,
      "grad_norm": 5.5625,
      "learning_rate": 2.6034535804938082e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9837133288383484,
      "num_tokens": 1499348.0,
      "step": 3437
    },
    {
      "epoch": 0.264441196831013,
      "grad_norm": 2.296875,
      "learning_rate": 2.6033382047534805e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 1499894.0,
      "step": 3438
    },
    {
      "epoch": 0.26451811399123143,
      "grad_norm": 4.0625,
      "learning_rate": 2.603222829013153e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 1500496.0,
      "step": 3439
    },
    {
      "epoch": 0.2645950311514499,
      "grad_norm": 2.0625,
      "learning_rate": 2.6031074532728254e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1501056.0,
      "step": 3440
    },
    {
      "epoch": 0.26467194831166835,
      "grad_norm": 3.109375,
      "learning_rate": 2.6029920775324973e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 1501636.0,
      "step": 3441
    },
    {
      "epoch": 0.26474886547188675,
      "grad_norm": 2.875,
      "learning_rate": 2.60287670179217e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1502192.0,
      "step": 3442
    },
    {
      "epoch": 0.2648257826321052,
      "grad_norm": 3.0,
      "learning_rate": 2.6027613260518422e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1502744.0,
      "step": 3443
    },
    {
      "epoch": 0.26490269979232367,
      "grad_norm": 7.6875,
      "learning_rate": 2.6026459503115145e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 1503036.0,
      "step": 3444
    },
    {
      "epoch": 0.26497961695254213,
      "grad_norm": 7.53125,
      "learning_rate": 2.602530574571187e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 1503392.0,
      "step": 3445
    },
    {
      "epoch": 0.26505653411276053,
      "grad_norm": 5.59375,
      "learning_rate": 2.602415198830859e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 1503706.0,
      "step": 3446
    },
    {
      "epoch": 0.265133451272979,
      "grad_norm": 6.21875,
      "learning_rate": 2.6022998230905317e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 1504070.0,
      "step": 3447
    },
    {
      "epoch": 0.26521036843319745,
      "grad_norm": 2.8125,
      "learning_rate": 2.602184447350204e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9910233616828918,
      "num_tokens": 1504628.0,
      "step": 3448
    },
    {
      "epoch": 0.2652872855934159,
      "grad_norm": 7.75,
      "learning_rate": 2.602069071609876e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 1504984.0,
      "step": 3449
    },
    {
      "epoch": 0.2653642027536343,
      "grad_norm": 3.078125,
      "learning_rate": 2.601953695869549e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1505540.0,
      "step": 3450
    },
    {
      "epoch": 0.2654411199138528,
      "grad_norm": 2.265625,
      "learning_rate": 2.6018383201292207e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1506088.0,
      "step": 3451
    },
    {
      "epoch": 0.26551803707407123,
      "grad_norm": 3.921875,
      "learning_rate": 2.601722944388893e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 1506440.0,
      "step": 3452
    },
    {
      "epoch": 0.2655949542342897,
      "grad_norm": 5.84375,
      "learning_rate": 2.6016075686485656e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1506760.0,
      "step": 3453
    },
    {
      "epoch": 0.2656718713945081,
      "grad_norm": 2.34375,
      "learning_rate": 2.601492192908238e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 1507584.0,
      "step": 3454
    },
    {
      "epoch": 0.26574878855472656,
      "grad_norm": 3.921875,
      "learning_rate": 2.60137681716791e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1508164.0,
      "step": 3455
    },
    {
      "epoch": 0.265825705714945,
      "grad_norm": 11.0,
      "learning_rate": 2.601261441427583e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 1508496.0,
      "step": 3456
    },
    {
      "epoch": 0.2659026228751635,
      "grad_norm": 5.71875,
      "learning_rate": 2.6011460656872547e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1508816.0,
      "step": 3457
    },
    {
      "epoch": 0.2659795400353819,
      "grad_norm": 8.625,
      "learning_rate": 2.6010306899469274e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1509140.0,
      "step": 3458
    },
    {
      "epoch": 0.26605645719560034,
      "grad_norm": 4.4375,
      "learning_rate": 2.6009153142065996e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1509658.0,
      "step": 3459
    },
    {
      "epoch": 0.2661333743558188,
      "grad_norm": 9.4375,
      "learning_rate": 2.600799938466272e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1509918.0,
      "step": 3460
    },
    {
      "epoch": 0.2662102915160372,
      "grad_norm": 3.703125,
      "learning_rate": 2.6006845627259446e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1510438.0,
      "step": 3461
    },
    {
      "epoch": 0.26628720867625566,
      "grad_norm": 3.359375,
      "learning_rate": 2.6005691869856164e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1510758.0,
      "step": 3462
    },
    {
      "epoch": 0.2663641258364741,
      "grad_norm": 6.5625,
      "learning_rate": 2.6004538112452886e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1511092.0,
      "step": 3463
    },
    {
      "epoch": 0.2664410429966926,
      "grad_norm": 7.84375,
      "learning_rate": 2.6003384355049613e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 1511444.0,
      "step": 3464
    },
    {
      "epoch": 0.266517960156911,
      "grad_norm": 8.125,
      "learning_rate": 2.6002230597646336e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1511706.0,
      "step": 3465
    },
    {
      "epoch": 0.26659487731712944,
      "grad_norm": 5.25,
      "learning_rate": 2.600107684024306e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 1512008.0,
      "step": 3466
    },
    {
      "epoch": 0.2666717944773479,
      "grad_norm": 4.3125,
      "learning_rate": 2.5999923082839785e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1512328.0,
      "step": 3467
    },
    {
      "epoch": 0.26674871163756636,
      "grad_norm": 2.53125,
      "learning_rate": 2.5998769325436504e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1513108.0,
      "step": 3468
    },
    {
      "epoch": 0.26682562879778476,
      "grad_norm": 3.25,
      "learning_rate": 2.599761556803323e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1513686.0,
      "step": 3469
    },
    {
      "epoch": 0.2669025459580032,
      "grad_norm": 2.375,
      "learning_rate": 2.5996461810629953e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 1514494.0,
      "step": 3470
    },
    {
      "epoch": 0.2669794631182217,
      "grad_norm": 2.53125,
      "learning_rate": 2.5995308053226676e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1515276.0,
      "step": 3471
    },
    {
      "epoch": 0.26705638027844014,
      "grad_norm": 2.21875,
      "learning_rate": 2.59941542958234e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1516056.0,
      "step": 3472
    },
    {
      "epoch": 0.26713329743865855,
      "grad_norm": 2.453125,
      "learning_rate": 2.599300053842012e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 1516812.0,
      "step": 3473
    },
    {
      "epoch": 0.267210214598877,
      "grad_norm": 3.171875,
      "learning_rate": 2.5991846781016843e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 1517400.0,
      "step": 3474
    },
    {
      "epoch": 0.26728713175909546,
      "grad_norm": 7.84375,
      "learning_rate": 2.599069302361357e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1517738.0,
      "step": 3475
    },
    {
      "epoch": 0.2673640489193139,
      "grad_norm": 4.3125,
      "learning_rate": 2.5989539266210293e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1518042.0,
      "step": 3476
    },
    {
      "epoch": 0.2674409660795323,
      "grad_norm": 2.03125,
      "learning_rate": 2.5988385508807015e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 1519086.0,
      "step": 3477
    },
    {
      "epoch": 0.2675178832397508,
      "grad_norm": 1.875,
      "learning_rate": 2.598723175140374e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9987593293190002,
      "num_tokens": 1519894.0,
      "step": 3478
    },
    {
      "epoch": 0.26759480039996925,
      "grad_norm": 2.03125,
      "learning_rate": 2.598607799400046e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1520462.0,
      "step": 3479
    },
    {
      "epoch": 0.2676717175601877,
      "grad_norm": 6.0,
      "learning_rate": 2.5984924236597183e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 1520756.0,
      "step": 3480
    },
    {
      "epoch": 0.2677486347204061,
      "grad_norm": 6.0625,
      "learning_rate": 2.598377047919391e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1521050.0,
      "step": 3481
    },
    {
      "epoch": 0.26782555188062457,
      "grad_norm": 9.625,
      "learning_rate": 2.5982616721790632e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 1521428.0,
      "step": 3482
    },
    {
      "epoch": 0.267902469040843,
      "grad_norm": 5.71875,
      "learning_rate": 2.5981462964387355e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1521740.0,
      "step": 3483
    },
    {
      "epoch": 0.26797938620106143,
      "grad_norm": 5.84375,
      "learning_rate": 2.5980309206984078e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 1522034.0,
      "step": 3484
    },
    {
      "epoch": 0.2680563033612799,
      "grad_norm": 3.453125,
      "learning_rate": 2.59791554495808e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1522576.0,
      "step": 3485
    },
    {
      "epoch": 0.26813322052149835,
      "grad_norm": 3.671875,
      "learning_rate": 2.5978001692177527e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1523156.0,
      "step": 3486
    },
    {
      "epoch": 0.2682101376817168,
      "grad_norm": 3.234375,
      "learning_rate": 2.597684793477425e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1523752.0,
      "step": 3487
    },
    {
      "epoch": 0.2682870548419352,
      "grad_norm": 2.671875,
      "learning_rate": 2.5975694177370972e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1524274.0,
      "step": 3488
    },
    {
      "epoch": 0.2683639720021537,
      "grad_norm": 5.21875,
      "learning_rate": 2.5974540419967695e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1524630.0,
      "step": 3489
    },
    {
      "epoch": 0.26844088916237213,
      "grad_norm": 7.65625,
      "learning_rate": 2.5973386662564417e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 1524948.0,
      "step": 3490
    },
    {
      "epoch": 0.2685178063225906,
      "grad_norm": 2.015625,
      "learning_rate": 2.597223290516114e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 1525734.0,
      "step": 3491
    },
    {
      "epoch": 0.268594723482809,
      "grad_norm": 5.0625,
      "learning_rate": 2.5971079147757867e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1526068.0,
      "step": 3492
    },
    {
      "epoch": 0.26867164064302745,
      "grad_norm": 12.0,
      "learning_rate": 2.596992539035459e-06,
      "loss": 0.0428,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 1526300.0,
      "step": 3493
    },
    {
      "epoch": 0.2687485578032459,
      "grad_norm": 3.734375,
      "learning_rate": 2.596877163295131e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1526622.0,
      "step": 3494
    },
    {
      "epoch": 0.2688254749634644,
      "grad_norm": 7.34375,
      "learning_rate": 2.5967617875548034e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9826989769935608,
      "num_tokens": 1526912.0,
      "step": 3495
    },
    {
      "epoch": 0.2689023921236828,
      "grad_norm": 2.046875,
      "learning_rate": 2.5966464118144757e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1527502.0,
      "step": 3496
    },
    {
      "epoch": 0.26897930928390124,
      "grad_norm": 2.046875,
      "learning_rate": 2.5965310360741484e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1528092.0,
      "step": 3497
    },
    {
      "epoch": 0.2690562264441197,
      "grad_norm": 4.21875,
      "learning_rate": 2.5964156603338206e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1528608.0,
      "step": 3498
    },
    {
      "epoch": 0.26913314360433815,
      "grad_norm": 3.125,
      "learning_rate": 2.596300284593493e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1528930.0,
      "step": 3499
    },
    {
      "epoch": 0.26921006076455656,
      "grad_norm": 4.6875,
      "learning_rate": 2.596184908853165e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 1529218.0,
      "step": 3500
    },
    {
      "epoch": 0.269286977924775,
      "grad_norm": 7.09375,
      "learning_rate": 2.5960695331128374e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 1529566.0,
      "step": 3501
    },
    {
      "epoch": 0.2693638950849935,
      "grad_norm": 2.453125,
      "learning_rate": 2.5959541573725097e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1530122.0,
      "step": 3502
    },
    {
      "epoch": 0.2694408122452119,
      "grad_norm": 8.875,
      "learning_rate": 2.5958387816321824e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 1530376.0,
      "step": 3503
    },
    {
      "epoch": 0.26951772940543034,
      "grad_norm": 5.15625,
      "learning_rate": 2.5957234058918546e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 1530868.0,
      "step": 3504
    },
    {
      "epoch": 0.2695946465656488,
      "grad_norm": 6.3125,
      "learning_rate": 2.595608030151527e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 1531214.0,
      "step": 3505
    },
    {
      "epoch": 0.26967156372586726,
      "grad_norm": 6.65625,
      "learning_rate": 2.595492654411199e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1531522.0,
      "step": 3506
    },
    {
      "epoch": 0.26974848088608566,
      "grad_norm": 5.90625,
      "learning_rate": 2.5953772786708714e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1531854.0,
      "step": 3507
    },
    {
      "epoch": 0.2698253980463041,
      "grad_norm": 2.96875,
      "learning_rate": 2.595261902930544e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1532432.0,
      "step": 3508
    },
    {
      "epoch": 0.2699023152065226,
      "grad_norm": 3.140625,
      "learning_rate": 2.5951465271902163e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1532984.0,
      "step": 3509
    },
    {
      "epoch": 0.26997923236674104,
      "grad_norm": 5.96875,
      "learning_rate": 2.5950311514498886e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1533320.0,
      "step": 3510
    },
    {
      "epoch": 0.27005614952695944,
      "grad_norm": 6.46875,
      "learning_rate": 2.594915775709561e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1533650.0,
      "step": 3511
    },
    {
      "epoch": 0.2701330666871779,
      "grad_norm": 4.03125,
      "learning_rate": 2.594800399969233e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1533962.0,
      "step": 3512
    },
    {
      "epoch": 0.27020998384739636,
      "grad_norm": 5.5,
      "learning_rate": 2.5946850242289054e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 1534298.0,
      "step": 3513
    },
    {
      "epoch": 0.2702869010076148,
      "grad_norm": 5.90625,
      "learning_rate": 2.594569648488578e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1534630.0,
      "step": 3514
    },
    {
      "epoch": 0.2703638181678332,
      "grad_norm": 6.5,
      "learning_rate": 2.5944542727482503e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1534916.0,
      "step": 3515
    },
    {
      "epoch": 0.2704407353280517,
      "grad_norm": 4.46875,
      "learning_rate": 2.5943388970079226e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1535260.0,
      "step": 3516
    },
    {
      "epoch": 0.27051765248827014,
      "grad_norm": 6.59375,
      "learning_rate": 2.594223521267595e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1535578.0,
      "step": 3517
    },
    {
      "epoch": 0.2705945696484886,
      "grad_norm": 6.4375,
      "learning_rate": 2.594108145527267e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1535896.0,
      "step": 3518
    },
    {
      "epoch": 0.270671486808707,
      "grad_norm": 6.21875,
      "learning_rate": 2.5939927697869398e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1536202.0,
      "step": 3519
    },
    {
      "epoch": 0.27074840396892547,
      "grad_norm": 5.21875,
      "learning_rate": 2.593877394046612e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 1536536.0,
      "step": 3520
    },
    {
      "epoch": 0.2708253211291439,
      "grad_norm": 2.453125,
      "learning_rate": 2.5937620183062843e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1537362.0,
      "step": 3521
    },
    {
      "epoch": 0.27090223828936233,
      "grad_norm": 3.734375,
      "learning_rate": 2.5936466425659565e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1537670.0,
      "step": 3522
    },
    {
      "epoch": 0.2709791554495808,
      "grad_norm": 8.0625,
      "learning_rate": 2.593531266825629e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.989276111125946,
      "num_tokens": 1538046.0,
      "step": 3523
    },
    {
      "epoch": 0.27105607260979925,
      "grad_norm": 4.125,
      "learning_rate": 2.593415891085301e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1538566.0,
      "step": 3524
    },
    {
      "epoch": 0.2711329897700177,
      "grad_norm": 10.5625,
      "learning_rate": 2.5933005153449737e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9858155846595764,
      "num_tokens": 1538850.0,
      "step": 3525
    },
    {
      "epoch": 0.2712099069302361,
      "grad_norm": 2.34375,
      "learning_rate": 2.593185139604646e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1539636.0,
      "step": 3526
    },
    {
      "epoch": 0.27128682409045457,
      "grad_norm": 5.46875,
      "learning_rate": 2.5930697638643182e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1539952.0,
      "step": 3527
    },
    {
      "epoch": 0.27136374125067303,
      "grad_norm": 4.71875,
      "learning_rate": 2.5929543881239905e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1540292.0,
      "step": 3528
    },
    {
      "epoch": 0.2714406584108915,
      "grad_norm": 2.921875,
      "learning_rate": 2.5928390123836628e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 1541048.0,
      "step": 3529
    },
    {
      "epoch": 0.2715175755711099,
      "grad_norm": 2.390625,
      "learning_rate": 2.592723636643335e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 1541828.0,
      "step": 3530
    },
    {
      "epoch": 0.27159449273132835,
      "grad_norm": 5.1875,
      "learning_rate": 2.5926082609030077e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1542136.0,
      "step": 3531
    },
    {
      "epoch": 0.2716714098915468,
      "grad_norm": 2.5,
      "learning_rate": 2.59249288516268e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1542722.0,
      "step": 3532
    },
    {
      "epoch": 0.27174832705176527,
      "grad_norm": 2.109375,
      "learning_rate": 2.5923775094223522e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1543264.0,
      "step": 3533
    },
    {
      "epoch": 0.2718252442119837,
      "grad_norm": 6.8125,
      "learning_rate": 2.5922621336820245e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1543600.0,
      "step": 3534
    },
    {
      "epoch": 0.27190216137220213,
      "grad_norm": 3.59375,
      "learning_rate": 2.5921467579416967e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1543922.0,
      "step": 3535
    },
    {
      "epoch": 0.2719790785324206,
      "grad_norm": 6.53125,
      "learning_rate": 2.5920313822013694e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1544256.0,
      "step": 3536
    },
    {
      "epoch": 0.27205599569263905,
      "grad_norm": 3.375,
      "learning_rate": 2.5919160064610417e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9908925294876099,
      "num_tokens": 1544806.0,
      "step": 3537
    },
    {
      "epoch": 0.27213291285285746,
      "grad_norm": 4.8125,
      "learning_rate": 2.5918006307207135e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1545108.0,
      "step": 3538
    },
    {
      "epoch": 0.2722098300130759,
      "grad_norm": 2.46875,
      "learning_rate": 2.591685254980386e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1545676.0,
      "step": 3539
    },
    {
      "epoch": 0.2722867471732944,
      "grad_norm": 2.359375,
      "learning_rate": 2.5915698792400585e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1546240.0,
      "step": 3540
    },
    {
      "epoch": 0.27236366433351283,
      "grad_norm": 3.46875,
      "learning_rate": 2.5914545034997307e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9915110468864441,
      "num_tokens": 1546830.0,
      "step": 3541
    },
    {
      "epoch": 0.27244058149373124,
      "grad_norm": 6.375,
      "learning_rate": 2.5913391277594034e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1547186.0,
      "step": 3542
    },
    {
      "epoch": 0.2725174986539497,
      "grad_norm": 5.96875,
      "learning_rate": 2.5912237520190756e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 1547500.0,
      "step": 3543
    },
    {
      "epoch": 0.27259441581416816,
      "grad_norm": 8.8125,
      "learning_rate": 2.591108376278748e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 1547864.0,
      "step": 3544
    },
    {
      "epoch": 0.27267133297438656,
      "grad_norm": 4.46875,
      "learning_rate": 2.59099300053842e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1548198.0,
      "step": 3545
    },
    {
      "epoch": 0.272748250134605,
      "grad_norm": 3.359375,
      "learning_rate": 2.5908776247980924e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 1548794.0,
      "step": 3546
    },
    {
      "epoch": 0.2728251672948235,
      "grad_norm": 3.4375,
      "learning_rate": 2.590762249057765e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 1549348.0,
      "step": 3547
    },
    {
      "epoch": 0.27290208445504194,
      "grad_norm": 6.46875,
      "learning_rate": 2.5906468733174374e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1549684.0,
      "step": 3548
    },
    {
      "epoch": 0.27297900161526034,
      "grad_norm": 4.625,
      "learning_rate": 2.590531497577109e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 1550014.0,
      "step": 3549
    },
    {
      "epoch": 0.2730559187754788,
      "grad_norm": 8.8125,
      "learning_rate": 2.590416121836782e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 1550324.0,
      "step": 3550
    },
    {
      "epoch": 0.27313283593569726,
      "grad_norm": 4.21875,
      "learning_rate": 2.590300746096454e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 1550650.0,
      "step": 3551
    },
    {
      "epoch": 0.2732097530959157,
      "grad_norm": 4.3125,
      "learning_rate": 2.5901853703561264e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1551222.0,
      "step": 3552
    },
    {
      "epoch": 0.2732866702561341,
      "grad_norm": 2.59375,
      "learning_rate": 2.590069994615799e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1551770.0,
      "step": 3553
    },
    {
      "epoch": 0.2733635874163526,
      "grad_norm": 7.6875,
      "learning_rate": 2.589954618875471e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 1552078.0,
      "step": 3554
    },
    {
      "epoch": 0.27344050457657104,
      "grad_norm": 5.4375,
      "learning_rate": 2.5898392431351436e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1552394.0,
      "step": 3555
    },
    {
      "epoch": 0.2735174217367895,
      "grad_norm": 5.0625,
      "learning_rate": 2.589723867394816e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1552712.0,
      "step": 3556
    },
    {
      "epoch": 0.2735943388970079,
      "grad_norm": 2.546875,
      "learning_rate": 2.589608491654488e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1553274.0,
      "step": 3557
    },
    {
      "epoch": 0.27367125605722636,
      "grad_norm": 8.4375,
      "learning_rate": 2.589493115914161e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 1553552.0,
      "step": 3558
    },
    {
      "epoch": 0.2737481732174448,
      "grad_norm": 2.09375,
      "learning_rate": 2.589377740173833e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1554136.0,
      "step": 3559
    },
    {
      "epoch": 0.2738250903776633,
      "grad_norm": 6.21875,
      "learning_rate": 2.589262364433505e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1554470.0,
      "step": 3560
    },
    {
      "epoch": 0.2739020075378817,
      "grad_norm": 4.46875,
      "learning_rate": 2.5891469886931776e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 1554780.0,
      "step": 3561
    },
    {
      "epoch": 0.27397892469810015,
      "grad_norm": 4.84375,
      "learning_rate": 2.58903161295285e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1555090.0,
      "step": 3562
    },
    {
      "epoch": 0.2740558418583186,
      "grad_norm": 7.0,
      "learning_rate": 2.588916237212522e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1555372.0,
      "step": 3563
    },
    {
      "epoch": 0.274132759018537,
      "grad_norm": 2.125,
      "learning_rate": 2.5888008614721948e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1555946.0,
      "step": 3564
    },
    {
      "epoch": 0.27420967617875547,
      "grad_norm": 5.53125,
      "learning_rate": 2.5886854857318666e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1556282.0,
      "step": 3565
    },
    {
      "epoch": 0.2742865933389739,
      "grad_norm": 2.578125,
      "learning_rate": 2.5885701099915393e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1556828.0,
      "step": 3566
    },
    {
      "epoch": 0.2743635104991924,
      "grad_norm": 5.46875,
      "learning_rate": 2.5884547342512115e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 1557164.0,
      "step": 3567
    },
    {
      "epoch": 0.2744404276594108,
      "grad_norm": 5.125,
      "learning_rate": 2.588339358510884e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1557464.0,
      "step": 3568
    },
    {
      "epoch": 0.27451734481962925,
      "grad_norm": 2.171875,
      "learning_rate": 2.5882239827705565e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 1558244.0,
      "step": 3569
    },
    {
      "epoch": 0.2745942619798477,
      "grad_norm": 2.359375,
      "learning_rate": 2.5881086070302287e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1558792.0,
      "step": 3570
    },
    {
      "epoch": 0.27467117914006617,
      "grad_norm": 6.78125,
      "learning_rate": 2.5879932312899006e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 1559122.0,
      "step": 3571
    },
    {
      "epoch": 0.27474809630028457,
      "grad_norm": 2.40625,
      "learning_rate": 2.5878778555495733e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1559942.0,
      "step": 3572
    },
    {
      "epoch": 0.27482501346050303,
      "grad_norm": 2.390625,
      "learning_rate": 2.5877624798092455e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1560762.0,
      "step": 3573
    },
    {
      "epoch": 0.2749019306207215,
      "grad_norm": 6.21875,
      "learning_rate": 2.5876471040689178e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 1561110.0,
      "step": 3574
    },
    {
      "epoch": 0.27497884778093995,
      "grad_norm": 1.6328125,
      "learning_rate": 2.5875317283285904e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1561656.0,
      "step": 3575
    },
    {
      "epoch": 0.27505576494115835,
      "grad_norm": 6.71875,
      "learning_rate": 2.5874163525882623e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 1561946.0,
      "step": 3576
    },
    {
      "epoch": 0.2751326821013768,
      "grad_norm": 3.5625,
      "learning_rate": 2.587300976847935e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1562292.0,
      "step": 3577
    },
    {
      "epoch": 0.2752095992615953,
      "grad_norm": 3.46875,
      "learning_rate": 2.5871856011076072e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1562598.0,
      "step": 3578
    },
    {
      "epoch": 0.27528651642181373,
      "grad_norm": 5.6875,
      "learning_rate": 2.5870702253672795e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1562904.0,
      "step": 3579
    },
    {
      "epoch": 0.27536343358203214,
      "grad_norm": 5.34375,
      "learning_rate": 2.5869548496269517e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 1563234.0,
      "step": 3580
    },
    {
      "epoch": 0.2754403507422506,
      "grad_norm": 5.0,
      "learning_rate": 2.586839473886624e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1563554.0,
      "step": 3581
    },
    {
      "epoch": 0.27551726790246905,
      "grad_norm": 2.84375,
      "learning_rate": 2.5867240981462963e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 1564108.0,
      "step": 3582
    },
    {
      "epoch": 0.27559418506268746,
      "grad_norm": 2.78125,
      "learning_rate": 2.586608722405969e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1564676.0,
      "step": 3583
    },
    {
      "epoch": 0.2756711022229059,
      "grad_norm": 6.28125,
      "learning_rate": 2.586493346665641e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1564958.0,
      "step": 3584
    },
    {
      "epoch": 0.2757480193831244,
      "grad_norm": 3.671875,
      "learning_rate": 2.5863779709253135e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1565266.0,
      "step": 3585
    },
    {
      "epoch": 0.27582493654334284,
      "grad_norm": 3.75,
      "learning_rate": 2.586262595184986e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1565612.0,
      "step": 3586
    },
    {
      "epoch": 0.27590185370356124,
      "grad_norm": 9.625,
      "learning_rate": 2.586147219444658e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 1565898.0,
      "step": 3587
    },
    {
      "epoch": 0.2759787708637797,
      "grad_norm": 4.5,
      "learning_rate": 2.5860318437043302e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1566210.0,
      "step": 3588
    },
    {
      "epoch": 0.27605568802399816,
      "grad_norm": 2.421875,
      "learning_rate": 2.585916467964003e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1566758.0,
      "step": 3589
    },
    {
      "epoch": 0.2761326051842166,
      "grad_norm": 4.46875,
      "learning_rate": 2.585801092223675e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 1567094.0,
      "step": 3590
    },
    {
      "epoch": 0.276209522344435,
      "grad_norm": 2.578125,
      "learning_rate": 2.5856857164833474e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 1567892.0,
      "step": 3591
    },
    {
      "epoch": 0.2762864395046535,
      "grad_norm": 4.875,
      "learning_rate": 2.5855703407430197e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 1568222.0,
      "step": 3592
    },
    {
      "epoch": 0.27636335666487194,
      "grad_norm": 4.78125,
      "learning_rate": 2.585454965002692e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1568542.0,
      "step": 3593
    },
    {
      "epoch": 0.2764402738250904,
      "grad_norm": 2.4375,
      "learning_rate": 2.5853395892623646e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 1569346.0,
      "step": 3594
    },
    {
      "epoch": 0.2765171909853088,
      "grad_norm": 5.21875,
      "learning_rate": 2.585224213522037e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1569652.0,
      "step": 3595
    },
    {
      "epoch": 0.27659410814552726,
      "grad_norm": 7.25,
      "learning_rate": 2.585108837781709e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 1569960.0,
      "step": 3596
    },
    {
      "epoch": 0.2766710253057457,
      "grad_norm": 2.171875,
      "learning_rate": 2.584993462041382e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1570508.0,
      "step": 3597
    },
    {
      "epoch": 0.2767479424659642,
      "grad_norm": 5.75,
      "learning_rate": 2.5848780863010537e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1570790.0,
      "step": 3598
    },
    {
      "epoch": 0.2768248596261826,
      "grad_norm": 3.296875,
      "learning_rate": 2.584762710560726e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1571358.0,
      "step": 3599
    },
    {
      "epoch": 0.27690177678640104,
      "grad_norm": 2.484375,
      "learning_rate": 2.5846473348203986e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1571910.0,
      "step": 3600
    },
    {
      "epoch": 0.2769786939466195,
      "grad_norm": 5.84375,
      "learning_rate": 2.584531959080071e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1572254.0,
      "step": 3601
    },
    {
      "epoch": 0.27705561110683796,
      "grad_norm": 3.171875,
      "learning_rate": 2.584416583339743e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 1572830.0,
      "step": 3602
    },
    {
      "epoch": 0.27713252826705637,
      "grad_norm": 3.890625,
      "learning_rate": 2.5843012075994154e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1573140.0,
      "step": 3603
    },
    {
      "epoch": 0.2772094454272748,
      "grad_norm": 3.859375,
      "learning_rate": 2.5841858318590876e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1573634.0,
      "step": 3604
    },
    {
      "epoch": 0.2772863625874933,
      "grad_norm": 5.59375,
      "learning_rate": 2.5840704561187603e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1573934.0,
      "step": 3605
    },
    {
      "epoch": 0.2773632797477117,
      "grad_norm": 5.9375,
      "learning_rate": 2.5839550803784326e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1574216.0,
      "step": 3606
    },
    {
      "epoch": 0.27744019690793015,
      "grad_norm": 2.5,
      "learning_rate": 2.583839704638105e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1574740.0,
      "step": 3607
    },
    {
      "epoch": 0.2775171140681486,
      "grad_norm": 2.640625,
      "learning_rate": 2.5837243288977775e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1575566.0,
      "step": 3608
    },
    {
      "epoch": 0.27759403122836707,
      "grad_norm": 8.375,
      "learning_rate": 2.5836089531574493e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 1575934.0,
      "step": 3609
    },
    {
      "epoch": 0.27767094838858547,
      "grad_norm": 1.796875,
      "learning_rate": 2.5834935774171216e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1576720.0,
      "step": 3610
    },
    {
      "epoch": 0.27774786554880393,
      "grad_norm": 8.0,
      "learning_rate": 2.5833782016767943e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 1577020.0,
      "step": 3611
    },
    {
      "epoch": 0.2778247827090224,
      "grad_norm": 7.3125,
      "learning_rate": 2.5832628259364665e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 1577324.0,
      "step": 3612
    },
    {
      "epoch": 0.27790169986924085,
      "grad_norm": 3.28125,
      "learning_rate": 2.583147450196139e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1577874.0,
      "step": 3613
    },
    {
      "epoch": 0.27797861702945925,
      "grad_norm": 9.1875,
      "learning_rate": 2.583032074455811e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 1578198.0,
      "step": 3614
    },
    {
      "epoch": 0.2780555341896777,
      "grad_norm": 4.96875,
      "learning_rate": 2.5829166987154833e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 1578526.0,
      "step": 3615
    },
    {
      "epoch": 0.27813245134989617,
      "grad_norm": 2.0,
      "learning_rate": 2.582801322975156e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 1579310.0,
      "step": 3616
    },
    {
      "epoch": 0.27820936851011463,
      "grad_norm": 6.625,
      "learning_rate": 2.5826859472348283e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1579626.0,
      "step": 3617
    },
    {
      "epoch": 0.27828628567033303,
      "grad_norm": 8.875,
      "learning_rate": 2.5825705714945005e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 1579902.0,
      "step": 3618
    },
    {
      "epoch": 0.2783632028305515,
      "grad_norm": 4.59375,
      "learning_rate": 2.5824551957541728e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1580200.0,
      "step": 3619
    },
    {
      "epoch": 0.27844011999076995,
      "grad_norm": 3.03125,
      "learning_rate": 2.582339820013845e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9959431886672974,
      "num_tokens": 1580694.0,
      "step": 3620
    },
    {
      "epoch": 0.2785170371509884,
      "grad_norm": 5.75,
      "learning_rate": 2.5822244442735173e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1581042.0,
      "step": 3621
    },
    {
      "epoch": 0.2785939543112068,
      "grad_norm": 4.9375,
      "learning_rate": 2.58210906853319e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1581324.0,
      "step": 3622
    },
    {
      "epoch": 0.2786708714714253,
      "grad_norm": 7.40625,
      "learning_rate": 2.5819936927928622e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 1581580.0,
      "step": 3623
    },
    {
      "epoch": 0.27874778863164373,
      "grad_norm": 3.78125,
      "learning_rate": 2.5818783170525345e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 1582152.0,
      "step": 3624
    },
    {
      "epoch": 0.27882470579186214,
      "grad_norm": 4.53125,
      "learning_rate": 2.5817629413122067e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 1582754.0,
      "step": 3625
    },
    {
      "epoch": 0.2789016229520806,
      "grad_norm": 3.390625,
      "learning_rate": 2.581647565571879e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 1583342.0,
      "step": 3626
    },
    {
      "epoch": 0.27897854011229906,
      "grad_norm": 2.578125,
      "learning_rate": 2.5815321898315517e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 1583866.0,
      "step": 3627
    },
    {
      "epoch": 0.2790554572725175,
      "grad_norm": 1.8828125,
      "learning_rate": 2.581416814091224e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9973474740982056,
      "num_tokens": 1584622.0,
      "step": 3628
    },
    {
      "epoch": 0.2791323744327359,
      "grad_norm": 5.625,
      "learning_rate": 2.581301438350896e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1584938.0,
      "step": 3629
    },
    {
      "epoch": 0.2792092915929544,
      "grad_norm": 5.78125,
      "learning_rate": 2.5811860626105685e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1585222.0,
      "step": 3630
    },
    {
      "epoch": 0.27928620875317284,
      "grad_norm": 3.421875,
      "learning_rate": 2.5810706868702407e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1585758.0,
      "step": 3631
    },
    {
      "epoch": 0.2793631259133913,
      "grad_norm": 3.296875,
      "learning_rate": 2.580955311129913e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 1586332.0,
      "step": 3632
    },
    {
      "epoch": 0.2794400430736097,
      "grad_norm": 2.9375,
      "learning_rate": 2.5808399353895857e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1586896.0,
      "step": 3633
    },
    {
      "epoch": 0.27951696023382816,
      "grad_norm": 3.40625,
      "learning_rate": 2.580724559649258e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1587216.0,
      "step": 3634
    },
    {
      "epoch": 0.2795938773940466,
      "grad_norm": 3.3125,
      "learning_rate": 2.58060918390893e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1587784.0,
      "step": 3635
    },
    {
      "epoch": 0.2796707945542651,
      "grad_norm": 2.03125,
      "learning_rate": 2.5804938081686024e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1588346.0,
      "step": 3636
    },
    {
      "epoch": 0.2797477117144835,
      "grad_norm": 6.71875,
      "learning_rate": 2.5803784324282747e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 1588690.0,
      "step": 3637
    },
    {
      "epoch": 0.27982462887470194,
      "grad_norm": 1.8359375,
      "learning_rate": 2.580263056687947e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1589274.0,
      "step": 3638
    },
    {
      "epoch": 0.2799015460349204,
      "grad_norm": 1.7578125,
      "learning_rate": 2.5801476809476196e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1589836.0,
      "step": 3639
    },
    {
      "epoch": 0.27997846319513886,
      "grad_norm": 5.78125,
      "learning_rate": 2.580032305207292e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1590136.0,
      "step": 3640
    },
    {
      "epoch": 0.28005538035535726,
      "grad_norm": 8.6875,
      "learning_rate": 2.579916929466964e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1590398.0,
      "step": 3641
    },
    {
      "epoch": 0.2801322975155757,
      "grad_norm": 2.484375,
      "learning_rate": 2.5798015537266364e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9948892593383789,
      "num_tokens": 1590986.0,
      "step": 3642
    },
    {
      "epoch": 0.2802092146757942,
      "grad_norm": 7.0625,
      "learning_rate": 2.5796861779863087e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 1591318.0,
      "step": 3643
    },
    {
      "epoch": 0.28028613183601264,
      "grad_norm": 4.1875,
      "learning_rate": 2.5795708022459813e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1591638.0,
      "step": 3644
    },
    {
      "epoch": 0.28036304899623105,
      "grad_norm": 5.53125,
      "learning_rate": 2.5794554265056536e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 1592000.0,
      "step": 3645
    },
    {
      "epoch": 0.2804399661564495,
      "grad_norm": 4.125,
      "learning_rate": 2.579340050765326e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1592540.0,
      "step": 3646
    },
    {
      "epoch": 0.28051688331666796,
      "grad_norm": 5.1875,
      "learning_rate": 2.579224675024998e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 1592892.0,
      "step": 3647
    },
    {
      "epoch": 0.28059380047688637,
      "grad_norm": 2.171875,
      "learning_rate": 2.5791092992846704e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948892593383789,
      "num_tokens": 1593480.0,
      "step": 3648
    },
    {
      "epoch": 0.2806707176371048,
      "grad_norm": 6.71875,
      "learning_rate": 2.5789939235443426e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1593776.0,
      "step": 3649
    },
    {
      "epoch": 0.2807476347973233,
      "grad_norm": 4.65625,
      "learning_rate": 2.5788785478040153e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1594088.0,
      "step": 3650
    },
    {
      "epoch": 0.28082455195754175,
      "grad_norm": 5.875,
      "learning_rate": 2.5787631720636876e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 1594404.0,
      "step": 3651
    },
    {
      "epoch": 0.28090146911776015,
      "grad_norm": 1.984375,
      "learning_rate": 2.57864779632336e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1594966.0,
      "step": 3652
    },
    {
      "epoch": 0.2809783862779786,
      "grad_norm": 5.15625,
      "learning_rate": 2.578532420583032e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1595278.0,
      "step": 3653
    },
    {
      "epoch": 0.28105530343819707,
      "grad_norm": 4.65625,
      "learning_rate": 2.5784170448427043e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1595612.0,
      "step": 3654
    },
    {
      "epoch": 0.2811322205984155,
      "grad_norm": 9.0625,
      "learning_rate": 2.578301669102377e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1595918.0,
      "step": 3655
    },
    {
      "epoch": 0.28120913775863393,
      "grad_norm": 3.0,
      "learning_rate": 2.5781862933620493e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1596476.0,
      "step": 3656
    },
    {
      "epoch": 0.2812860549188524,
      "grad_norm": 1.9609375,
      "learning_rate": 2.578070917621721e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1597028.0,
      "step": 3657
    },
    {
      "epoch": 0.28136297207907085,
      "grad_norm": 3.0625,
      "learning_rate": 2.577955541881394e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1597572.0,
      "step": 3658
    },
    {
      "epoch": 0.2814398892392893,
      "grad_norm": 4.8125,
      "learning_rate": 2.577840166141066e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1597870.0,
      "step": 3659
    },
    {
      "epoch": 0.2815168063995077,
      "grad_norm": 8.6875,
      "learning_rate": 2.5777247904007383e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 1598126.0,
      "step": 3660
    },
    {
      "epoch": 0.2815937235597262,
      "grad_norm": 2.65625,
      "learning_rate": 2.577609414660411e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1598946.0,
      "step": 3661
    },
    {
      "epoch": 0.28167064071994463,
      "grad_norm": 3.0,
      "learning_rate": 2.5774940389200833e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1599532.0,
      "step": 3662
    },
    {
      "epoch": 0.2817475578801631,
      "grad_norm": 8.3125,
      "learning_rate": 2.5773786631797555e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1599810.0,
      "step": 3663
    },
    {
      "epoch": 0.2818244750403815,
      "grad_norm": 3.328125,
      "learning_rate": 2.5772632874394278e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1600360.0,
      "step": 3664
    },
    {
      "epoch": 0.28190139220059995,
      "grad_norm": 2.625,
      "learning_rate": 2.5771479116991e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1600918.0,
      "step": 3665
    },
    {
      "epoch": 0.2819783093608184,
      "grad_norm": 4.25,
      "learning_rate": 2.5770325359587727e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1601262.0,
      "step": 3666
    },
    {
      "epoch": 0.2820552265210368,
      "grad_norm": 4.3125,
      "learning_rate": 2.576917160218445e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 1601862.0,
      "step": 3667
    },
    {
      "epoch": 0.2821321436812553,
      "grad_norm": 6.75,
      "learning_rate": 2.576801784478117e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 1602218.0,
      "step": 3668
    },
    {
      "epoch": 0.28220906084147374,
      "grad_norm": 6.40625,
      "learning_rate": 2.5766864087377895e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 1602520.0,
      "step": 3669
    },
    {
      "epoch": 0.2822859780016922,
      "grad_norm": 6.65625,
      "learning_rate": 2.5765710329974617e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 1602886.0,
      "step": 3670
    },
    {
      "epoch": 0.2823628951619106,
      "grad_norm": 3.375,
      "learning_rate": 2.576455657257134e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 1603450.0,
      "step": 3671
    },
    {
      "epoch": 0.28243981232212906,
      "grad_norm": 3.109375,
      "learning_rate": 2.5763402815168067e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 1604046.0,
      "step": 3672
    },
    {
      "epoch": 0.2825167294823475,
      "grad_norm": 7.03125,
      "learning_rate": 2.576224905776479e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1604332.0,
      "step": 3673
    },
    {
      "epoch": 0.282593646642566,
      "grad_norm": 2.015625,
      "learning_rate": 2.576109530036151e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1604918.0,
      "step": 3674
    },
    {
      "epoch": 0.2826705638027844,
      "grad_norm": 8.5625,
      "learning_rate": 2.5759941542958235e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 1605226.0,
      "step": 3675
    },
    {
      "epoch": 0.28274748096300284,
      "grad_norm": 5.375,
      "learning_rate": 2.5758787785554957e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 1605578.0,
      "step": 3676
    },
    {
      "epoch": 0.2828243981232213,
      "grad_norm": 4.15625,
      "learning_rate": 2.5757634028151684e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1605928.0,
      "step": 3677
    },
    {
      "epoch": 0.28290131528343976,
      "grad_norm": 5.3125,
      "learning_rate": 2.5756480270748407e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1606424.0,
      "step": 3678
    },
    {
      "epoch": 0.28297823244365816,
      "grad_norm": 3.1875,
      "learning_rate": 2.5755326513345125e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 1606998.0,
      "step": 3679
    },
    {
      "epoch": 0.2830551496038766,
      "grad_norm": 6.1875,
      "learning_rate": 2.575417275594185e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1607306.0,
      "step": 3680
    },
    {
      "epoch": 0.2831320667640951,
      "grad_norm": 8.0625,
      "learning_rate": 2.5753018998538574e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1607608.0,
      "step": 3681
    },
    {
      "epoch": 0.28320898392431354,
      "grad_norm": 11.1875,
      "learning_rate": 2.5751865241135297e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 1607866.0,
      "step": 3682
    },
    {
      "epoch": 0.28328590108453194,
      "grad_norm": 6.125,
      "learning_rate": 2.5750711483732024e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9915013909339905,
      "num_tokens": 1608220.0,
      "step": 3683
    },
    {
      "epoch": 0.2833628182447504,
      "grad_norm": 3.34375,
      "learning_rate": 2.574955772632874e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 1608760.0,
      "step": 3684
    },
    {
      "epoch": 0.28343973540496886,
      "grad_norm": 2.171875,
      "learning_rate": 2.574840396892547e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 1609584.0,
      "step": 3685
    },
    {
      "epoch": 0.28351665256518727,
      "grad_norm": 2.46875,
      "learning_rate": 2.574725021152219e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1610170.0,
      "step": 3686
    },
    {
      "epoch": 0.2835935697254057,
      "grad_norm": 2.984375,
      "learning_rate": 2.5746096454118914e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1610750.0,
      "step": 3687
    },
    {
      "epoch": 0.2836704868856242,
      "grad_norm": 3.640625,
      "learning_rate": 2.5744942696715637e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1611314.0,
      "step": 3688
    },
    {
      "epoch": 0.28374740404584264,
      "grad_norm": 7.59375,
      "learning_rate": 2.5743788939312363e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1611594.0,
      "step": 3689
    },
    {
      "epoch": 0.28382432120606105,
      "grad_norm": 8.6875,
      "learning_rate": 2.574263518190908e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9967845678329468,
      "num_tokens": 1611908.0,
      "step": 3690
    },
    {
      "epoch": 0.2839012383662795,
      "grad_norm": 4.78125,
      "learning_rate": 2.574148142450581e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 1612196.0,
      "step": 3691
    },
    {
      "epoch": 0.28397815552649797,
      "grad_norm": 2.25,
      "learning_rate": 2.574032766710253e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1612742.0,
      "step": 3692
    },
    {
      "epoch": 0.2840550726867164,
      "grad_norm": 3.3125,
      "learning_rate": 2.5739173909699254e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1613310.0,
      "step": 3693
    },
    {
      "epoch": 0.28413198984693483,
      "grad_norm": 6.75,
      "learning_rate": 2.573802015229598e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9929577708244324,
      "num_tokens": 1613596.0,
      "step": 3694
    },
    {
      "epoch": 0.2842089070071533,
      "grad_norm": 2.109375,
      "learning_rate": 2.57368663948927e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 1614646.0,
      "step": 3695
    },
    {
      "epoch": 0.28428582416737175,
      "grad_norm": 2.03125,
      "learning_rate": 2.573571263748942e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 1615198.0,
      "step": 3696
    },
    {
      "epoch": 0.2843627413275902,
      "grad_norm": 3.296875,
      "learning_rate": 2.573455888008615e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1615762.0,
      "step": 3697
    },
    {
      "epoch": 0.2844396584878086,
      "grad_norm": 3.71875,
      "learning_rate": 2.573340512268287e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1616256.0,
      "step": 3698
    },
    {
      "epoch": 0.28451657564802707,
      "grad_norm": 7.21875,
      "learning_rate": 2.5732251365279593e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1616548.0,
      "step": 3699
    },
    {
      "epoch": 0.28459349280824553,
      "grad_norm": 3.71875,
      "learning_rate": 2.573109760787632e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1616858.0,
      "step": 3700
    },
    {
      "epoch": 0.284670409968464,
      "grad_norm": 7.875,
      "learning_rate": 2.572994385047304e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1617190.0,
      "step": 3701
    },
    {
      "epoch": 0.2847473271286824,
      "grad_norm": 2.546875,
      "learning_rate": 2.5728790093069765e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1617752.0,
      "step": 3702
    },
    {
      "epoch": 0.28482424428890085,
      "grad_norm": 4.34375,
      "learning_rate": 2.572763633566649e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1618036.0,
      "step": 3703
    },
    {
      "epoch": 0.2849011614491193,
      "grad_norm": 7.53125,
      "learning_rate": 2.572648257826321e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1618346.0,
      "step": 3704
    },
    {
      "epoch": 0.28497807860933777,
      "grad_norm": 8.1875,
      "learning_rate": 2.5725328820859937e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1618646.0,
      "step": 3705
    },
    {
      "epoch": 0.2850549957695562,
      "grad_norm": 5.28125,
      "learning_rate": 2.5724175063456656e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1618986.0,
      "step": 3706
    },
    {
      "epoch": 0.28513191292977463,
      "grad_norm": 2.875,
      "learning_rate": 2.572302130605338e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1619550.0,
      "step": 3707
    },
    {
      "epoch": 0.2852088300899931,
      "grad_norm": 2.8125,
      "learning_rate": 2.5721867548650105e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1620114.0,
      "step": 3708
    },
    {
      "epoch": 0.2852857472502115,
      "grad_norm": 3.1875,
      "learning_rate": 2.5720713791246828e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1620662.0,
      "step": 3709
    },
    {
      "epoch": 0.28536266441042996,
      "grad_norm": 7.1875,
      "learning_rate": 2.571956003384355e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9858757257461548,
      "num_tokens": 1621018.0,
      "step": 3710
    },
    {
      "epoch": 0.2854395815706484,
      "grad_norm": 3.203125,
      "learning_rate": 2.5718406276440277e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 1621330.0,
      "step": 3711
    },
    {
      "epoch": 0.2855164987308669,
      "grad_norm": 2.796875,
      "learning_rate": 2.5717252519036995e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9948006868362427,
      "num_tokens": 1621908.0,
      "step": 3712
    },
    {
      "epoch": 0.2855934158910853,
      "grad_norm": 5.1875,
      "learning_rate": 2.5716098761633722e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9930555820465088,
      "num_tokens": 1622198.0,
      "step": 3713
    },
    {
      "epoch": 0.28567033305130374,
      "grad_norm": 2.609375,
      "learning_rate": 2.5714945004230445e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 1623010.0,
      "step": 3714
    },
    {
      "epoch": 0.2857472502115222,
      "grad_norm": 6.96875,
      "learning_rate": 2.5713791246827167e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1623316.0,
      "step": 3715
    },
    {
      "epoch": 0.28582416737174066,
      "grad_norm": 3.640625,
      "learning_rate": 2.5712637489423894e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1623862.0,
      "step": 3716
    },
    {
      "epoch": 0.28590108453195906,
      "grad_norm": 3.84375,
      "learning_rate": 2.5711483732020613e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9907235503196716,
      "num_tokens": 1624402.0,
      "step": 3717
    },
    {
      "epoch": 0.2859780016921775,
      "grad_norm": 8.875,
      "learning_rate": 2.5710329974617335e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1624682.0,
      "step": 3718
    },
    {
      "epoch": 0.286054918852396,
      "grad_norm": 3.03125,
      "learning_rate": 2.570917621721406e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9910233616828918,
      "num_tokens": 1625240.0,
      "step": 3719
    },
    {
      "epoch": 0.28613183601261444,
      "grad_norm": 5.0,
      "learning_rate": 2.5708022459810785e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 1625562.0,
      "step": 3720
    },
    {
      "epoch": 0.28620875317283284,
      "grad_norm": 5.1875,
      "learning_rate": 2.5706868702407507e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1625910.0,
      "step": 3721
    },
    {
      "epoch": 0.2862856703330513,
      "grad_norm": 2.265625,
      "learning_rate": 2.570571494500423e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1626456.0,
      "step": 3722
    },
    {
      "epoch": 0.28636258749326976,
      "grad_norm": 6.15625,
      "learning_rate": 2.5704561187600952e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9857549667358398,
      "num_tokens": 1626808.0,
      "step": 3723
    },
    {
      "epoch": 0.2864395046534882,
      "grad_norm": 4.28125,
      "learning_rate": 2.570340743019768e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 1627134.0,
      "step": 3724
    },
    {
      "epoch": 0.2865164218137066,
      "grad_norm": 6.8125,
      "learning_rate": 2.57022536727944e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 1627442.0,
      "step": 3725
    },
    {
      "epoch": 0.2865933389739251,
      "grad_norm": 3.578125,
      "learning_rate": 2.5701099915391124e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1627766.0,
      "step": 3726
    },
    {
      "epoch": 0.28667025613414354,
      "grad_norm": 5.46875,
      "learning_rate": 2.569994615798785e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 1628106.0,
      "step": 3727
    },
    {
      "epoch": 0.28674717329436195,
      "grad_norm": 2.515625,
      "learning_rate": 2.569879240058457e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 1628696.0,
      "step": 3728
    },
    {
      "epoch": 0.2868240904545804,
      "grad_norm": 2.28125,
      "learning_rate": 2.569763864318129e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1629264.0,
      "step": 3729
    },
    {
      "epoch": 0.28690100761479886,
      "grad_norm": 6.3125,
      "learning_rate": 2.569648488577802e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1629578.0,
      "step": 3730
    },
    {
      "epoch": 0.2869779247750173,
      "grad_norm": 5.125,
      "learning_rate": 2.569533112837474e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 1629944.0,
      "step": 3731
    },
    {
      "epoch": 0.2870548419352357,
      "grad_norm": 7.46875,
      "learning_rate": 2.5694177370971464e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 1630280.0,
      "step": 3732
    },
    {
      "epoch": 0.2871317590954542,
      "grad_norm": 2.5625,
      "learning_rate": 2.5693023613568187e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1630844.0,
      "step": 3733
    },
    {
      "epoch": 0.28720867625567265,
      "grad_norm": 3.6875,
      "learning_rate": 2.569186985616491e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1631408.0,
      "step": 3734
    },
    {
      "epoch": 0.2872855934158911,
      "grad_norm": 2.203125,
      "learning_rate": 2.5690716098761636e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1631992.0,
      "step": 3735
    },
    {
      "epoch": 0.2873625105761095,
      "grad_norm": 2.34375,
      "learning_rate": 2.568956234135836e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 1632776.0,
      "step": 3736
    },
    {
      "epoch": 0.28743942773632797,
      "grad_norm": 3.609375,
      "learning_rate": 2.568840858395508e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1633120.0,
      "step": 3737
    },
    {
      "epoch": 0.2875163448965464,
      "grad_norm": 2.296875,
      "learning_rate": 2.568725482655181e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1633690.0,
      "step": 3738
    },
    {
      "epoch": 0.2875932620567649,
      "grad_norm": 3.515625,
      "learning_rate": 2.5686101069148526e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 1634292.0,
      "step": 3739
    },
    {
      "epoch": 0.2876701792169833,
      "grad_norm": 5.71875,
      "learning_rate": 2.568494731174525e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1634586.0,
      "step": 3740
    },
    {
      "epoch": 0.28774709637720175,
      "grad_norm": 5.34375,
      "learning_rate": 2.5683793554341976e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1635150.0,
      "step": 3741
    },
    {
      "epoch": 0.2878240135374202,
      "grad_norm": 6.09375,
      "learning_rate": 2.56826397969387e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1635478.0,
      "step": 3742
    },
    {
      "epoch": 0.28790093069763867,
      "grad_norm": 6.9375,
      "learning_rate": 2.568148603953542e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 1635774.0,
      "step": 3743
    },
    {
      "epoch": 0.28797784785785707,
      "grad_norm": 4.1875,
      "learning_rate": 2.5680332282132143e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1636096.0,
      "step": 3744
    },
    {
      "epoch": 0.28805476501807553,
      "grad_norm": 1.875,
      "learning_rate": 2.5679178524728866e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1636916.0,
      "step": 3745
    },
    {
      "epoch": 0.288131682178294,
      "grad_norm": 4.46875,
      "learning_rate": 2.5678024767325593e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 1637242.0,
      "step": 3746
    },
    {
      "epoch": 0.2882085993385124,
      "grad_norm": 2.953125,
      "learning_rate": 2.5676871009922315e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1637786.0,
      "step": 3747
    },
    {
      "epoch": 0.28828551649873085,
      "grad_norm": 2.6875,
      "learning_rate": 2.567571725251904e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1638350.0,
      "step": 3748
    },
    {
      "epoch": 0.2883624336589493,
      "grad_norm": 2.59375,
      "learning_rate": 2.567456349511576e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 1639136.0,
      "step": 3749
    },
    {
      "epoch": 0.2884393508191678,
      "grad_norm": 6.65625,
      "learning_rate": 2.5673409737712483e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1639442.0,
      "step": 3750
    },
    {
      "epoch": 0.2885162679793862,
      "grad_norm": 5.09375,
      "learning_rate": 2.5672255980309206e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1639736.0,
      "step": 3751
    },
    {
      "epoch": 0.28859318513960464,
      "grad_norm": 3.046875,
      "learning_rate": 2.5671102222905933e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1640288.0,
      "step": 3752
    },
    {
      "epoch": 0.2886701022998231,
      "grad_norm": 6.34375,
      "learning_rate": 2.5669948465502655e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 1640606.0,
      "step": 3753
    },
    {
      "epoch": 0.28874701946004155,
      "grad_norm": 3.4375,
      "learning_rate": 2.5668794708099378e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 1641206.0,
      "step": 3754
    },
    {
      "epoch": 0.28882393662025996,
      "grad_norm": 6.84375,
      "learning_rate": 2.56676409506961e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1641492.0,
      "step": 3755
    },
    {
      "epoch": 0.2889008537804784,
      "grad_norm": 2.0625,
      "learning_rate": 2.5666487193292823e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 1642306.0,
      "step": 3756
    },
    {
      "epoch": 0.2889777709406969,
      "grad_norm": 4.15625,
      "learning_rate": 2.5665333435889546e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1642630.0,
      "step": 3757
    },
    {
      "epoch": 0.28905468810091534,
      "grad_norm": 5.78125,
      "learning_rate": 2.5664179678486272e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1642964.0,
      "step": 3758
    },
    {
      "epoch": 0.28913160526113374,
      "grad_norm": 4.46875,
      "learning_rate": 2.5663025921082995e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1643284.0,
      "step": 3759
    },
    {
      "epoch": 0.2892085224213522,
      "grad_norm": 8.625,
      "learning_rate": 2.5661872163679717e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1643602.0,
      "step": 3760
    },
    {
      "epoch": 0.28928543958157066,
      "grad_norm": 7.5625,
      "learning_rate": 2.566071840627644e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9937499761581421,
      "num_tokens": 1643926.0,
      "step": 3761
    },
    {
      "epoch": 0.2893623567417891,
      "grad_norm": 3.375,
      "learning_rate": 2.5659564648873163e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1644490.0,
      "step": 3762
    },
    {
      "epoch": 0.2894392739020075,
      "grad_norm": 3.375,
      "learning_rate": 2.565841089146989e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 1645058.0,
      "step": 3763
    },
    {
      "epoch": 0.289516191062226,
      "grad_norm": 2.28125,
      "learning_rate": 2.565725713406661e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1645604.0,
      "step": 3764
    },
    {
      "epoch": 0.28959310822244444,
      "grad_norm": 4.03125,
      "learning_rate": 2.5656103376663335e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1645954.0,
      "step": 3765
    },
    {
      "epoch": 0.2896700253826629,
      "grad_norm": 4.4375,
      "learning_rate": 2.5654949619260057e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 1646530.0,
      "step": 3766
    },
    {
      "epoch": 0.2897469425428813,
      "grad_norm": 3.65625,
      "learning_rate": 2.565379586185678e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1646852.0,
      "step": 3767
    },
    {
      "epoch": 0.28982385970309976,
      "grad_norm": 5.78125,
      "learning_rate": 2.5652642104453502e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1647138.0,
      "step": 3768
    },
    {
      "epoch": 0.2899007768633182,
      "grad_norm": 6.46875,
      "learning_rate": 2.565148834705023e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1647418.0,
      "step": 3769
    },
    {
      "epoch": 0.2899776940235366,
      "grad_norm": 4.71875,
      "learning_rate": 2.565033458964695e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1647734.0,
      "step": 3770
    },
    {
      "epoch": 0.2900546111837551,
      "grad_norm": 7.4375,
      "learning_rate": 2.5649180832243674e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1647996.0,
      "step": 3771
    },
    {
      "epoch": 0.29013152834397354,
      "grad_norm": 3.046875,
      "learning_rate": 2.5648027074840397e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1648560.0,
      "step": 3772
    },
    {
      "epoch": 0.290208445504192,
      "grad_norm": 5.65625,
      "learning_rate": 2.564687331743712e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9970674514770508,
      "num_tokens": 1648902.0,
      "step": 3773
    },
    {
      "epoch": 0.2902853626644104,
      "grad_norm": 5.5625,
      "learning_rate": 2.5645719560033846e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1649184.0,
      "step": 3774
    },
    {
      "epoch": 0.29036227982462887,
      "grad_norm": 8.9375,
      "learning_rate": 2.564456580263057e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 1649556.0,
      "step": 3775
    },
    {
      "epoch": 0.2904391969848473,
      "grad_norm": 5.84375,
      "learning_rate": 2.564341204522729e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1649884.0,
      "step": 3776
    },
    {
      "epoch": 0.2905161141450658,
      "grad_norm": 2.359375,
      "learning_rate": 2.5642258287824014e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1650468.0,
      "step": 3777
    },
    {
      "epoch": 0.2905930313052842,
      "grad_norm": 2.546875,
      "learning_rate": 2.5641104530420737e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 1651018.0,
      "step": 3778
    },
    {
      "epoch": 0.29066994846550265,
      "grad_norm": 2.65625,
      "learning_rate": 2.563995077301746e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1651576.0,
      "step": 3779
    },
    {
      "epoch": 0.2907468656257211,
      "grad_norm": 3.984375,
      "learning_rate": 2.5638797015614186e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1652096.0,
      "step": 3780
    },
    {
      "epoch": 0.29082378278593957,
      "grad_norm": 2.9375,
      "learning_rate": 2.563764325821091e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1652682.0,
      "step": 3781
    },
    {
      "epoch": 0.29090069994615797,
      "grad_norm": 6.3125,
      "learning_rate": 2.563648950080763e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1653038.0,
      "step": 3782
    },
    {
      "epoch": 0.29097761710637643,
      "grad_norm": 5.75,
      "learning_rate": 2.5635335743404354e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1653320.0,
      "step": 3783
    },
    {
      "epoch": 0.2910545342665949,
      "grad_norm": 3.109375,
      "learning_rate": 2.5634181986001076e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 1654156.0,
      "step": 3784
    },
    {
      "epoch": 0.29113145142681335,
      "grad_norm": 2.484375,
      "learning_rate": 2.5633028228597803e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1654936.0,
      "step": 3785
    },
    {
      "epoch": 0.29120836858703175,
      "grad_norm": 2.921875,
      "learning_rate": 2.5631874471194526e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 1655522.0,
      "step": 3786
    },
    {
      "epoch": 0.2912852857472502,
      "grad_norm": 2.171875,
      "learning_rate": 2.5630720713791244e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1656080.0,
      "step": 3787
    },
    {
      "epoch": 0.29136220290746867,
      "grad_norm": 3.59375,
      "learning_rate": 2.562956695638797e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 1656430.0,
      "step": 3788
    },
    {
      "epoch": 0.2914391200676871,
      "grad_norm": 6.375,
      "learning_rate": 2.5628413198984694e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 1656726.0,
      "step": 3789
    },
    {
      "epoch": 0.29151603722790553,
      "grad_norm": 3.390625,
      "learning_rate": 2.5627259441581416e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1657270.0,
      "step": 3790
    },
    {
      "epoch": 0.291592954388124,
      "grad_norm": 8.3125,
      "learning_rate": 2.5626105684178143e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 1657594.0,
      "step": 3791
    },
    {
      "epoch": 0.29166987154834245,
      "grad_norm": 9.3125,
      "learning_rate": 2.5624951926774866e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 1657870.0,
      "step": 3792
    },
    {
      "epoch": 0.29174678870856086,
      "grad_norm": 6.25,
      "learning_rate": 2.562379816937159e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1658200.0,
      "step": 3793
    },
    {
      "epoch": 0.2918237058687793,
      "grad_norm": 3.0,
      "learning_rate": 2.562264441196831e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1658780.0,
      "step": 3794
    },
    {
      "epoch": 0.2919006230289978,
      "grad_norm": 8.3125,
      "learning_rate": 2.5621490654565033e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 1659088.0,
      "step": 3795
    },
    {
      "epoch": 0.29197754018921623,
      "grad_norm": 4.625,
      "learning_rate": 2.562033689716176e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1659396.0,
      "step": 3796
    },
    {
      "epoch": 0.29205445734943464,
      "grad_norm": 8.8125,
      "learning_rate": 2.5619183139758483e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 1659688.0,
      "step": 3797
    },
    {
      "epoch": 0.2921313745096531,
      "grad_norm": 5.9375,
      "learning_rate": 2.56180293823552e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1659994.0,
      "step": 3798
    },
    {
      "epoch": 0.29220829166987156,
      "grad_norm": 4.34375,
      "learning_rate": 2.5616875624951928e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1660318.0,
      "step": 3799
    },
    {
      "epoch": 0.29228520883009,
      "grad_norm": 3.640625,
      "learning_rate": 2.561572186754865e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1660894.0,
      "step": 3800
    },
    {
      "epoch": 0.2923621259903084,
      "grad_norm": 3.953125,
      "learning_rate": 2.5614568110145373e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1661188.0,
      "step": 3801
    },
    {
      "epoch": 0.2924390431505269,
      "grad_norm": 6.15625,
      "learning_rate": 2.56134143527421e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 1661476.0,
      "step": 3802
    },
    {
      "epoch": 0.29251596031074534,
      "grad_norm": 1.796875,
      "learning_rate": 2.5612260595338822e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1662022.0,
      "step": 3803
    },
    {
      "epoch": 0.2925928774709638,
      "grad_norm": 3.765625,
      "learning_rate": 2.5611106837935545e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1662518.0,
      "step": 3804
    },
    {
      "epoch": 0.2926697946311822,
      "grad_norm": 6.84375,
      "learning_rate": 2.5609953080532268e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 1662822.0,
      "step": 3805
    },
    {
      "epoch": 0.29274671179140066,
      "grad_norm": 7.46875,
      "learning_rate": 2.560879932312899e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1663162.0,
      "step": 3806
    },
    {
      "epoch": 0.2928236289516191,
      "grad_norm": 2.171875,
      "learning_rate": 2.5607645565725713e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1663470.0,
      "step": 3807
    },
    {
      "epoch": 0.2929005461118376,
      "grad_norm": 4.25,
      "learning_rate": 2.560649180832244e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1663764.0,
      "step": 3808
    },
    {
      "epoch": 0.292977463272056,
      "grad_norm": 3.734375,
      "learning_rate": 2.5605338050919158e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1664048.0,
      "step": 3809
    },
    {
      "epoch": 0.29305438043227444,
      "grad_norm": 6.21875,
      "learning_rate": 2.5604184293515885e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1664376.0,
      "step": 3810
    },
    {
      "epoch": 0.2931312975924929,
      "grad_norm": 2.75,
      "learning_rate": 2.5603030536112607e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 1664950.0,
      "step": 3811
    },
    {
      "epoch": 0.2932082147527113,
      "grad_norm": 3.015625,
      "learning_rate": 2.560187677870933e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1665530.0,
      "step": 3812
    },
    {
      "epoch": 0.29328513191292976,
      "grad_norm": 5.59375,
      "learning_rate": 2.5600723021306057e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1665880.0,
      "step": 3813
    },
    {
      "epoch": 0.2933620490731482,
      "grad_norm": 2.59375,
      "learning_rate": 2.559956926390278e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1666164.0,
      "step": 3814
    },
    {
      "epoch": 0.2934389662333667,
      "grad_norm": 2.734375,
      "learning_rate": 2.5598415506499498e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9980952143669128,
      "num_tokens": 1666690.0,
      "step": 3815
    },
    {
      "epoch": 0.2935158833935851,
      "grad_norm": 6.0625,
      "learning_rate": 2.5597261749096224e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9926199316978455,
      "num_tokens": 1666962.0,
      "step": 3816
    },
    {
      "epoch": 0.29359280055380355,
      "grad_norm": 2.59375,
      "learning_rate": 2.5596107991692947e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 1667760.0,
      "step": 3817
    },
    {
      "epoch": 0.293669717714022,
      "grad_norm": 7.40625,
      "learning_rate": 2.559495423428967e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9808306694030762,
      "num_tokens": 1668074.0,
      "step": 3818
    },
    {
      "epoch": 0.29374663487424046,
      "grad_norm": 2.875,
      "learning_rate": 2.5593800476886396e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1668644.0,
      "step": 3819
    },
    {
      "epoch": 0.29382355203445887,
      "grad_norm": 4.84375,
      "learning_rate": 2.5592646719483115e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1668966.0,
      "step": 3820
    },
    {
      "epoch": 0.2939004691946773,
      "grad_norm": 6.96875,
      "learning_rate": 2.559149296207984e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 1669222.0,
      "step": 3821
    },
    {
      "epoch": 0.2939773863548958,
      "grad_norm": 2.28125,
      "learning_rate": 2.5590339204676564e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1669764.0,
      "step": 3822
    },
    {
      "epoch": 0.29405430351511425,
      "grad_norm": 5.96875,
      "learning_rate": 2.5589185447273287e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1670092.0,
      "step": 3823
    },
    {
      "epoch": 0.29413122067533265,
      "grad_norm": 4.625,
      "learning_rate": 2.5588031689870014e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1670414.0,
      "step": 3824
    },
    {
      "epoch": 0.2942081378355511,
      "grad_norm": 7.40625,
      "learning_rate": 2.558687793246673e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 1670694.0,
      "step": 3825
    },
    {
      "epoch": 0.29428505499576957,
      "grad_norm": 3.734375,
      "learning_rate": 2.5585724175063454e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 1671262.0,
      "step": 3826
    },
    {
      "epoch": 0.294361972155988,
      "grad_norm": 8.8125,
      "learning_rate": 2.558457041766018e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.992337167263031,
      "num_tokens": 1671524.0,
      "step": 3827
    },
    {
      "epoch": 0.29443888931620643,
      "grad_norm": 2.296875,
      "learning_rate": 2.5583416660256904e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1672082.0,
      "step": 3828
    },
    {
      "epoch": 0.2945158064764249,
      "grad_norm": 3.890625,
      "learning_rate": 2.5582262902853626e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1672662.0,
      "step": 3829
    },
    {
      "epoch": 0.29459272363664335,
      "grad_norm": 3.5625,
      "learning_rate": 2.5581109145450353e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1673226.0,
      "step": 3830
    },
    {
      "epoch": 0.29466964079686175,
      "grad_norm": 7.0625,
      "learning_rate": 2.557995538804707e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 1673482.0,
      "step": 3831
    },
    {
      "epoch": 0.2947465579570802,
      "grad_norm": 3.28125,
      "learning_rate": 2.55788016306438e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 1674070.0,
      "step": 3832
    },
    {
      "epoch": 0.2948234751172987,
      "grad_norm": 2.171875,
      "learning_rate": 2.557764787324052e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1674616.0,
      "step": 3833
    },
    {
      "epoch": 0.29490039227751713,
      "grad_norm": 3.71875,
      "learning_rate": 2.5576494115837244e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1675190.0,
      "step": 3834
    },
    {
      "epoch": 0.29497730943773554,
      "grad_norm": 4.21875,
      "learning_rate": 2.557534035843397e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1675510.0,
      "step": 3835
    },
    {
      "epoch": 0.295054226597954,
      "grad_norm": 3.796875,
      "learning_rate": 2.557418660103069e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1675828.0,
      "step": 3836
    },
    {
      "epoch": 0.29513114375817245,
      "grad_norm": 2.890625,
      "learning_rate": 2.557303284362741e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 1676400.0,
      "step": 3837
    },
    {
      "epoch": 0.2952080609183909,
      "grad_norm": 4.625,
      "learning_rate": 2.557187908622414e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1676712.0,
      "step": 3838
    },
    {
      "epoch": 0.2952849780786093,
      "grad_norm": 1.953125,
      "learning_rate": 2.557072532882086e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1677258.0,
      "step": 3839
    },
    {
      "epoch": 0.2953618952388278,
      "grad_norm": 1.8671875,
      "learning_rate": 2.5569571571417583e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1677820.0,
      "step": 3840
    },
    {
      "epoch": 0.29543881239904624,
      "grad_norm": 8.5,
      "learning_rate": 2.556841781401431e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 1678192.0,
      "step": 3841
    },
    {
      "epoch": 0.2955157295592647,
      "grad_norm": 4.96875,
      "learning_rate": 2.556726405661103e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 1678530.0,
      "step": 3842
    },
    {
      "epoch": 0.2955926467194831,
      "grad_norm": 3.671875,
      "learning_rate": 2.5566110299207755e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1679072.0,
      "step": 3843
    },
    {
      "epoch": 0.29566956387970156,
      "grad_norm": 8.6875,
      "learning_rate": 2.5564956541804478e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 1679450.0,
      "step": 3844
    },
    {
      "epoch": 0.29574648103992,
      "grad_norm": 2.234375,
      "learning_rate": 2.55638027844012e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1679996.0,
      "step": 3845
    },
    {
      "epoch": 0.2958233982001385,
      "grad_norm": 3.90625,
      "learning_rate": 2.5562649026997927e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1680346.0,
      "step": 3846
    },
    {
      "epoch": 0.2959003153603569,
      "grad_norm": 2.96875,
      "learning_rate": 2.5561495269594646e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1680668.0,
      "step": 3847
    },
    {
      "epoch": 0.29597723252057534,
      "grad_norm": 3.046875,
      "learning_rate": 2.556034151219137e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1681218.0,
      "step": 3848
    },
    {
      "epoch": 0.2960541496807938,
      "grad_norm": 2.328125,
      "learning_rate": 2.5559187754788095e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1681998.0,
      "step": 3849
    },
    {
      "epoch": 0.2961310668410122,
      "grad_norm": 7.25,
      "learning_rate": 2.5558033997384818e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9941520690917969,
      "num_tokens": 1682342.0,
      "step": 3850
    },
    {
      "epoch": 0.29620798400123066,
      "grad_norm": 2.78125,
      "learning_rate": 2.555688023998154e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1682900.0,
      "step": 3851
    },
    {
      "epoch": 0.2962849011614491,
      "grad_norm": 3.1875,
      "learning_rate": 2.5555726482578267e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1683442.0,
      "step": 3852
    },
    {
      "epoch": 0.2963618183216676,
      "grad_norm": 2.90625,
      "learning_rate": 2.5554572725174985e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1684000.0,
      "step": 3853
    },
    {
      "epoch": 0.296438735481886,
      "grad_norm": 6.9375,
      "learning_rate": 2.555341896777171e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1684310.0,
      "step": 3854
    },
    {
      "epoch": 0.29651565264210444,
      "grad_norm": 1.9921875,
      "learning_rate": 2.5552265210368435e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1685096.0,
      "step": 3855
    },
    {
      "epoch": 0.2965925698023229,
      "grad_norm": 4.46875,
      "learning_rate": 2.5551111452965157e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 1685688.0,
      "step": 3856
    },
    {
      "epoch": 0.29666948696254136,
      "grad_norm": 7.53125,
      "learning_rate": 2.554995769556188e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1685970.0,
      "step": 3857
    },
    {
      "epoch": 0.29674640412275977,
      "grad_norm": 2.734375,
      "learning_rate": 2.5548803938158602e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1686512.0,
      "step": 3858
    },
    {
      "epoch": 0.2968233212829782,
      "grad_norm": 1.6484375,
      "learning_rate": 2.5547650180755325e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1687298.0,
      "step": 3859
    },
    {
      "epoch": 0.2969002384431967,
      "grad_norm": 7.125,
      "learning_rate": 2.554649642335205e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1687590.0,
      "step": 3860
    },
    {
      "epoch": 0.29697715560341514,
      "grad_norm": 7.9375,
      "learning_rate": 2.5545342665948774e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1687898.0,
      "step": 3861
    },
    {
      "epoch": 0.29705407276363355,
      "grad_norm": 7.0,
      "learning_rate": 2.5544188908545497e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1688214.0,
      "step": 3862
    },
    {
      "epoch": 0.297130989923852,
      "grad_norm": 7.90625,
      "learning_rate": 2.554303515114222e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1688518.0,
      "step": 3863
    },
    {
      "epoch": 0.29720790708407047,
      "grad_norm": 4.84375,
      "learning_rate": 2.5541881393738942e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1688838.0,
      "step": 3864
    },
    {
      "epoch": 0.2972848242442889,
      "grad_norm": 2.90625,
      "learning_rate": 2.5540727636335665e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1689374.0,
      "step": 3865
    },
    {
      "epoch": 0.29736174140450733,
      "grad_norm": 2.390625,
      "learning_rate": 2.553957387893239e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1689924.0,
      "step": 3866
    },
    {
      "epoch": 0.2974386585647258,
      "grad_norm": 6.875,
      "learning_rate": 2.5538420121529114e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 1690264.0,
      "step": 3867
    },
    {
      "epoch": 0.29751557572494425,
      "grad_norm": 6.09375,
      "learning_rate": 2.5537266364125837e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 1690596.0,
      "step": 3868
    },
    {
      "epoch": 0.2975924928851627,
      "grad_norm": 2.375,
      "learning_rate": 2.553611260672256e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1691416.0,
      "step": 3869
    },
    {
      "epoch": 0.2976694100453811,
      "grad_norm": 2.6875,
      "learning_rate": 2.553495884931928e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 1691956.0,
      "step": 3870
    },
    {
      "epoch": 0.29774632720559957,
      "grad_norm": 1.7890625,
      "learning_rate": 2.553380509191601e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1692498.0,
      "step": 3871
    },
    {
      "epoch": 0.29782324436581803,
      "grad_norm": 4.34375,
      "learning_rate": 2.553265133451273e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1692810.0,
      "step": 3872
    },
    {
      "epoch": 0.29790016152603643,
      "grad_norm": 4.5,
      "learning_rate": 2.5531497577109454e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 1693402.0,
      "step": 3873
    },
    {
      "epoch": 0.2979770786862549,
      "grad_norm": 4.4375,
      "learning_rate": 2.5530343819706176e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1693708.0,
      "step": 3874
    },
    {
      "epoch": 0.29805399584647335,
      "grad_norm": 2.59375,
      "learning_rate": 2.55291900623029e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1694282.0,
      "step": 3875
    },
    {
      "epoch": 0.2981309130066918,
      "grad_norm": 8.6875,
      "learning_rate": 2.552803630489962e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1694622.0,
      "step": 3876
    },
    {
      "epoch": 0.2982078301669102,
      "grad_norm": 6.875,
      "learning_rate": 2.552688254749635e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 1694932.0,
      "step": 3877
    },
    {
      "epoch": 0.2982847473271287,
      "grad_norm": 2.375,
      "learning_rate": 2.552572879009307e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1695478.0,
      "step": 3878
    },
    {
      "epoch": 0.29836166448734713,
      "grad_norm": 4.125,
      "learning_rate": 2.5524575032689794e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 1695796.0,
      "step": 3879
    },
    {
      "epoch": 0.2984385816475656,
      "grad_norm": 3.703125,
      "learning_rate": 2.5523421275286516e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1696130.0,
      "step": 3880
    },
    {
      "epoch": 0.298515498807784,
      "grad_norm": 2.875,
      "learning_rate": 2.552226751788324e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1696654.0,
      "step": 3881
    },
    {
      "epoch": 0.29859241596800246,
      "grad_norm": 8.6875,
      "learning_rate": 2.5521113760479966e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 1696910.0,
      "step": 3882
    },
    {
      "epoch": 0.2986693331282209,
      "grad_norm": 2.296875,
      "learning_rate": 2.551996000307669e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1697466.0,
      "step": 3883
    },
    {
      "epoch": 0.2987462502884394,
      "grad_norm": 3.640625,
      "learning_rate": 2.551880624567341e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1698046.0,
      "step": 3884
    },
    {
      "epoch": 0.2988231674486578,
      "grad_norm": 5.75,
      "learning_rate": 2.5517652488270133e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1698342.0,
      "step": 3885
    },
    {
      "epoch": 0.29890008460887624,
      "grad_norm": 6.25,
      "learning_rate": 2.5516498730866856e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 1698656.0,
      "step": 3886
    },
    {
      "epoch": 0.2989770017690947,
      "grad_norm": 3.890625,
      "learning_rate": 2.551534497346358e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1699208.0,
      "step": 3887
    },
    {
      "epoch": 0.29905391892931316,
      "grad_norm": 2.375,
      "learning_rate": 2.5514191216060305e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1699782.0,
      "step": 3888
    },
    {
      "epoch": 0.29913083608953156,
      "grad_norm": 3.4375,
      "learning_rate": 2.5513037458657028e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1700308.0,
      "step": 3889
    },
    {
      "epoch": 0.29920775324975,
      "grad_norm": 6.125,
      "learning_rate": 2.551188370125375e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 1700670.0,
      "step": 3890
    },
    {
      "epoch": 0.2992846704099685,
      "grad_norm": 4.15625,
      "learning_rate": 2.5510729943850473e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1700980.0,
      "step": 3891
    },
    {
      "epoch": 0.2993615875701869,
      "grad_norm": 2.1875,
      "learning_rate": 2.5509576186447196e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1701504.0,
      "step": 3892
    },
    {
      "epoch": 0.29943850473040534,
      "grad_norm": 5.84375,
      "learning_rate": 2.5508422429043922e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1701822.0,
      "step": 3893
    },
    {
      "epoch": 0.2995154218906238,
      "grad_norm": 4.3125,
      "learning_rate": 2.5507268671640645e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 1702162.0,
      "step": 3894
    },
    {
      "epoch": 0.29959233905084226,
      "grad_norm": 2.65625,
      "learning_rate": 2.5506114914237368e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1702920.0,
      "step": 3895
    },
    {
      "epoch": 0.29966925621106066,
      "grad_norm": 3.359375,
      "learning_rate": 2.550496115683409e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1703462.0,
      "step": 3896
    },
    {
      "epoch": 0.2997461733712791,
      "grad_norm": 7.1875,
      "learning_rate": 2.5503807399430813e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1703754.0,
      "step": 3897
    },
    {
      "epoch": 0.2998230905314976,
      "grad_norm": 7.8125,
      "learning_rate": 2.5502653642027535e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1704082.0,
      "step": 3898
    },
    {
      "epoch": 0.29990000769171604,
      "grad_norm": 3.015625,
      "learning_rate": 2.5501499884624262e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 1704650.0,
      "step": 3899
    },
    {
      "epoch": 0.29997692485193445,
      "grad_norm": 4.21875,
      "learning_rate": 2.5500346127220985e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1704960.0,
      "step": 3900
    },
    {
      "epoch": 0.3000538420121529,
      "grad_norm": 2.46875,
      "learning_rate": 2.5499192369817707e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1705522.0,
      "step": 3901
    },
    {
      "epoch": 0.30013075917237136,
      "grad_norm": 3.765625,
      "learning_rate": 2.549803861241443e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1705846.0,
      "step": 3902
    },
    {
      "epoch": 0.3002076763325898,
      "grad_norm": 2.078125,
      "learning_rate": 2.5496884855011152e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1706420.0,
      "step": 3903
    },
    {
      "epoch": 0.3002845934928082,
      "grad_norm": 5.0625,
      "learning_rate": 2.549573109760788e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1706714.0,
      "step": 3904
    },
    {
      "epoch": 0.3003615106530267,
      "grad_norm": 5.875,
      "learning_rate": 2.54945773402046e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1706992.0,
      "step": 3905
    },
    {
      "epoch": 0.30043842781324515,
      "grad_norm": 3.65625,
      "learning_rate": 2.5493423582801324e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1707564.0,
      "step": 3906
    },
    {
      "epoch": 0.3005153449734636,
      "grad_norm": 5.375,
      "learning_rate": 2.5492269825398047e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 1707892.0,
      "step": 3907
    },
    {
      "epoch": 0.300592262133682,
      "grad_norm": 5.40625,
      "learning_rate": 2.549111606799477e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1708174.0,
      "step": 3908
    },
    {
      "epoch": 0.30066917929390047,
      "grad_norm": 5.375,
      "learning_rate": 2.5489962310591492e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1708486.0,
      "step": 3909
    },
    {
      "epoch": 0.3007460964541189,
      "grad_norm": 7.6875,
      "learning_rate": 2.548880855318822e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 1708772.0,
      "step": 3910
    },
    {
      "epoch": 0.30082301361433733,
      "grad_norm": 4.09375,
      "learning_rate": 2.548765479578494e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1709082.0,
      "step": 3911
    },
    {
      "epoch": 0.3008999307745558,
      "grad_norm": 2.1875,
      "learning_rate": 2.5486501038381664e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1709624.0,
      "step": 3912
    },
    {
      "epoch": 0.30097684793477425,
      "grad_norm": 5.28125,
      "learning_rate": 2.5485347280978387e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1709930.0,
      "step": 3913
    },
    {
      "epoch": 0.3010537650949927,
      "grad_norm": 4.90625,
      "learning_rate": 2.548419352357511e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1710216.0,
      "step": 3914
    },
    {
      "epoch": 0.3011306822552111,
      "grad_norm": 7.1875,
      "learning_rate": 2.548303976617183e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1710548.0,
      "step": 3915
    },
    {
      "epoch": 0.30120759941542957,
      "grad_norm": 5.125,
      "learning_rate": 2.548188600876856e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 1710888.0,
      "step": 3916
    },
    {
      "epoch": 0.30128451657564803,
      "grad_norm": 8.3125,
      "learning_rate": 2.548073225136528e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 1711190.0,
      "step": 3917
    },
    {
      "epoch": 0.3013614337358665,
      "grad_norm": 2.828125,
      "learning_rate": 2.5479578493962004e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 1711750.0,
      "step": 3918
    },
    {
      "epoch": 0.3014383508960849,
      "grad_norm": 3.515625,
      "learning_rate": 2.5478424736558726e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1712056.0,
      "step": 3919
    },
    {
      "epoch": 0.30151526805630335,
      "grad_norm": 2.75,
      "learning_rate": 2.547727097915545e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 1712844.0,
      "step": 3920
    },
    {
      "epoch": 0.3015921852165218,
      "grad_norm": 2.140625,
      "learning_rate": 2.5476117221752176e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1713388.0,
      "step": 3921
    },
    {
      "epoch": 0.3016691023767403,
      "grad_norm": 5.5625,
      "learning_rate": 2.54749634643489e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1713722.0,
      "step": 3922
    },
    {
      "epoch": 0.3017460195369587,
      "grad_norm": 5.9375,
      "learning_rate": 2.5473809706945617e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1714024.0,
      "step": 3923
    },
    {
      "epoch": 0.30182293669717714,
      "grad_norm": 7.3125,
      "learning_rate": 2.5472655949542344e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1714352.0,
      "step": 3924
    },
    {
      "epoch": 0.3018998538573956,
      "grad_norm": 5.625,
      "learning_rate": 2.5471502192139066e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 1714660.0,
      "step": 3925
    },
    {
      "epoch": 0.30197677101761405,
      "grad_norm": 3.828125,
      "learning_rate": 2.547034843473579e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1715212.0,
      "step": 3926
    },
    {
      "epoch": 0.30205368817783246,
      "grad_norm": 4.1875,
      "learning_rate": 2.5469194677332516e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1715738.0,
      "step": 3927
    },
    {
      "epoch": 0.3021306053380509,
      "grad_norm": 3.6875,
      "learning_rate": 2.5468040919929234e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1716082.0,
      "step": 3928
    },
    {
      "epoch": 0.3022075224982694,
      "grad_norm": 2.65625,
      "learning_rate": 2.546688716252596e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1716902.0,
      "step": 3929
    },
    {
      "epoch": 0.30228443965848784,
      "grad_norm": 2.484375,
      "learning_rate": 2.5465733405122683e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1717448.0,
      "step": 3930
    },
    {
      "epoch": 0.30236135681870624,
      "grad_norm": 3.328125,
      "learning_rate": 2.5464579647719406e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1717964.0,
      "step": 3931
    },
    {
      "epoch": 0.3024382739789247,
      "grad_norm": 5.34375,
      "learning_rate": 2.5463425890316133e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1718308.0,
      "step": 3932
    },
    {
      "epoch": 0.30251519113914316,
      "grad_norm": 2.34375,
      "learning_rate": 2.5462272132912855e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1718856.0,
      "step": 3933
    },
    {
      "epoch": 0.30259210829936156,
      "grad_norm": 2.3125,
      "learning_rate": 2.5461118375509574e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1719402.0,
      "step": 3934
    },
    {
      "epoch": 0.30266902545958,
      "grad_norm": 4.5625,
      "learning_rate": 2.54599646181063e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1719708.0,
      "step": 3935
    },
    {
      "epoch": 0.3027459426197985,
      "grad_norm": 5.65625,
      "learning_rate": 2.5458810860703023e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1720000.0,
      "step": 3936
    },
    {
      "epoch": 0.30282285978001694,
      "grad_norm": 2.9375,
      "learning_rate": 2.5457657103299746e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 1720522.0,
      "step": 3937
    },
    {
      "epoch": 0.30289977694023534,
      "grad_norm": 2.296875,
      "learning_rate": 2.5456503345896472e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1721072.0,
      "step": 3938
    },
    {
      "epoch": 0.3029766941004538,
      "grad_norm": 4.34375,
      "learning_rate": 2.545534958849319e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 1721390.0,
      "step": 3939
    },
    {
      "epoch": 0.30305361126067226,
      "grad_norm": 7.40625,
      "learning_rate": 2.5454195831089918e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 1721694.0,
      "step": 3940
    },
    {
      "epoch": 0.3031305284208907,
      "grad_norm": 2.8125,
      "learning_rate": 2.545304207368664e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1722258.0,
      "step": 3941
    },
    {
      "epoch": 0.3032074455811091,
      "grad_norm": 10.0,
      "learning_rate": 2.5451888316283363e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 1722532.0,
      "step": 3942
    },
    {
      "epoch": 0.3032843627413276,
      "grad_norm": 6.34375,
      "learning_rate": 2.545073455888009e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1722832.0,
      "step": 3943
    },
    {
      "epoch": 0.30336127990154604,
      "grad_norm": 7.46875,
      "learning_rate": 2.5449580801476812e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 1723188.0,
      "step": 3944
    },
    {
      "epoch": 0.3034381970617645,
      "grad_norm": 6.78125,
      "learning_rate": 2.544842704407353e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 1723460.0,
      "step": 3945
    },
    {
      "epoch": 0.3035151142219829,
      "grad_norm": 3.296875,
      "learning_rate": 2.5447273286670257e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 1724048.0,
      "step": 3946
    },
    {
      "epoch": 0.30359203138220137,
      "grad_norm": 6.625,
      "learning_rate": 2.544611952926698e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1724386.0,
      "step": 3947
    },
    {
      "epoch": 0.3036689485424198,
      "grad_norm": 5.75,
      "learning_rate": 2.5444965771863702e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1724702.0,
      "step": 3948
    },
    {
      "epoch": 0.3037458657026383,
      "grad_norm": 3.953125,
      "learning_rate": 2.544381201446043e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1725278.0,
      "step": 3949
    },
    {
      "epoch": 0.3038227828628567,
      "grad_norm": 4.78125,
      "learning_rate": 2.5442658257057148e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 1725612.0,
      "step": 3950
    },
    {
      "epoch": 0.30389970002307515,
      "grad_norm": 6.5625,
      "learning_rate": 2.5441504499653874e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1725918.0,
      "step": 3951
    },
    {
      "epoch": 0.3039766171832936,
      "grad_norm": 3.953125,
      "learning_rate": 2.5440350742250597e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 1726246.0,
      "step": 3952
    },
    {
      "epoch": 0.304053534343512,
      "grad_norm": 9.125,
      "learning_rate": 2.543919698484732e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 1726522.0,
      "step": 3953
    },
    {
      "epoch": 0.30413045150373047,
      "grad_norm": 2.703125,
      "learning_rate": 2.5438043227444046e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1727090.0,
      "step": 3954
    },
    {
      "epoch": 0.30420736866394893,
      "grad_norm": 2.46875,
      "learning_rate": 2.543688947004077e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 1727632.0,
      "step": 3955
    },
    {
      "epoch": 0.3042842858241674,
      "grad_norm": 13.3125,
      "learning_rate": 2.5435735712637487e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9750000238418579,
      "num_tokens": 1727956.0,
      "step": 3956
    },
    {
      "epoch": 0.3043612029843858,
      "grad_norm": 9.3125,
      "learning_rate": 2.5434581955234214e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9858155846595764,
      "num_tokens": 1728240.0,
      "step": 3957
    },
    {
      "epoch": 0.30443812014460425,
      "grad_norm": 2.765625,
      "learning_rate": 2.5433428197830937e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1728790.0,
      "step": 3958
    },
    {
      "epoch": 0.3045150373048227,
      "grad_norm": 5.625,
      "learning_rate": 2.543227444042766e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1729118.0,
      "step": 3959
    },
    {
      "epoch": 0.30459195446504117,
      "grad_norm": 2.515625,
      "learning_rate": 2.5431120683024386e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1729678.0,
      "step": 3960
    },
    {
      "epoch": 0.3046688716252596,
      "grad_norm": 2.515625,
      "learning_rate": 2.5429966925621105e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1730212.0,
      "step": 3961
    },
    {
      "epoch": 0.30474578878547803,
      "grad_norm": 4.40625,
      "learning_rate": 2.542881316821783e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1730534.0,
      "step": 3962
    },
    {
      "epoch": 0.3048227059456965,
      "grad_norm": 2.5,
      "learning_rate": 2.5427659410814554e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1731104.0,
      "step": 3963
    },
    {
      "epoch": 0.30489962310591495,
      "grad_norm": 4.8125,
      "learning_rate": 2.5426505653411276e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1731386.0,
      "step": 3964
    },
    {
      "epoch": 0.30497654026613336,
      "grad_norm": 6.03125,
      "learning_rate": 2.5425351896008e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9943342804908752,
      "num_tokens": 1731740.0,
      "step": 3965
    },
    {
      "epoch": 0.3050534574263518,
      "grad_norm": 7.0,
      "learning_rate": 2.542419813860472e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 1732048.0,
      "step": 3966
    },
    {
      "epoch": 0.3051303745865703,
      "grad_norm": 4.125,
      "learning_rate": 2.5423044381201444e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1732370.0,
      "step": 3967
    },
    {
      "epoch": 0.30520729174678873,
      "grad_norm": 1.9609375,
      "learning_rate": 2.542189062379817e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 1733178.0,
      "step": 3968
    },
    {
      "epoch": 0.30528420890700714,
      "grad_norm": 5.65625,
      "learning_rate": 2.5420736866394894e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 1733522.0,
      "step": 3969
    },
    {
      "epoch": 0.3053611260672256,
      "grad_norm": 8.0625,
      "learning_rate": 2.5419583108991616e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1733784.0,
      "step": 3970
    },
    {
      "epoch": 0.30543804322744406,
      "grad_norm": 2.640625,
      "learning_rate": 2.5418429351588343e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1734326.0,
      "step": 3971
    },
    {
      "epoch": 0.30551496038766246,
      "grad_norm": 8.625,
      "learning_rate": 2.541727559418506e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1734626.0,
      "step": 3972
    },
    {
      "epoch": 0.3055918775478809,
      "grad_norm": 8.125,
      "learning_rate": 2.5416121836781784e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1734928.0,
      "step": 3973
    },
    {
      "epoch": 0.3056687947080994,
      "grad_norm": 7.0625,
      "learning_rate": 2.541496807937851e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1735212.0,
      "step": 3974
    },
    {
      "epoch": 0.30574571186831784,
      "grad_norm": 3.6875,
      "learning_rate": 2.5413814321975233e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1735764.0,
      "step": 3975
    },
    {
      "epoch": 0.30582262902853624,
      "grad_norm": 8.8125,
      "learning_rate": 2.5412660564571956e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 1736116.0,
      "step": 3976
    },
    {
      "epoch": 0.3058995461887547,
      "grad_norm": 2.46875,
      "learning_rate": 2.541150680716868e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 1736904.0,
      "step": 3977
    },
    {
      "epoch": 0.30597646334897316,
      "grad_norm": 3.375,
      "learning_rate": 2.54103530497654e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1737224.0,
      "step": 3978
    },
    {
      "epoch": 0.3060533805091916,
      "grad_norm": 3.1875,
      "learning_rate": 2.540919929236213e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1737794.0,
      "step": 3979
    },
    {
      "epoch": 0.30613029766941,
      "grad_norm": 3.03125,
      "learning_rate": 2.540804553495885e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1738326.0,
      "step": 3980
    },
    {
      "epoch": 0.3062072148296285,
      "grad_norm": 3.421875,
      "learning_rate": 2.5406891777555573e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1738906.0,
      "step": 3981
    },
    {
      "epoch": 0.30628413198984694,
      "grad_norm": 4.75,
      "learning_rate": 2.54057380201523e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1739216.0,
      "step": 3982
    },
    {
      "epoch": 0.3063610491500654,
      "grad_norm": 2.390625,
      "learning_rate": 2.540458426274902e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1739764.0,
      "step": 3983
    },
    {
      "epoch": 0.3064379663102838,
      "grad_norm": 9.125,
      "learning_rate": 2.540343050534574e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1740044.0,
      "step": 3984
    },
    {
      "epoch": 0.30651488347050226,
      "grad_norm": 3.765625,
      "learning_rate": 2.5402276747942468e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1740360.0,
      "step": 3985
    },
    {
      "epoch": 0.3065918006307207,
      "grad_norm": 5.15625,
      "learning_rate": 2.540112299053919e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 1740690.0,
      "step": 3986
    },
    {
      "epoch": 0.3066687177909392,
      "grad_norm": 3.65625,
      "learning_rate": 2.5399969233135913e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1741012.0,
      "step": 3987
    },
    {
      "epoch": 0.3067456349511576,
      "grad_norm": 5.0,
      "learning_rate": 2.5398815475732635e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1741336.0,
      "step": 3988
    },
    {
      "epoch": 0.30682255211137605,
      "grad_norm": 6.28125,
      "learning_rate": 2.539766171832936e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1741622.0,
      "step": 3989
    },
    {
      "epoch": 0.3068994692715945,
      "grad_norm": 1.9609375,
      "learning_rate": 2.5396507960926085e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 1742404.0,
      "step": 3990
    },
    {
      "epoch": 0.30697638643181296,
      "grad_norm": 6.4375,
      "learning_rate": 2.5395354203522807e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1742674.0,
      "step": 3991
    },
    {
      "epoch": 0.30705330359203137,
      "grad_norm": 3.203125,
      "learning_rate": 2.539420044611953e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1743220.0,
      "step": 3992
    },
    {
      "epoch": 0.3071302207522498,
      "grad_norm": 2.25,
      "learning_rate": 2.5393046688716257e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 1743790.0,
      "step": 3993
    },
    {
      "epoch": 0.3072071379124683,
      "grad_norm": 3.203125,
      "learning_rate": 2.5391892931312975e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 1744364.0,
      "step": 3994
    },
    {
      "epoch": 0.3072840550726867,
      "grad_norm": 6.875,
      "learning_rate": 2.5390739173909698e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 1744650.0,
      "step": 3995
    },
    {
      "epoch": 0.30736097223290515,
      "grad_norm": 3.34375,
      "learning_rate": 2.5389585416506424e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1744972.0,
      "step": 3996
    },
    {
      "epoch": 0.3074378893931236,
      "grad_norm": 6.46875,
      "learning_rate": 2.5388431659103147e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 1745244.0,
      "step": 3997
    },
    {
      "epoch": 0.30751480655334207,
      "grad_norm": 6.375,
      "learning_rate": 2.538727790169987e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1745572.0,
      "step": 3998
    },
    {
      "epoch": 0.30759172371356047,
      "grad_norm": 4.53125,
      "learning_rate": 2.5386124144296592e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1745866.0,
      "step": 3999
    },
    {
      "epoch": 0.30766864087377893,
      "grad_norm": 2.515625,
      "learning_rate": 2.5384970386893315e-06,
      "loss": 0.0104,
      "step": 4000
    },
    {
      "epoch": 0.30766864087377893,
      "eval_loss": 0.020453371107578278,
      "eval_mean_token_accuracy": 0.9915181255340576,
      "eval_num_tokens": 1746686.0,
      "eval_runtime": 21.0653,
      "eval_samples_per_second": 9.494,
      "eval_steps_per_second": 1.187,
      "step": 4000
    },
    {
      "epoch": 0.3077455580339974,
      "grad_norm": 6.53125,
      "learning_rate": 2.538381662949004e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9893041849136353,
      "num_tokens": 1746990.0,
      "step": 4001
    },
    {
      "epoch": 0.30782247519421585,
      "grad_norm": 8.375,
      "learning_rate": 2.5382662872086764e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9963369965553284,
      "num_tokens": 1747264.0,
      "step": 4002
    },
    {
      "epoch": 0.30789939235443425,
      "grad_norm": 7.96875,
      "learning_rate": 2.5381509114683487e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1747542.0,
      "step": 4003
    },
    {
      "epoch": 0.3079763095146527,
      "grad_norm": 5.4375,
      "learning_rate": 2.538035535728021e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1747892.0,
      "step": 4004
    },
    {
      "epoch": 0.3080532266748712,
      "grad_norm": 5.5,
      "learning_rate": 2.537920159987693e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 1748226.0,
      "step": 4005
    },
    {
      "epoch": 0.30813014383508963,
      "grad_norm": 2.296875,
      "learning_rate": 2.5378047842473655e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1748772.0,
      "step": 4006
    },
    {
      "epoch": 0.30820706099530804,
      "grad_norm": 2.515625,
      "learning_rate": 2.537689408507038e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1749318.0,
      "step": 4007
    },
    {
      "epoch": 0.3082839781555265,
      "grad_norm": 7.5,
      "learning_rate": 2.5375740327667104e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1749596.0,
      "step": 4008
    },
    {
      "epoch": 0.30836089531574495,
      "grad_norm": 3.625,
      "learning_rate": 2.5374586570263827e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 1750112.0,
      "step": 4009
    },
    {
      "epoch": 0.3084378124759634,
      "grad_norm": 2.1875,
      "learning_rate": 2.537343281286055e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9970559477806091,
      "num_tokens": 1751134.0,
      "step": 4010
    },
    {
      "epoch": 0.3085147296361818,
      "grad_norm": 4.875,
      "learning_rate": 2.537227905545727e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9897785186767578,
      "num_tokens": 1751722.0,
      "step": 4011
    },
    {
      "epoch": 0.3085916467964003,
      "grad_norm": 7.75,
      "learning_rate": 2.5371125298054e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 1752002.0,
      "step": 4012
    },
    {
      "epoch": 0.30866856395661874,
      "grad_norm": 3.21875,
      "learning_rate": 2.536997154065072e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1752542.0,
      "step": 4013
    },
    {
      "epoch": 0.30874548111683714,
      "grad_norm": 5.5,
      "learning_rate": 2.5368817783247444e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1752872.0,
      "step": 4014
    },
    {
      "epoch": 0.3088223982770556,
      "grad_norm": 3.265625,
      "learning_rate": 2.5367664025844166e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 1753684.0,
      "step": 4015
    },
    {
      "epoch": 0.30889931543727406,
      "grad_norm": 5.4375,
      "learning_rate": 2.536651026844089e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1753990.0,
      "step": 4016
    },
    {
      "epoch": 0.3089762325974925,
      "grad_norm": 2.640625,
      "learning_rate": 2.536535651103761e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1754810.0,
      "step": 4017
    },
    {
      "epoch": 0.3090531497577109,
      "grad_norm": 2.078125,
      "learning_rate": 2.536420275363434e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 1755352.0,
      "step": 4018
    },
    {
      "epoch": 0.3091300669179294,
      "grad_norm": 3.109375,
      "learning_rate": 2.536304899623106e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1756178.0,
      "step": 4019
    },
    {
      "epoch": 0.30920698407814784,
      "grad_norm": 7.34375,
      "learning_rate": 2.5361895238827783e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 1756480.0,
      "step": 4020
    },
    {
      "epoch": 0.3092839012383663,
      "grad_norm": 1.953125,
      "learning_rate": 2.5360741481424506e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 1757288.0,
      "step": 4021
    },
    {
      "epoch": 0.3093608183985847,
      "grad_norm": 6.78125,
      "learning_rate": 2.535958772402123e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 1757620.0,
      "step": 4022
    },
    {
      "epoch": 0.30943773555880316,
      "grad_norm": 1.8046875,
      "learning_rate": 2.535843396661795e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 1758160.0,
      "step": 4023
    },
    {
      "epoch": 0.3095146527190216,
      "grad_norm": 6.59375,
      "learning_rate": 2.535728020921468e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1758496.0,
      "step": 4024
    },
    {
      "epoch": 0.3095915698792401,
      "grad_norm": 3.1875,
      "learning_rate": 2.53561264518114e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 1759012.0,
      "step": 4025
    },
    {
      "epoch": 0.3096684870394585,
      "grad_norm": 7.5,
      "learning_rate": 2.5354972694408123e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1759326.0,
      "step": 4026
    },
    {
      "epoch": 0.30974540419967694,
      "grad_norm": 1.9609375,
      "learning_rate": 2.5353818937004846e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9973545074462891,
      "num_tokens": 1760084.0,
      "step": 4027
    },
    {
      "epoch": 0.3098223213598954,
      "grad_norm": 5.21875,
      "learning_rate": 2.535266517960157e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9895651936531067,
      "num_tokens": 1760660.0,
      "step": 4028
    },
    {
      "epoch": 0.30989923852011386,
      "grad_norm": 7.375,
      "learning_rate": 2.5351511422198295e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9768977165222168,
      "num_tokens": 1760964.0,
      "step": 4029
    },
    {
      "epoch": 0.30997615568033227,
      "grad_norm": 9.25,
      "learning_rate": 2.5350357664795018e-06,
      "loss": 0.0383,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 1761256.0,
      "step": 4030
    },
    {
      "epoch": 0.3100530728405507,
      "grad_norm": 10.0625,
      "learning_rate": 2.5349203907391736e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 1761508.0,
      "step": 4031
    },
    {
      "epoch": 0.3101299900007692,
      "grad_norm": 6.0,
      "learning_rate": 2.5348050149988463e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1761792.0,
      "step": 4032
    },
    {
      "epoch": 0.31020690716098764,
      "grad_norm": 1.953125,
      "learning_rate": 2.5346896392585185e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1762336.0,
      "step": 4033
    },
    {
      "epoch": 0.31028382432120605,
      "grad_norm": 2.171875,
      "learning_rate": 2.534574263518191e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1762882.0,
      "step": 4034
    },
    {
      "epoch": 0.3103607414814245,
      "grad_norm": 3.46875,
      "learning_rate": 2.5344588877778635e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 1763460.0,
      "step": 4035
    },
    {
      "epoch": 0.31043765864164297,
      "grad_norm": 2.921875,
      "learning_rate": 2.5343435120375357e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 1764018.0,
      "step": 4036
    },
    {
      "epoch": 0.31051457580186137,
      "grad_norm": 7.71875,
      "learning_rate": 2.534228136297208e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9850187301635742,
      "num_tokens": 1764286.0,
      "step": 4037
    },
    {
      "epoch": 0.31059149296207983,
      "grad_norm": 5.71875,
      "learning_rate": 2.5341127605568803e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9967213273048401,
      "num_tokens": 1764592.0,
      "step": 4038
    },
    {
      "epoch": 0.3106684101222983,
      "grad_norm": 2.765625,
      "learning_rate": 2.5339973848165525e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1765110.0,
      "step": 4039
    },
    {
      "epoch": 0.31074532728251675,
      "grad_norm": 3.984375,
      "learning_rate": 2.533882009076225e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1765430.0,
      "step": 4040
    },
    {
      "epoch": 0.31082224444273515,
      "grad_norm": 2.734375,
      "learning_rate": 2.5337666333358975e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1765976.0,
      "step": 4041
    },
    {
      "epoch": 0.3108991616029536,
      "grad_norm": 7.71875,
      "learning_rate": 2.5336512575955693e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1766286.0,
      "step": 4042
    },
    {
      "epoch": 0.31097607876317207,
      "grad_norm": 2.03125,
      "learning_rate": 2.533535881855242e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1767066.0,
      "step": 4043
    },
    {
      "epoch": 0.31105299592339053,
      "grad_norm": 4.15625,
      "learning_rate": 2.5334205061149142e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 1767400.0,
      "step": 4044
    },
    {
      "epoch": 0.31112991308360893,
      "grad_norm": 2.03125,
      "learning_rate": 2.5333051303745865e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 1768208.0,
      "step": 4045
    },
    {
      "epoch": 0.3112068302438274,
      "grad_norm": 5.53125,
      "learning_rate": 2.533189754634259e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 1768532.0,
      "step": 4046
    },
    {
      "epoch": 0.31128374740404585,
      "grad_norm": 6.03125,
      "learning_rate": 2.5330743788939314e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 1768872.0,
      "step": 4047
    },
    {
      "epoch": 0.3113606645642643,
      "grad_norm": 3.09375,
      "learning_rate": 2.5329590031536037e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1769430.0,
      "step": 4048
    },
    {
      "epoch": 0.3114375817244827,
      "grad_norm": 5.375,
      "learning_rate": 2.532843627413276e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1769754.0,
      "step": 4049
    },
    {
      "epoch": 0.3115144988847012,
      "grad_norm": 5.9375,
      "learning_rate": 2.532728251672948e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9943342804908752,
      "num_tokens": 1770108.0,
      "step": 4050
    },
    {
      "epoch": 0.31159141604491963,
      "grad_norm": 4.71875,
      "learning_rate": 2.532612875932621e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1770408.0,
      "step": 4051
    },
    {
      "epoch": 0.3116683332051381,
      "grad_norm": 3.421875,
      "learning_rate": 2.532497500192293e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 1770988.0,
      "step": 4052
    },
    {
      "epoch": 0.3117452503653565,
      "grad_norm": 7.03125,
      "learning_rate": 2.532382124451965e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 1771278.0,
      "step": 4053
    },
    {
      "epoch": 0.31182216752557496,
      "grad_norm": 5.90625,
      "learning_rate": 2.5322667487116377e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1771594.0,
      "step": 4054
    },
    {
      "epoch": 0.3118990846857934,
      "grad_norm": 6.8125,
      "learning_rate": 2.53215137297131e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1771900.0,
      "step": 4055
    },
    {
      "epoch": 0.3119760018460118,
      "grad_norm": 2.21875,
      "learning_rate": 2.532035997230982e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1772486.0,
      "step": 4056
    },
    {
      "epoch": 0.3120529190062303,
      "grad_norm": 5.6875,
      "learning_rate": 2.531920621490655e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 1772832.0,
      "step": 4057
    },
    {
      "epoch": 0.31212983616644874,
      "grad_norm": 3.1875,
      "learning_rate": 2.531805245750327e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1773356.0,
      "step": 4058
    },
    {
      "epoch": 0.3122067533266672,
      "grad_norm": 4.96875,
      "learning_rate": 2.5316898700099994e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1773690.0,
      "step": 4059
    },
    {
      "epoch": 0.3122836704868856,
      "grad_norm": 6.0,
      "learning_rate": 2.5315744942696716e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 1774254.0,
      "step": 4060
    },
    {
      "epoch": 0.31236058764710406,
      "grad_norm": 4.34375,
      "learning_rate": 2.531459118529344e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 1774594.0,
      "step": 4061
    },
    {
      "epoch": 0.3124375048073225,
      "grad_norm": 2.359375,
      "learning_rate": 2.5313437427890166e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1775178.0,
      "step": 4062
    },
    {
      "epoch": 0.312514421967541,
      "grad_norm": 3.265625,
      "learning_rate": 2.531228367048689e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9909256100654602,
      "num_tokens": 1775730.0,
      "step": 4063
    },
    {
      "epoch": 0.3125913391277594,
      "grad_norm": 6.5,
      "learning_rate": 2.5311129913083607e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 1776024.0,
      "step": 4064
    },
    {
      "epoch": 0.31266825628797784,
      "grad_norm": 1.9296875,
      "learning_rate": 2.5309976155680333e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 1776610.0,
      "step": 4065
    },
    {
      "epoch": 0.3127451734481963,
      "grad_norm": 5.5,
      "learning_rate": 2.5308822398277056e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1776944.0,
      "step": 4066
    },
    {
      "epoch": 0.31282209060841476,
      "grad_norm": 4.625,
      "learning_rate": 2.530766864087378e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 1777300.0,
      "step": 4067
    },
    {
      "epoch": 0.31289900776863316,
      "grad_norm": 6.96875,
      "learning_rate": 2.5306514883470505e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 1777612.0,
      "step": 4068
    },
    {
      "epoch": 0.3129759249288516,
      "grad_norm": 10.625,
      "learning_rate": 2.5305361126067224e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 1777938.0,
      "step": 4069
    },
    {
      "epoch": 0.3130528420890701,
      "grad_norm": 7.75,
      "learning_rate": 2.530420736866395e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9855907559394836,
      "num_tokens": 1778286.0,
      "step": 4070
    },
    {
      "epoch": 0.31312975924928854,
      "grad_norm": 8.6875,
      "learning_rate": 2.5303053611260673e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1778566.0,
      "step": 4071
    },
    {
      "epoch": 0.31320667640950695,
      "grad_norm": 2.203125,
      "learning_rate": 2.5301899853857396e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 1779376.0,
      "step": 4072
    },
    {
      "epoch": 0.3132835935697254,
      "grad_norm": 6.09375,
      "learning_rate": 2.530074609645412e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9971671104431152,
      "num_tokens": 1779730.0,
      "step": 4073
    },
    {
      "epoch": 0.31336051072994386,
      "grad_norm": 3.96875,
      "learning_rate": 2.5299592339050845e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 1780052.0,
      "step": 4074
    },
    {
      "epoch": 0.31343742789016227,
      "grad_norm": 6.15625,
      "learning_rate": 2.5298438581647563e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1780392.0,
      "step": 4075
    },
    {
      "epoch": 0.3135143450503807,
      "grad_norm": 6.28125,
      "learning_rate": 2.529728482424429e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1780688.0,
      "step": 4076
    },
    {
      "epoch": 0.3135912622105992,
      "grad_norm": 3.171875,
      "learning_rate": 2.5296131066841013e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1781230.0,
      "step": 4077
    },
    {
      "epoch": 0.31366817937081765,
      "grad_norm": 6.0625,
      "learning_rate": 2.5294977309437735e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1781540.0,
      "step": 4078
    },
    {
      "epoch": 0.31374509653103605,
      "grad_norm": 2.140625,
      "learning_rate": 2.5293823552034462e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1782124.0,
      "step": 4079
    },
    {
      "epoch": 0.3138220136912545,
      "grad_norm": 8.25,
      "learning_rate": 2.529266979463118e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1782466.0,
      "step": 4080
    },
    {
      "epoch": 0.31389893085147297,
      "grad_norm": 6.78125,
      "learning_rate": 2.5291516037227903e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1782750.0,
      "step": 4081
    },
    {
      "epoch": 0.3139758480116914,
      "grad_norm": 2.140625,
      "learning_rate": 2.529036227982463e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1783308.0,
      "step": 4082
    },
    {
      "epoch": 0.31405276517190983,
      "grad_norm": 3.890625,
      "learning_rate": 2.5289208522421353e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1783828.0,
      "step": 4083
    },
    {
      "epoch": 0.3141296823321283,
      "grad_norm": 10.5625,
      "learning_rate": 2.5288054765018075e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1784106.0,
      "step": 4084
    },
    {
      "epoch": 0.31420659949234675,
      "grad_norm": 7.6875,
      "learning_rate": 2.52869010076148e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1784446.0,
      "step": 4085
    },
    {
      "epoch": 0.3142835166525652,
      "grad_norm": 3.46875,
      "learning_rate": 2.528574725021152e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1784972.0,
      "step": 4086
    },
    {
      "epoch": 0.3143604338127836,
      "grad_norm": 6.96875,
      "learning_rate": 2.5284593492808247e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1785304.0,
      "step": 4087
    },
    {
      "epoch": 0.31443735097300207,
      "grad_norm": 8.25,
      "learning_rate": 2.528343973540497e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 1785590.0,
      "step": 4088
    },
    {
      "epoch": 0.31451426813322053,
      "grad_norm": 7.34375,
      "learning_rate": 2.5282285978001692e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1785850.0,
      "step": 4089
    },
    {
      "epoch": 0.314591185293439,
      "grad_norm": 3.390625,
      "learning_rate": 2.528113222059842e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1786162.0,
      "step": 4090
    },
    {
      "epoch": 0.3146681024536574,
      "grad_norm": 3.75,
      "learning_rate": 2.5279978463195137e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1786656.0,
      "step": 4091
    },
    {
      "epoch": 0.31474501961387585,
      "grad_norm": 2.359375,
      "learning_rate": 2.527882470579186e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1787214.0,
      "step": 4092
    },
    {
      "epoch": 0.3148219367740943,
      "grad_norm": 3.234375,
      "learning_rate": 2.5277670948388587e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1787784.0,
      "step": 4093
    },
    {
      "epoch": 0.3148988539343128,
      "grad_norm": 3.578125,
      "learning_rate": 2.527651719098531e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1788090.0,
      "step": 4094
    },
    {
      "epoch": 0.3149757710945312,
      "grad_norm": 7.0625,
      "learning_rate": 2.527536343358203e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1788370.0,
      "step": 4095
    },
    {
      "epoch": 0.31505268825474964,
      "grad_norm": 1.7109375,
      "learning_rate": 2.527420967617876e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1788916.0,
      "step": 4096
    },
    {
      "epoch": 0.3151296054149681,
      "grad_norm": 2.421875,
      "learning_rate": 2.5273055918775477e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 1789484.0,
      "step": 4097
    },
    {
      "epoch": 0.3152065225751865,
      "grad_norm": 7.46875,
      "learning_rate": 2.5271902161372204e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 1789794.0,
      "step": 4098
    },
    {
      "epoch": 0.31528343973540496,
      "grad_norm": 3.453125,
      "learning_rate": 2.5270748403968927e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1790378.0,
      "step": 4099
    },
    {
      "epoch": 0.3153603568956234,
      "grad_norm": 3.15625,
      "learning_rate": 2.526959464656565e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1790924.0,
      "step": 4100
    },
    {
      "epoch": 0.3154372740558419,
      "grad_norm": 4.09375,
      "learning_rate": 2.5268440889162376e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 1791510.0,
      "step": 4101
    },
    {
      "epoch": 0.3155141912160603,
      "grad_norm": 2.375,
      "learning_rate": 2.5267287131759094e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1792290.0,
      "step": 4102
    },
    {
      "epoch": 0.31559110837627874,
      "grad_norm": 2.046875,
      "learning_rate": 2.5266133374355817e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 1793076.0,
      "step": 4103
    },
    {
      "epoch": 0.3156680255364972,
      "grad_norm": 5.40625,
      "learning_rate": 2.5264979616952544e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1793386.0,
      "step": 4104
    },
    {
      "epoch": 0.31574494269671566,
      "grad_norm": 7.15625,
      "learning_rate": 2.5263825859549266e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 1793666.0,
      "step": 4105
    },
    {
      "epoch": 0.31582185985693406,
      "grad_norm": 3.953125,
      "learning_rate": 2.526267210214599e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1793988.0,
      "step": 4106
    },
    {
      "epoch": 0.3158987770171525,
      "grad_norm": 2.421875,
      "learning_rate": 2.526151834474271e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 1794528.0,
      "step": 4107
    },
    {
      "epoch": 0.315975694177371,
      "grad_norm": 2.875,
      "learning_rate": 2.5260364587339434e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1795076.0,
      "step": 4108
    },
    {
      "epoch": 0.31605261133758944,
      "grad_norm": 6.75,
      "learning_rate": 2.525921082993616e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1795394.0,
      "step": 4109
    },
    {
      "epoch": 0.31612952849780784,
      "grad_norm": 3.84375,
      "learning_rate": 2.5258057072532883e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1795966.0,
      "step": 4110
    },
    {
      "epoch": 0.3162064456580263,
      "grad_norm": 1.875,
      "learning_rate": 2.5256903315129606e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1796512.0,
      "step": 4111
    },
    {
      "epoch": 0.31628336281824476,
      "grad_norm": 8.4375,
      "learning_rate": 2.5255749557726333e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1796824.0,
      "step": 4112
    },
    {
      "epoch": 0.3163602799784632,
      "grad_norm": 3.421875,
      "learning_rate": 2.525459580032305e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 1797366.0,
      "step": 4113
    },
    {
      "epoch": 0.3164371971386816,
      "grad_norm": 1.7421875,
      "learning_rate": 2.5253442042919774e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 1798146.0,
      "step": 4114
    },
    {
      "epoch": 0.3165141142989001,
      "grad_norm": 2.671875,
      "learning_rate": 2.52522882855165e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1798972.0,
      "step": 4115
    },
    {
      "epoch": 0.31659103145911854,
      "grad_norm": 9.8125,
      "learning_rate": 2.5251134528113223e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1799292.0,
      "step": 4116
    },
    {
      "epoch": 0.31666794861933695,
      "grad_norm": 6.875,
      "learning_rate": 2.5249980770709946e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1799574.0,
      "step": 4117
    },
    {
      "epoch": 0.3167448657795554,
      "grad_norm": 7.03125,
      "learning_rate": 2.524882701330667e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1799922.0,
      "step": 4118
    },
    {
      "epoch": 0.31682178293977387,
      "grad_norm": 5.84375,
      "learning_rate": 2.524767325590339e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1800272.0,
      "step": 4119
    },
    {
      "epoch": 0.3168987000999923,
      "grad_norm": 6.15625,
      "learning_rate": 2.5246519498500118e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 1800612.0,
      "step": 4120
    },
    {
      "epoch": 0.31697561726021073,
      "grad_norm": 2.546875,
      "learning_rate": 2.524536574109684e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1801170.0,
      "step": 4121
    },
    {
      "epoch": 0.3170525344204292,
      "grad_norm": 1.859375,
      "learning_rate": 2.5244211983693563e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 1801980.0,
      "step": 4122
    },
    {
      "epoch": 0.31712945158064765,
      "grad_norm": 3.390625,
      "learning_rate": 2.524305822629029e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1802532.0,
      "step": 4123
    },
    {
      "epoch": 0.3172063687408661,
      "grad_norm": 2.59375,
      "learning_rate": 2.524190446888701e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 1803078.0,
      "step": 4124
    },
    {
      "epoch": 0.3172832859010845,
      "grad_norm": 3.28125,
      "learning_rate": 2.524075071148373e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 1803674.0,
      "step": 4125
    },
    {
      "epoch": 0.31736020306130297,
      "grad_norm": 4.4375,
      "learning_rate": 2.5239596954080457e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1804194.0,
      "step": 4126
    },
    {
      "epoch": 0.31743712022152143,
      "grad_norm": 2.9375,
      "learning_rate": 2.523844319667718e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1804778.0,
      "step": 4127
    },
    {
      "epoch": 0.3175140373817399,
      "grad_norm": 9.0625,
      "learning_rate": 2.5237289439273903e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1805040.0,
      "step": 4128
    },
    {
      "epoch": 0.3175909545419583,
      "grad_norm": 6.03125,
      "learning_rate": 2.5236135681870625e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 1805396.0,
      "step": 4129
    },
    {
      "epoch": 0.31766787170217675,
      "grad_norm": 3.765625,
      "learning_rate": 2.5234981924467348e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 1805914.0,
      "step": 4130
    },
    {
      "epoch": 0.3177447888623952,
      "grad_norm": 3.0625,
      "learning_rate": 2.5233828167064075e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 1806458.0,
      "step": 4131
    },
    {
      "epoch": 0.31782170602261367,
      "grad_norm": 5.59375,
      "learning_rate": 2.5232674409660797e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 1806806.0,
      "step": 4132
    },
    {
      "epoch": 0.3178986231828321,
      "grad_norm": 2.078125,
      "learning_rate": 2.523152065225752e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 1807620.0,
      "step": 4133
    },
    {
      "epoch": 0.31797554034305053,
      "grad_norm": 5.96875,
      "learning_rate": 2.5230366894854242e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 1807976.0,
      "step": 4134
    },
    {
      "epoch": 0.318052457503269,
      "grad_norm": 5.6875,
      "learning_rate": 2.5229213137450965e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1808282.0,
      "step": 4135
    },
    {
      "epoch": 0.3181293746634874,
      "grad_norm": 7.1875,
      "learning_rate": 2.5228059380047687e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1808552.0,
      "step": 4136
    },
    {
      "epoch": 0.31820629182370586,
      "grad_norm": 4.0,
      "learning_rate": 2.5226905622644414e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 1809130.0,
      "step": 4137
    },
    {
      "epoch": 0.3182832089839243,
      "grad_norm": 6.46875,
      "learning_rate": 2.5225751865241137e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1809470.0,
      "step": 4138
    },
    {
      "epoch": 0.3183601261441428,
      "grad_norm": 3.171875,
      "learning_rate": 2.522459810783786e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1810046.0,
      "step": 4139
    },
    {
      "epoch": 0.3184370433043612,
      "grad_norm": 4.3125,
      "learning_rate": 2.522344435043458e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1810368.0,
      "step": 4140
    },
    {
      "epoch": 0.31851396046457964,
      "grad_norm": 3.78125,
      "learning_rate": 2.5222290593031305e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 1810694.0,
      "step": 4141
    },
    {
      "epoch": 0.3185908776247981,
      "grad_norm": 3.46875,
      "learning_rate": 2.5221136835628027e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1811022.0,
      "step": 4142
    },
    {
      "epoch": 0.31866779478501656,
      "grad_norm": 7.90625,
      "learning_rate": 2.5219983078224754e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1811328.0,
      "step": 4143
    },
    {
      "epoch": 0.31874471194523496,
      "grad_norm": 3.5,
      "learning_rate": 2.5218829320821477e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1811874.0,
      "step": 4144
    },
    {
      "epoch": 0.3188216291054534,
      "grad_norm": 3.609375,
      "learning_rate": 2.52176755634182e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1812174.0,
      "step": 4145
    },
    {
      "epoch": 0.3188985462656719,
      "grad_norm": 5.59375,
      "learning_rate": 2.521652180601492e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1812742.0,
      "step": 4146
    },
    {
      "epoch": 0.31897546342589034,
      "grad_norm": 7.03125,
      "learning_rate": 2.5215368048611644e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1813026.0,
      "step": 4147
    },
    {
      "epoch": 0.31905238058610874,
      "grad_norm": 3.421875,
      "learning_rate": 2.521421429120837e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1813574.0,
      "step": 4148
    },
    {
      "epoch": 0.3191292977463272,
      "grad_norm": 3.90625,
      "learning_rate": 2.5213060533805094e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1814116.0,
      "step": 4149
    },
    {
      "epoch": 0.31920621490654566,
      "grad_norm": 2.46875,
      "learning_rate": 2.5211906776401816e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 1814650.0,
      "step": 4150
    },
    {
      "epoch": 0.3192831320667641,
      "grad_norm": 7.5,
      "learning_rate": 2.521075301899854e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1815012.0,
      "step": 4151
    },
    {
      "epoch": 0.3193600492269825,
      "grad_norm": 2.84375,
      "learning_rate": 2.520959926159526e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 1815582.0,
      "step": 4152
    },
    {
      "epoch": 0.319436966387201,
      "grad_norm": 5.84375,
      "learning_rate": 2.5208445504191984e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1815890.0,
      "step": 4153
    },
    {
      "epoch": 0.31951388354741944,
      "grad_norm": 12.375,
      "learning_rate": 2.520729174678871e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1816226.0,
      "step": 4154
    },
    {
      "epoch": 0.3195908007076379,
      "grad_norm": 2.90625,
      "learning_rate": 2.5206137989385433e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1816790.0,
      "step": 4155
    },
    {
      "epoch": 0.3196677178678563,
      "grad_norm": 9.5625,
      "learning_rate": 2.5204984231982156e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9918699264526367,
      "num_tokens": 1817162.0,
      "step": 4156
    },
    {
      "epoch": 0.31974463502807476,
      "grad_norm": 1.96875,
      "learning_rate": 2.520383047457888e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1817710.0,
      "step": 4157
    },
    {
      "epoch": 0.3198215521882932,
      "grad_norm": 3.25,
      "learning_rate": 2.52026767171756e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1818010.0,
      "step": 4158
    },
    {
      "epoch": 0.3198984693485116,
      "grad_norm": 7.125,
      "learning_rate": 2.520152295977233e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 1818320.0,
      "step": 4159
    },
    {
      "epoch": 0.3199753865087301,
      "grad_norm": 4.1875,
      "learning_rate": 2.520036920236905e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1818866.0,
      "step": 4160
    },
    {
      "epoch": 0.32005230366894855,
      "grad_norm": 9.125,
      "learning_rate": 2.5199215444965773e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1819172.0,
      "step": 4161
    },
    {
      "epoch": 0.320129220829167,
      "grad_norm": 3.90625,
      "learning_rate": 2.5198061687562496e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1819494.0,
      "step": 4162
    },
    {
      "epoch": 0.3202061379893854,
      "grad_norm": 4.71875,
      "learning_rate": 2.519690793015922e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9967948794364929,
      "num_tokens": 1819808.0,
      "step": 4163
    },
    {
      "epoch": 0.32028305514960387,
      "grad_norm": 6.03125,
      "learning_rate": 2.519575417275594e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 1820142.0,
      "step": 4164
    },
    {
      "epoch": 0.3203599723098223,
      "grad_norm": 8.375,
      "learning_rate": 2.5194600415352668e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1820442.0,
      "step": 4165
    },
    {
      "epoch": 0.3204368894700408,
      "grad_norm": 10.0625,
      "learning_rate": 2.519344665794939e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1820722.0,
      "step": 4166
    },
    {
      "epoch": 0.3205138066302592,
      "grad_norm": 5.71875,
      "learning_rate": 2.5192292900546113e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1821020.0,
      "step": 4167
    },
    {
      "epoch": 0.32059072379047765,
      "grad_norm": 8.0,
      "learning_rate": 2.5191139143142835e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 1821298.0,
      "step": 4168
    },
    {
      "epoch": 0.3206676409506961,
      "grad_norm": 5.1875,
      "learning_rate": 2.518998538573956e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 1821618.0,
      "step": 4169
    },
    {
      "epoch": 0.32074455811091457,
      "grad_norm": 4.96875,
      "learning_rate": 2.5188831628336285e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 1822138.0,
      "step": 4170
    },
    {
      "epoch": 0.32082147527113297,
      "grad_norm": 8.9375,
      "learning_rate": 2.5187677870933007e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 1822476.0,
      "step": 4171
    },
    {
      "epoch": 0.32089839243135143,
      "grad_norm": 7.96875,
      "learning_rate": 2.5186524113529726e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 1822786.0,
      "step": 4172
    },
    {
      "epoch": 0.3209753095915699,
      "grad_norm": 6.84375,
      "learning_rate": 2.5185370356126453e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 1823102.0,
      "step": 4173
    },
    {
      "epoch": 0.32105222675178835,
      "grad_norm": 8.0625,
      "learning_rate": 2.5184216598723175e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 1823462.0,
      "step": 4174
    },
    {
      "epoch": 0.32112914391200675,
      "grad_norm": 3.390625,
      "learning_rate": 2.5183062841319898e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 1824018.0,
      "step": 4175
    },
    {
      "epoch": 0.3212060610722252,
      "grad_norm": 5.46875,
      "learning_rate": 2.5181909083916625e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1824592.0,
      "step": 4176
    },
    {
      "epoch": 0.3212829782324437,
      "grad_norm": 2.4375,
      "learning_rate": 2.5180755326513347e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 1825166.0,
      "step": 4177
    },
    {
      "epoch": 0.3213598953926621,
      "grad_norm": 5.21875,
      "learning_rate": 2.517960156911007e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1825730.0,
      "step": 4178
    },
    {
      "epoch": 0.32143681255288054,
      "grad_norm": 3.140625,
      "learning_rate": 2.5178447811706792e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1826300.0,
      "step": 4179
    },
    {
      "epoch": 0.321513729713099,
      "grad_norm": 9.875,
      "learning_rate": 2.5177294054303515e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 1826668.0,
      "step": 4180
    },
    {
      "epoch": 0.32159064687331745,
      "grad_norm": 9.25,
      "learning_rate": 2.517614029690024e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1826974.0,
      "step": 4181
    },
    {
      "epoch": 0.32166756403353586,
      "grad_norm": 3.296875,
      "learning_rate": 2.5174986539496964e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1827522.0,
      "step": 4182
    },
    {
      "epoch": 0.3217444811937543,
      "grad_norm": 3.875,
      "learning_rate": 2.5173832782093683e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 1828098.0,
      "step": 4183
    },
    {
      "epoch": 0.3218213983539728,
      "grad_norm": 6.21875,
      "learning_rate": 2.517267902469041e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1828410.0,
      "step": 4184
    },
    {
      "epoch": 0.32189831551419124,
      "grad_norm": 7.78125,
      "learning_rate": 2.517152526728713e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1828752.0,
      "step": 4185
    },
    {
      "epoch": 0.32197523267440964,
      "grad_norm": 6.625,
      "learning_rate": 2.5170371509883855e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1829036.0,
      "step": 4186
    },
    {
      "epoch": 0.3220521498346281,
      "grad_norm": 6.84375,
      "learning_rate": 2.516921775248058e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 1829346.0,
      "step": 4187
    },
    {
      "epoch": 0.32212906699484656,
      "grad_norm": 5.1875,
      "learning_rate": 2.5168063995077304e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 1829660.0,
      "step": 4188
    },
    {
      "epoch": 0.322205984155065,
      "grad_norm": 6.25,
      "learning_rate": 2.5166910237674027e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 1829996.0,
      "step": 4189
    },
    {
      "epoch": 0.3222829013152834,
      "grad_norm": 2.671875,
      "learning_rate": 2.516575648027075e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1830552.0,
      "step": 4190
    },
    {
      "epoch": 0.3223598184755019,
      "grad_norm": 6.875,
      "learning_rate": 2.516460272286747e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1830892.0,
      "step": 4191
    },
    {
      "epoch": 0.32243673563572034,
      "grad_norm": 7.90625,
      "learning_rate": 2.5163448965464194e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1831248.0,
      "step": 4192
    },
    {
      "epoch": 0.3225136527959388,
      "grad_norm": 5.3125,
      "learning_rate": 2.516229520806092e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1831744.0,
      "step": 4193
    },
    {
      "epoch": 0.3225905699561572,
      "grad_norm": 3.03125,
      "learning_rate": 2.516114145065764e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1832302.0,
      "step": 4194
    },
    {
      "epoch": 0.32266748711637566,
      "grad_norm": 2.390625,
      "learning_rate": 2.5159987693254366e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1832850.0,
      "step": 4195
    },
    {
      "epoch": 0.3227444042765941,
      "grad_norm": 2.21875,
      "learning_rate": 2.515883393585109e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1833396.0,
      "step": 4196
    },
    {
      "epoch": 0.3228213214368126,
      "grad_norm": 4.53125,
      "learning_rate": 2.515768017844781e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1833890.0,
      "step": 4197
    },
    {
      "epoch": 0.322898238597031,
      "grad_norm": 6.75,
      "learning_rate": 2.515652642104454e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1834196.0,
      "step": 4198
    },
    {
      "epoch": 0.32297515575724944,
      "grad_norm": 6.5,
      "learning_rate": 2.515537266364126e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9967319965362549,
      "num_tokens": 1834504.0,
      "step": 4199
    },
    {
      "epoch": 0.3230520729174679,
      "grad_norm": 4.15625,
      "learning_rate": 2.515421890623798e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1835052.0,
      "step": 4200
    },
    {
      "epoch": 0.3231289900776863,
      "grad_norm": 6.4375,
      "learning_rate": 2.5153065148834706e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 1835400.0,
      "step": 4201
    },
    {
      "epoch": 0.32320590723790477,
      "grad_norm": 3.25,
      "learning_rate": 2.515191139143143e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 1835980.0,
      "step": 4202
    },
    {
      "epoch": 0.3232828243981232,
      "grad_norm": 9.3125,
      "learning_rate": 2.515075763402815e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1836294.0,
      "step": 4203
    },
    {
      "epoch": 0.3233597415583417,
      "grad_norm": 1.8515625,
      "learning_rate": 2.514960387662488e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 1836812.0,
      "step": 4204
    },
    {
      "epoch": 0.3234366587185601,
      "grad_norm": 3.984375,
      "learning_rate": 2.5148450119221596e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1837402.0,
      "step": 4205
    },
    {
      "epoch": 0.32351357587877855,
      "grad_norm": 3.828125,
      "learning_rate": 2.5147296361818323e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 1837722.0,
      "step": 4206
    },
    {
      "epoch": 0.323590493038997,
      "grad_norm": 8.25,
      "learning_rate": 2.5146142604415046e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 1838056.0,
      "step": 4207
    },
    {
      "epoch": 0.32366741019921547,
      "grad_norm": 6.03125,
      "learning_rate": 2.514498884701177e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1838384.0,
      "step": 4208
    },
    {
      "epoch": 0.32374432735943387,
      "grad_norm": 6.125,
      "learning_rate": 2.5143835089608495e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 1838724.0,
      "step": 4209
    },
    {
      "epoch": 0.32382124451965233,
      "grad_norm": 3.71875,
      "learning_rate": 2.5142681332205214e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1839046.0,
      "step": 4210
    },
    {
      "epoch": 0.3238981616798708,
      "grad_norm": 3.703125,
      "learning_rate": 2.5141527574801936e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1839566.0,
      "step": 4211
    },
    {
      "epoch": 0.32397507884008925,
      "grad_norm": 2.921875,
      "learning_rate": 2.5140373817398663e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1840124.0,
      "step": 4212
    },
    {
      "epoch": 0.32405199600030765,
      "grad_norm": 1.8671875,
      "learning_rate": 2.5139220059995386e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 1840932.0,
      "step": 4213
    },
    {
      "epoch": 0.3241289131605261,
      "grad_norm": 4.15625,
      "learning_rate": 2.513806630259211e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 1841528.0,
      "step": 4214
    },
    {
      "epoch": 0.32420583032074457,
      "grad_norm": 2.984375,
      "learning_rate": 2.5136912545188835e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 1842050.0,
      "step": 4215
    },
    {
      "epoch": 0.32428274748096303,
      "grad_norm": 2.78125,
      "learning_rate": 2.5135758787785553e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1842596.0,
      "step": 4216
    },
    {
      "epoch": 0.32435966464118143,
      "grad_norm": 2.640625,
      "learning_rate": 2.513460503038228e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 1843186.0,
      "step": 4217
    },
    {
      "epoch": 0.3244365818013999,
      "grad_norm": 1.9296875,
      "learning_rate": 2.5133451272979003e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 1843968.0,
      "step": 4218
    },
    {
      "epoch": 0.32451349896161835,
      "grad_norm": 6.9375,
      "learning_rate": 2.5132297515575725e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1844302.0,
      "step": 4219
    },
    {
      "epoch": 0.32459041612183676,
      "grad_norm": 3.03125,
      "learning_rate": 2.513114375817245e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1844848.0,
      "step": 4220
    },
    {
      "epoch": 0.3246673332820552,
      "grad_norm": 8.8125,
      "learning_rate": 2.512999000076917e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1845190.0,
      "step": 4221
    },
    {
      "epoch": 0.3247442504422737,
      "grad_norm": 7.0,
      "learning_rate": 2.5128836243365893e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1845460.0,
      "step": 4222
    },
    {
      "epoch": 0.32482116760249213,
      "grad_norm": 3.421875,
      "learning_rate": 2.512768248596262e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9942857027053833,
      "num_tokens": 1845986.0,
      "step": 4223
    },
    {
      "epoch": 0.32489808476271054,
      "grad_norm": 5.40625,
      "learning_rate": 2.5126528728559342e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1846322.0,
      "step": 4224
    },
    {
      "epoch": 0.324975001922929,
      "grad_norm": 2.53125,
      "learning_rate": 2.5125374971156065e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1846908.0,
      "step": 4225
    },
    {
      "epoch": 0.32505191908314746,
      "grad_norm": 4.6875,
      "learning_rate": 2.512422121375279e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1847228.0,
      "step": 4226
    },
    {
      "epoch": 0.3251288362433659,
      "grad_norm": 5.8125,
      "learning_rate": 2.512306745634951e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1847538.0,
      "step": 4227
    },
    {
      "epoch": 0.3252057534035843,
      "grad_norm": 3.03125,
      "learning_rate": 2.5121913698946237e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 1848108.0,
      "step": 4228
    },
    {
      "epoch": 0.3252826705638028,
      "grad_norm": 8.4375,
      "learning_rate": 2.512075994154296e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9838709831237793,
      "num_tokens": 1848420.0,
      "step": 4229
    },
    {
      "epoch": 0.32535958772402124,
      "grad_norm": 3.890625,
      "learning_rate": 2.511960618413968e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 1848936.0,
      "step": 4230
    },
    {
      "epoch": 0.3254365048842397,
      "grad_norm": 8.1875,
      "learning_rate": 2.511845242673641e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9838709831237793,
      "num_tokens": 1849248.0,
      "step": 4231
    },
    {
      "epoch": 0.3255134220444581,
      "grad_norm": 2.359375,
      "learning_rate": 2.5117298669333127e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1849772.0,
      "step": 4232
    },
    {
      "epoch": 0.32559033920467656,
      "grad_norm": 12.8125,
      "learning_rate": 2.511614491192985e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 1850076.0,
      "step": 4233
    },
    {
      "epoch": 0.325667256364895,
      "grad_norm": 8.625,
      "learning_rate": 2.5114991154526577e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 1850354.0,
      "step": 4234
    },
    {
      "epoch": 0.3257441735251135,
      "grad_norm": 9.8125,
      "learning_rate": 2.51138373971233e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9848024249076843,
      "num_tokens": 1850684.0,
      "step": 4235
    },
    {
      "epoch": 0.3258210906853319,
      "grad_norm": 5.375,
      "learning_rate": 2.511268363972002e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 1851014.0,
      "step": 4236
    },
    {
      "epoch": 0.32589800784555034,
      "grad_norm": 3.75,
      "learning_rate": 2.5111529882316744e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1851320.0,
      "step": 4237
    },
    {
      "epoch": 0.3259749250057688,
      "grad_norm": 4.4375,
      "learning_rate": 2.5110376124913467e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1851856.0,
      "step": 4238
    },
    {
      "epoch": 0.3260518421659872,
      "grad_norm": 6.40625,
      "learning_rate": 2.5109222367510194e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 1852138.0,
      "step": 4239
    },
    {
      "epoch": 0.32612875932620566,
      "grad_norm": 6.46875,
      "learning_rate": 2.5108068610106916e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 1852442.0,
      "step": 4240
    },
    {
      "epoch": 0.3262056764864241,
      "grad_norm": 6.96875,
      "learning_rate": 2.510691485270364e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 1852782.0,
      "step": 4241
    },
    {
      "epoch": 0.3262825936466426,
      "grad_norm": 2.296875,
      "learning_rate": 2.510576109530036e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1853352.0,
      "step": 4242
    },
    {
      "epoch": 0.326359510806861,
      "grad_norm": 2.78125,
      "learning_rate": 2.5104607337897084e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 1854140.0,
      "step": 4243
    },
    {
      "epoch": 0.32643642796707945,
      "grad_norm": 7.875,
      "learning_rate": 2.5103453580493807e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1854470.0,
      "step": 4244
    },
    {
      "epoch": 0.3265133451272979,
      "grad_norm": 5.875,
      "learning_rate": 2.5102299823090534e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9890310764312744,
      "num_tokens": 1855018.0,
      "step": 4245
    },
    {
      "epoch": 0.32659026228751636,
      "grad_norm": 5.4375,
      "learning_rate": 2.5101146065687256e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 1855326.0,
      "step": 4246
    },
    {
      "epoch": 0.32666717944773477,
      "grad_norm": 9.875,
      "learning_rate": 2.509999230828398e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 1855584.0,
      "step": 4247
    },
    {
      "epoch": 0.3267440966079532,
      "grad_norm": 7.125,
      "learning_rate": 2.50988385508807e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1855886.0,
      "step": 4248
    },
    {
      "epoch": 0.3268210137681717,
      "grad_norm": 4.5625,
      "learning_rate": 2.5097684793477424e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 1856200.0,
      "step": 4249
    },
    {
      "epoch": 0.32689793092839015,
      "grad_norm": 2.890625,
      "learning_rate": 2.5096531036074146e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1856740.0,
      "step": 4250
    },
    {
      "epoch": 0.32697484808860855,
      "grad_norm": 3.96875,
      "learning_rate": 2.5095377278670873e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1857330.0,
      "step": 4251
    },
    {
      "epoch": 0.327051765248827,
      "grad_norm": 5.9375,
      "learning_rate": 2.5094223521267596e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 1857640.0,
      "step": 4252
    },
    {
      "epoch": 0.32712868240904547,
      "grad_norm": 2.3125,
      "learning_rate": 2.509306976386432e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 1858188.0,
      "step": 4253
    },
    {
      "epoch": 0.3272055995692639,
      "grad_norm": 4.03125,
      "learning_rate": 2.509191600646104e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 1858500.0,
      "step": 4254
    },
    {
      "epoch": 0.32728251672948233,
      "grad_norm": 5.21875,
      "learning_rate": 2.5090762249057764e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9890310764312744,
      "num_tokens": 1859048.0,
      "step": 4255
    },
    {
      "epoch": 0.3273594338897008,
      "grad_norm": 3.109375,
      "learning_rate": 2.508960849165449e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1859604.0,
      "step": 4256
    },
    {
      "epoch": 0.32743635104991925,
      "grad_norm": 3.375,
      "learning_rate": 2.5088454734251213e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1860168.0,
      "step": 4257
    },
    {
      "epoch": 0.3275132682101377,
      "grad_norm": 5.96875,
      "learning_rate": 2.5087300976847936e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 1860472.0,
      "step": 4258
    },
    {
      "epoch": 0.3275901853703561,
      "grad_norm": 1.9140625,
      "learning_rate": 2.508614721944466e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1861018.0,
      "step": 4259
    },
    {
      "epoch": 0.32766710253057457,
      "grad_norm": 6.78125,
      "learning_rate": 2.508499346204138e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 1861350.0,
      "step": 4260
    },
    {
      "epoch": 0.32774401969079303,
      "grad_norm": 4.09375,
      "learning_rate": 2.5083839704638103e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1861634.0,
      "step": 4261
    },
    {
      "epoch": 0.32782093685101144,
      "grad_norm": 6.375,
      "learning_rate": 2.508268594723483e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 1.0,
      "num_tokens": 1861970.0,
      "step": 4262
    },
    {
      "epoch": 0.3278978540112299,
      "grad_norm": 2.8125,
      "learning_rate": 2.5081532189831553e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1862488.0,
      "step": 4263
    },
    {
      "epoch": 0.32797477117144835,
      "grad_norm": 6.28125,
      "learning_rate": 2.5080378432428275e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 1862804.0,
      "step": 4264
    },
    {
      "epoch": 0.3280516883316668,
      "grad_norm": 3.359375,
      "learning_rate": 2.5079224675024998e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1863324.0,
      "step": 4265
    },
    {
      "epoch": 0.3281286054918852,
      "grad_norm": 6.6875,
      "learning_rate": 2.507807091762172e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 1863594.0,
      "step": 4266
    },
    {
      "epoch": 0.3282055226521037,
      "grad_norm": 5.21875,
      "learning_rate": 2.5076917160218447e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1863918.0,
      "step": 4267
    },
    {
      "epoch": 0.32828243981232214,
      "grad_norm": 4.5625,
      "learning_rate": 2.507576340281517e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1864490.0,
      "step": 4268
    },
    {
      "epoch": 0.3283593569725406,
      "grad_norm": 2.328125,
      "learning_rate": 2.5074609645411892e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1865038.0,
      "step": 4269
    },
    {
      "epoch": 0.328436274132759,
      "grad_norm": 3.921875,
      "learning_rate": 2.5073455888008615e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 1865614.0,
      "step": 4270
    },
    {
      "epoch": 0.32851319129297746,
      "grad_norm": 4.4375,
      "learning_rate": 2.5072302130605338e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 1866202.0,
      "step": 4271
    },
    {
      "epoch": 0.3285901084531959,
      "grad_norm": 1.9296875,
      "learning_rate": 2.507114837320206e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 1866982.0,
      "step": 4272
    },
    {
      "epoch": 0.3286670256134144,
      "grad_norm": 3.71875,
      "learning_rate": 2.5069994615798787e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1867502.0,
      "step": 4273
    },
    {
      "epoch": 0.3287439427736328,
      "grad_norm": 5.21875,
      "learning_rate": 2.506884085839551e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 1867786.0,
      "step": 4274
    },
    {
      "epoch": 0.32882085993385124,
      "grad_norm": 2.75,
      "learning_rate": 2.506768710099223e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1868372.0,
      "step": 4275
    },
    {
      "epoch": 0.3288977770940697,
      "grad_norm": 5.9375,
      "learning_rate": 2.5066533343588955e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 1868732.0,
      "step": 4276
    },
    {
      "epoch": 0.32897469425428816,
      "grad_norm": 5.90625,
      "learning_rate": 2.5065379586185677e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1869046.0,
      "step": 4277
    },
    {
      "epoch": 0.32905161141450656,
      "grad_norm": 5.9375,
      "learning_rate": 2.5064225828782404e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1869360.0,
      "step": 4278
    },
    {
      "epoch": 0.329128528574725,
      "grad_norm": 2.140625,
      "learning_rate": 2.5063072071379127e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 1869908.0,
      "step": 4279
    },
    {
      "epoch": 0.3292054457349435,
      "grad_norm": 5.375,
      "learning_rate": 2.506191831397585e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1870248.0,
      "step": 4280
    },
    {
      "epoch": 0.3292823628951619,
      "grad_norm": 5.53125,
      "learning_rate": 2.506076455657257e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 1870592.0,
      "step": 4281
    },
    {
      "epoch": 0.32935928005538034,
      "grad_norm": 4.125,
      "learning_rate": 2.5059610799169294e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9904030561447144,
      "num_tokens": 1871114.0,
      "step": 4282
    },
    {
      "epoch": 0.3294361972155988,
      "grad_norm": 3.546875,
      "learning_rate": 2.5058457041766017e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1871436.0,
      "step": 4283
    },
    {
      "epoch": 0.32951311437581726,
      "grad_norm": 6.4375,
      "learning_rate": 2.5057303284362744e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1871734.0,
      "step": 4284
    },
    {
      "epoch": 0.32959003153603567,
      "grad_norm": 3.71875,
      "learning_rate": 2.5056149526959466e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1872252.0,
      "step": 4285
    },
    {
      "epoch": 0.3296669486962541,
      "grad_norm": 4.0625,
      "learning_rate": 2.505499576955619e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 1872602.0,
      "step": 4286
    },
    {
      "epoch": 0.3297438658564726,
      "grad_norm": 4.84375,
      "learning_rate": 2.505384201215291e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 1872952.0,
      "step": 4287
    },
    {
      "epoch": 0.32982078301669104,
      "grad_norm": 6.0,
      "learning_rate": 2.5052688254749634e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 1873244.0,
      "step": 4288
    },
    {
      "epoch": 0.32989770017690945,
      "grad_norm": 2.09375,
      "learning_rate": 2.505153449734636e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 1873800.0,
      "step": 4289
    },
    {
      "epoch": 0.3299746173371279,
      "grad_norm": 2.921875,
      "learning_rate": 2.5050380739943084e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1874604.0,
      "step": 4290
    },
    {
      "epoch": 0.33005153449734637,
      "grad_norm": 7.625,
      "learning_rate": 2.5049226982539806e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1874886.0,
      "step": 4291
    },
    {
      "epoch": 0.3301284516575648,
      "grad_norm": 4.28125,
      "learning_rate": 2.504807322513653e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1875202.0,
      "step": 4292
    },
    {
      "epoch": 0.33020536881778323,
      "grad_norm": 7.5,
      "learning_rate": 2.504691946773325e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1875464.0,
      "step": 4293
    },
    {
      "epoch": 0.3302822859780017,
      "grad_norm": 4.96875,
      "learning_rate": 2.5045765710329974e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 1875776.0,
      "step": 4294
    },
    {
      "epoch": 0.33035920313822015,
      "grad_norm": 3.734375,
      "learning_rate": 2.50446119529267e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 1876270.0,
      "step": 4295
    },
    {
      "epoch": 0.3304361202984386,
      "grad_norm": 3.078125,
      "learning_rate": 2.5043458195523423e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 1876790.0,
      "step": 4296
    },
    {
      "epoch": 0.330513037458657,
      "grad_norm": 2.859375,
      "learning_rate": 2.5042304438120146e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1877610.0,
      "step": 4297
    },
    {
      "epoch": 0.33058995461887547,
      "grad_norm": 7.375,
      "learning_rate": 2.504115068071687e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9929577708244324,
      "num_tokens": 1877896.0,
      "step": 4298
    },
    {
      "epoch": 0.33066687177909393,
      "grad_norm": 4.15625,
      "learning_rate": 2.503999692331359e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1878180.0,
      "step": 4299
    },
    {
      "epoch": 0.33074378893931233,
      "grad_norm": 5.25,
      "learning_rate": 2.5038843165910314e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1878488.0,
      "step": 4300
    },
    {
      "epoch": 0.3308207060995308,
      "grad_norm": 7.8125,
      "learning_rate": 2.503768940850704e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 1878806.0,
      "step": 4301
    },
    {
      "epoch": 0.33089762325974925,
      "grad_norm": 4.375,
      "learning_rate": 2.5036535651103763e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1879126.0,
      "step": 4302
    },
    {
      "epoch": 0.3309745404199677,
      "grad_norm": 3.734375,
      "learning_rate": 2.5035381893700486e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 1879952.0,
      "step": 4303
    },
    {
      "epoch": 0.3310514575801861,
      "grad_norm": 4.4375,
      "learning_rate": 2.503422813629721e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1880264.0,
      "step": 4304
    },
    {
      "epoch": 0.3311283747404046,
      "grad_norm": 4.78125,
      "learning_rate": 2.503307437889393e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 1880574.0,
      "step": 4305
    },
    {
      "epoch": 0.33120529190062303,
      "grad_norm": 2.625,
      "learning_rate": 2.5031920621490658e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9980732202529907,
      "num_tokens": 1881094.0,
      "step": 4306
    },
    {
      "epoch": 0.3312822090608415,
      "grad_norm": 2.609375,
      "learning_rate": 2.503076686408738e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 1881654.0,
      "step": 4307
    },
    {
      "epoch": 0.3313591262210599,
      "grad_norm": 3.34375,
      "learning_rate": 2.50296131066841e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 1882242.0,
      "step": 4308
    },
    {
      "epoch": 0.33143604338127836,
      "grad_norm": 5.40625,
      "learning_rate": 2.5028459349280825e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1882822.0,
      "step": 4309
    },
    {
      "epoch": 0.3315129605414968,
      "grad_norm": 5.71875,
      "learning_rate": 2.5027305591877548e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 1883154.0,
      "step": 4310
    },
    {
      "epoch": 0.3315898777017153,
      "grad_norm": 1.734375,
      "learning_rate": 2.502615183447427e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1883704.0,
      "step": 4311
    },
    {
      "epoch": 0.3316667948619337,
      "grad_norm": 3.734375,
      "learning_rate": 2.5024998077070997e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1884284.0,
      "step": 4312
    },
    {
      "epoch": 0.33174371202215214,
      "grad_norm": 6.0,
      "learning_rate": 2.5023844319667716e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1884608.0,
      "step": 4313
    },
    {
      "epoch": 0.3318206291823706,
      "grad_norm": 4.3125,
      "learning_rate": 2.5022690562264442e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 1885132.0,
      "step": 4314
    },
    {
      "epoch": 0.33189754634258906,
      "grad_norm": 5.8125,
      "learning_rate": 2.5021536804861165e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 1885442.0,
      "step": 4315
    },
    {
      "epoch": 0.33197446350280746,
      "grad_norm": 2.3125,
      "learning_rate": 2.5020383047457888e-06,
      "loss": 0.0072,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 1886502.0,
      "step": 4316
    },
    {
      "epoch": 0.3320513806630259,
      "grad_norm": 11.0,
      "learning_rate": 2.5019229290054614e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1886820.0,
      "step": 4317
    },
    {
      "epoch": 0.3321282978232444,
      "grad_norm": 5.65625,
      "learning_rate": 2.5018075532651337e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 1887176.0,
      "step": 4318
    },
    {
      "epoch": 0.33220521498346284,
      "grad_norm": 3.078125,
      "learning_rate": 2.5016921775248055e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1887746.0,
      "step": 4319
    },
    {
      "epoch": 0.33228213214368124,
      "grad_norm": 10.75,
      "learning_rate": 2.5015768017844782e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 1888024.0,
      "step": 4320
    },
    {
      "epoch": 0.3323590493038997,
      "grad_norm": 7.09375,
      "learning_rate": 2.5014614260441505e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 1888320.0,
      "step": 4321
    },
    {
      "epoch": 0.33243596646411816,
      "grad_norm": 11.9375,
      "learning_rate": 2.5013460503038227e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 1888578.0,
      "step": 4322
    },
    {
      "epoch": 0.33251288362433656,
      "grad_norm": 7.21875,
      "learning_rate": 2.5012306745634954e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1888878.0,
      "step": 4323
    },
    {
      "epoch": 0.332589800784555,
      "grad_norm": 4.5,
      "learning_rate": 2.5011152988231672e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 1889226.0,
      "step": 4324
    },
    {
      "epoch": 0.3326667179447735,
      "grad_norm": 7.4375,
      "learning_rate": 2.50099992308284e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 1889540.0,
      "step": 4325
    },
    {
      "epoch": 0.33274363510499194,
      "grad_norm": 2.171875,
      "learning_rate": 2.500884547342512e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 1890364.0,
      "step": 4326
    },
    {
      "epoch": 0.33282055226521035,
      "grad_norm": 6.6875,
      "learning_rate": 2.5007691716021844e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1890720.0,
      "step": 4327
    },
    {
      "epoch": 0.3328974694254288,
      "grad_norm": 2.890625,
      "learning_rate": 2.500653795861857e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1891264.0,
      "step": 4328
    },
    {
      "epoch": 0.33297438658564726,
      "grad_norm": 2.03125,
      "learning_rate": 2.5005384201215294e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1891848.0,
      "step": 4329
    },
    {
      "epoch": 0.3330513037458657,
      "grad_norm": 6.375,
      "learning_rate": 2.5004230443812012e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1892162.0,
      "step": 4330
    },
    {
      "epoch": 0.3331282209060841,
      "grad_norm": 6.71875,
      "learning_rate": 2.500307668640874e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1892490.0,
      "step": 4331
    },
    {
      "epoch": 0.3332051380663026,
      "grad_norm": 7.59375,
      "learning_rate": 2.500192292900546e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 1892798.0,
      "step": 4332
    },
    {
      "epoch": 0.33328205522652105,
      "grad_norm": 2.8125,
      "learning_rate": 2.5000769171602184e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 1893602.0,
      "step": 4333
    },
    {
      "epoch": 0.3333589723867395,
      "grad_norm": 4.25,
      "learning_rate": 2.499961541419891e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 1894202.0,
      "step": 4334
    },
    {
      "epoch": 0.3334358895469579,
      "grad_norm": 4.4375,
      "learning_rate": 2.499846165679563e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1894524.0,
      "step": 4335
    },
    {
      "epoch": 0.33351280670717637,
      "grad_norm": 3.390625,
      "learning_rate": 2.4997307899392356e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1895064.0,
      "step": 4336
    },
    {
      "epoch": 0.3335897238673948,
      "grad_norm": 7.53125,
      "learning_rate": 2.499615414198908e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1895348.0,
      "step": 4337
    },
    {
      "epoch": 0.3336666410276133,
      "grad_norm": 3.578125,
      "learning_rate": 2.49950003845858e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 1895884.0,
      "step": 4338
    },
    {
      "epoch": 0.3337435581878317,
      "grad_norm": 7.3125,
      "learning_rate": 2.499384662718253e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1896168.0,
      "step": 4339
    },
    {
      "epoch": 0.33382047534805015,
      "grad_norm": 6.15625,
      "learning_rate": 2.4992692869779246e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1896502.0,
      "step": 4340
    },
    {
      "epoch": 0.3338973925082686,
      "grad_norm": 6.375,
      "learning_rate": 2.499153911237597e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1896810.0,
      "step": 4341
    },
    {
      "epoch": 0.333974309668487,
      "grad_norm": 5.6875,
      "learning_rate": 2.4990385354972696e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1897138.0,
      "step": 4342
    },
    {
      "epoch": 0.33405122682870547,
      "grad_norm": 2.40625,
      "learning_rate": 2.498923159756942e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1897958.0,
      "step": 4343
    },
    {
      "epoch": 0.33412814398892393,
      "grad_norm": 2.671875,
      "learning_rate": 2.498807784016614e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 1898476.0,
      "step": 4344
    },
    {
      "epoch": 0.3342050611491424,
      "grad_norm": 4.09375,
      "learning_rate": 2.4986924082762868e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1898792.0,
      "step": 4345
    },
    {
      "epoch": 0.3342819783093608,
      "grad_norm": 5.09375,
      "learning_rate": 2.4985770325359586e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1899114.0,
      "step": 4346
    },
    {
      "epoch": 0.33435889546957925,
      "grad_norm": 4.3125,
      "learning_rate": 2.4984616567956313e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1899658.0,
      "step": 4347
    },
    {
      "epoch": 0.3344358126297977,
      "grad_norm": 3.96875,
      "learning_rate": 2.4983462810553036e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1900242.0,
      "step": 4348
    },
    {
      "epoch": 0.3345127297900162,
      "grad_norm": 24.625,
      "learning_rate": 2.498230905314976e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 1900582.0,
      "step": 4349
    },
    {
      "epoch": 0.3345896469502346,
      "grad_norm": 6.0,
      "learning_rate": 2.498115529574648e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1900894.0,
      "step": 4350
    },
    {
      "epoch": 0.33466656411045304,
      "grad_norm": 2.171875,
      "learning_rate": 2.4980001538343203e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 1901676.0,
      "step": 4351
    },
    {
      "epoch": 0.3347434812706715,
      "grad_norm": 4.90625,
      "learning_rate": 2.4978847780939926e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1901998.0,
      "step": 4352
    },
    {
      "epoch": 0.33482039843088995,
      "grad_norm": 2.234375,
      "learning_rate": 2.4977694023536653e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 1902568.0,
      "step": 4353
    },
    {
      "epoch": 0.33489731559110836,
      "grad_norm": 5.96875,
      "learning_rate": 2.4976540266133375e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 1903156.0,
      "step": 4354
    },
    {
      "epoch": 0.3349742327513268,
      "grad_norm": 4.84375,
      "learning_rate": 2.49753865087301e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 1903756.0,
      "step": 4355
    },
    {
      "epoch": 0.3350511499115453,
      "grad_norm": 5.125,
      "learning_rate": 2.4974232751326825e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 1904054.0,
      "step": 4356
    },
    {
      "epoch": 0.33512806707176374,
      "grad_norm": 6.34375,
      "learning_rate": 2.4973078993923543e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9970845580101013,
      "num_tokens": 1904398.0,
      "step": 4357
    },
    {
      "epoch": 0.33520498423198214,
      "grad_norm": 8.0,
      "learning_rate": 2.4971925236520266e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1904660.0,
      "step": 4358
    },
    {
      "epoch": 0.3352819013922006,
      "grad_norm": 2.875,
      "learning_rate": 2.4970771479116992e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 1905210.0,
      "step": 4359
    },
    {
      "epoch": 0.33535881855241906,
      "grad_norm": 2.4375,
      "learning_rate": 2.4969617721713715e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1905992.0,
      "step": 4360
    },
    {
      "epoch": 0.33543573571263746,
      "grad_norm": 2.84375,
      "learning_rate": 2.4968463964310438e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 1906536.0,
      "step": 4361
    },
    {
      "epoch": 0.3355126528728559,
      "grad_norm": 3.421875,
      "learning_rate": 2.496731020690716e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1907078.0,
      "step": 4362
    },
    {
      "epoch": 0.3355895700330744,
      "grad_norm": 3.9375,
      "learning_rate": 2.4966156449503883e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1907598.0,
      "step": 4363
    },
    {
      "epoch": 0.33566648719329284,
      "grad_norm": 6.5625,
      "learning_rate": 2.496500269210061e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1907908.0,
      "step": 4364
    },
    {
      "epoch": 0.33574340435351124,
      "grad_norm": 7.5,
      "learning_rate": 2.4963848934697332e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 1908196.0,
      "step": 4365
    },
    {
      "epoch": 0.3358203215137297,
      "grad_norm": 8.375,
      "learning_rate": 2.4962695177294055e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 1908476.0,
      "step": 4366
    },
    {
      "epoch": 0.33589723867394816,
      "grad_norm": 4.75,
      "learning_rate": 2.496154141989078e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1908804.0,
      "step": 4367
    },
    {
      "epoch": 0.3359741558341666,
      "grad_norm": 2.28125,
      "learning_rate": 2.49603876624875e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 1909348.0,
      "step": 4368
    },
    {
      "epoch": 0.336051072994385,
      "grad_norm": 4.9375,
      "learning_rate": 2.4959233905084222e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 1909686.0,
      "step": 4369
    },
    {
      "epoch": 0.3361279901546035,
      "grad_norm": 2.28125,
      "learning_rate": 2.495808014768095e-06,
      "loss": 0.0072,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 1910730.0,
      "step": 4370
    },
    {
      "epoch": 0.33620490731482194,
      "grad_norm": 5.46875,
      "learning_rate": 2.495692639027767e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1911046.0,
      "step": 4371
    },
    {
      "epoch": 0.3362818244750404,
      "grad_norm": 7.0625,
      "learning_rate": 2.4955772632874394e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 1911376.0,
      "step": 4372
    },
    {
      "epoch": 0.3363587416352588,
      "grad_norm": 6.75,
      "learning_rate": 2.4954618875471117e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1911658.0,
      "step": 4373
    },
    {
      "epoch": 0.33643565879547727,
      "grad_norm": 5.5,
      "learning_rate": 2.495346511806784e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 1911986.0,
      "step": 4374
    },
    {
      "epoch": 0.3365125759556957,
      "grad_norm": 5.625,
      "learning_rate": 2.4952311360664566e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1912310.0,
      "step": 4375
    },
    {
      "epoch": 0.3365894931159142,
      "grad_norm": 3.578125,
      "learning_rate": 2.495115760326129e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1912862.0,
      "step": 4376
    },
    {
      "epoch": 0.3366664102761326,
      "grad_norm": 2.34375,
      "learning_rate": 2.495000384585801e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 1913672.0,
      "step": 4377
    },
    {
      "epoch": 0.33674332743635105,
      "grad_norm": 3.765625,
      "learning_rate": 2.4948850088454734e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 1914248.0,
      "step": 4378
    },
    {
      "epoch": 0.3368202445965695,
      "grad_norm": 5.34375,
      "learning_rate": 2.4947696331051457e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1914530.0,
      "step": 4379
    },
    {
      "epoch": 0.33689716175678797,
      "grad_norm": 2.28125,
      "learning_rate": 2.494654257364818e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 1915340.0,
      "step": 4380
    },
    {
      "epoch": 0.33697407891700637,
      "grad_norm": 2.0,
      "learning_rate": 2.4945388816244906e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 1916150.0,
      "step": 4381
    },
    {
      "epoch": 0.33705099607722483,
      "grad_norm": 3.1875,
      "learning_rate": 2.494423505884163e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1916720.0,
      "step": 4382
    },
    {
      "epoch": 0.3371279132374433,
      "grad_norm": 6.90625,
      "learning_rate": 2.494308130143835e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1917038.0,
      "step": 4383
    },
    {
      "epoch": 0.3372048303976617,
      "grad_norm": 8.1875,
      "learning_rate": 2.4941927544035074e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 1917334.0,
      "step": 4384
    },
    {
      "epoch": 0.33728174755788015,
      "grad_norm": 3.15625,
      "learning_rate": 2.4940773786631796e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 1917884.0,
      "step": 4385
    },
    {
      "epoch": 0.3373586647180986,
      "grad_norm": 6.34375,
      "learning_rate": 2.4939620029228523e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 1918224.0,
      "step": 4386
    },
    {
      "epoch": 0.33743558187831707,
      "grad_norm": 3.671875,
      "learning_rate": 2.4938466271825246e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 1918784.0,
      "step": 4387
    },
    {
      "epoch": 0.3375124990385355,
      "grad_norm": 5.3125,
      "learning_rate": 2.493731251442197e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 1919098.0,
      "step": 4388
    },
    {
      "epoch": 0.33758941619875393,
      "grad_norm": 4.15625,
      "learning_rate": 2.493615875701869e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 1919686.0,
      "step": 4389
    },
    {
      "epoch": 0.3376663333589724,
      "grad_norm": 6.75,
      "learning_rate": 2.4935004999615414e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 1920020.0,
      "step": 4390
    },
    {
      "epoch": 0.33774325051919085,
      "grad_norm": 12.1875,
      "learning_rate": 2.4933851242212136e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 1920274.0,
      "step": 4391
    },
    {
      "epoch": 0.33782016767940926,
      "grad_norm": 2.375,
      "learning_rate": 2.4932697484808863e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1920858.0,
      "step": 4392
    },
    {
      "epoch": 0.3378970848396277,
      "grad_norm": 9.875,
      "learning_rate": 2.4931543727405586e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 1921144.0,
      "step": 4393
    },
    {
      "epoch": 0.3379740019998462,
      "grad_norm": 3.4375,
      "learning_rate": 2.493038997000231e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 1921684.0,
      "step": 4394
    },
    {
      "epoch": 0.33805091916006463,
      "grad_norm": 4.34375,
      "learning_rate": 2.492923621259903e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1921996.0,
      "step": 4395
    },
    {
      "epoch": 0.33812783632028304,
      "grad_norm": 5.8125,
      "learning_rate": 2.4928082455195753e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1922328.0,
      "step": 4396
    },
    {
      "epoch": 0.3382047534805015,
      "grad_norm": 5.375,
      "learning_rate": 2.492692869779248e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1922644.0,
      "step": 4397
    },
    {
      "epoch": 0.33828167064071996,
      "grad_norm": 1.6640625,
      "learning_rate": 2.4925774940389203e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1923190.0,
      "step": 4398
    },
    {
      "epoch": 0.3383585878009384,
      "grad_norm": 6.53125,
      "learning_rate": 2.4924621182985925e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 1923502.0,
      "step": 4399
    },
    {
      "epoch": 0.3384355049611568,
      "grad_norm": 7.8125,
      "learning_rate": 2.492346742558265e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 1923804.0,
      "step": 4400
    },
    {
      "epoch": 0.3385124221213753,
      "grad_norm": 5.9375,
      "learning_rate": 2.492231366817937e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 1924126.0,
      "step": 4401
    },
    {
      "epoch": 0.33858933928159374,
      "grad_norm": 4.375,
      "learning_rate": 2.4921159910776093e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 1924464.0,
      "step": 4402
    },
    {
      "epoch": 0.33866625644181214,
      "grad_norm": 7.84375,
      "learning_rate": 2.492000615337282e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 1924754.0,
      "step": 4403
    },
    {
      "epoch": 0.3387431736020306,
      "grad_norm": 2.1875,
      "learning_rate": 2.4918852395969542e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 1925564.0,
      "step": 4404
    },
    {
      "epoch": 0.33882009076224906,
      "grad_norm": 7.34375,
      "learning_rate": 2.4917698638566265e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 1925820.0,
      "step": 4405
    },
    {
      "epoch": 0.3388970079224675,
      "grad_norm": 4.9375,
      "learning_rate": 2.4916544881162988e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1926142.0,
      "step": 4406
    },
    {
      "epoch": 0.3389739250826859,
      "grad_norm": 7.5,
      "learning_rate": 2.491539112375971e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9845201373100281,
      "num_tokens": 1926466.0,
      "step": 4407
    },
    {
      "epoch": 0.3390508422429044,
      "grad_norm": 3.625,
      "learning_rate": 2.4914237366356433e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1927018.0,
      "step": 4408
    },
    {
      "epoch": 0.33912775940312284,
      "grad_norm": 2.546875,
      "learning_rate": 2.491308360895316e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1927838.0,
      "step": 4409
    },
    {
      "epoch": 0.3392046765633413,
      "grad_norm": 5.5625,
      "learning_rate": 2.4911929851549882e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 1928142.0,
      "step": 4410
    },
    {
      "epoch": 0.3392815937235597,
      "grad_norm": 5.125,
      "learning_rate": 2.4910776094146605e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 1928448.0,
      "step": 4411
    },
    {
      "epoch": 0.33935851088377816,
      "grad_norm": 2.78125,
      "learning_rate": 2.4909622336743327e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 1928992.0,
      "step": 4412
    },
    {
      "epoch": 0.3394354280439966,
      "grad_norm": 7.25,
      "learning_rate": 2.490846857934005e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1929302.0,
      "step": 4413
    },
    {
      "epoch": 0.3395123452042151,
      "grad_norm": 5.21875,
      "learning_rate": 2.4907314821936777e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9936708807945251,
      "num_tokens": 1929620.0,
      "step": 4414
    },
    {
      "epoch": 0.3395892623644335,
      "grad_norm": 7.84375,
      "learning_rate": 2.49061610645335e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 1929878.0,
      "step": 4415
    },
    {
      "epoch": 0.33966617952465195,
      "grad_norm": 4.34375,
      "learning_rate": 2.4905007307130218e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 1930172.0,
      "step": 4416
    },
    {
      "epoch": 0.3397430966848704,
      "grad_norm": 5.25,
      "learning_rate": 2.4903853549726944e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1930458.0,
      "step": 4417
    },
    {
      "epoch": 0.33982001384508886,
      "grad_norm": 8.3125,
      "learning_rate": 2.4902699792323667e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 1930738.0,
      "step": 4418
    },
    {
      "epoch": 0.33989693100530727,
      "grad_norm": 3.421875,
      "learning_rate": 2.490154603492039e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 1931310.0,
      "step": 4419
    },
    {
      "epoch": 0.3399738481655257,
      "grad_norm": 2.78125,
      "learning_rate": 2.4900392277517116e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 1931868.0,
      "step": 4420
    },
    {
      "epoch": 0.3400507653257442,
      "grad_norm": 3.359375,
      "learning_rate": 2.489923852011384e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 1932440.0,
      "step": 4421
    },
    {
      "epoch": 0.34012768248596265,
      "grad_norm": 8.4375,
      "learning_rate": 2.489808476271056e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1932796.0,
      "step": 4422
    },
    {
      "epoch": 0.34020459964618105,
      "grad_norm": 4.625,
      "learning_rate": 2.4896931005307284e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1933130.0,
      "step": 4423
    },
    {
      "epoch": 0.3402815168063995,
      "grad_norm": 2.453125,
      "learning_rate": 2.4895777247904007e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 1933886.0,
      "step": 4424
    },
    {
      "epoch": 0.34035843396661797,
      "grad_norm": 6.875,
      "learning_rate": 2.4894623490500734e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1934178.0,
      "step": 4425
    },
    {
      "epoch": 0.34043535112683637,
      "grad_norm": 5.4375,
      "learning_rate": 2.4893469733097456e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 1934502.0,
      "step": 4426
    },
    {
      "epoch": 0.34051226828705483,
      "grad_norm": 7.6875,
      "learning_rate": 2.4892315975694175e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 1934812.0,
      "step": 4427
    },
    {
      "epoch": 0.3405891854472733,
      "grad_norm": 7.25,
      "learning_rate": 2.48911622182909e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 1935114.0,
      "step": 4428
    },
    {
      "epoch": 0.34066610260749175,
      "grad_norm": 2.890625,
      "learning_rate": 2.4890008460887624e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1935684.0,
      "step": 4429
    },
    {
      "epoch": 0.34074301976771015,
      "grad_norm": 2.578125,
      "learning_rate": 2.4888854703484347e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 1936498.0,
      "step": 4430
    },
    {
      "epoch": 0.3408199369279286,
      "grad_norm": 2.1875,
      "learning_rate": 2.4887700946081073e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 1937296.0,
      "step": 4431
    },
    {
      "epoch": 0.34089685408814707,
      "grad_norm": 2.28125,
      "learning_rate": 2.4886547188677796e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1938078.0,
      "step": 4432
    },
    {
      "epoch": 0.34097377124836553,
      "grad_norm": 2.15625,
      "learning_rate": 2.488539343127452e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 1938662.0,
      "step": 4433
    },
    {
      "epoch": 0.34105068840858394,
      "grad_norm": 3.640625,
      "learning_rate": 2.488423967387124e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 1939230.0,
      "step": 4434
    },
    {
      "epoch": 0.3411276055688024,
      "grad_norm": 4.09375,
      "learning_rate": 2.4883085916467964e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 1939554.0,
      "step": 4435
    },
    {
      "epoch": 0.34120452272902085,
      "grad_norm": 12.0,
      "learning_rate": 2.488193215906469e-06,
      "loss": 0.038,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 1939810.0,
      "step": 4436
    },
    {
      "epoch": 0.3412814398892393,
      "grad_norm": 4.21875,
      "learning_rate": 2.4880778401661413e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1940132.0,
      "step": 4437
    },
    {
      "epoch": 0.3413583570494577,
      "grad_norm": 12.75,
      "learning_rate": 2.487962464425813e-06,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 1940392.0,
      "step": 4438
    },
    {
      "epoch": 0.3414352742096762,
      "grad_norm": 3.578125,
      "learning_rate": 2.487847088685486e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 1940984.0,
      "step": 4439
    },
    {
      "epoch": 0.34151219136989464,
      "grad_norm": 7.90625,
      "learning_rate": 2.487731712945158e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 1941290.0,
      "step": 4440
    },
    {
      "epoch": 0.3415891085301131,
      "grad_norm": 7.75,
      "learning_rate": 2.4876163372048303e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1941618.0,
      "step": 4441
    },
    {
      "epoch": 0.3416660256903315,
      "grad_norm": 2.59375,
      "learning_rate": 2.487500961464503e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 1942164.0,
      "step": 4442
    },
    {
      "epoch": 0.34174294285054996,
      "grad_norm": 13.6875,
      "learning_rate": 2.4873855857241753e-06,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 1942398.0,
      "step": 4443
    },
    {
      "epoch": 0.3418198600107684,
      "grad_norm": 6.84375,
      "learning_rate": 2.4872702099838475e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 1942758.0,
      "step": 4444
    },
    {
      "epoch": 0.3418967771709868,
      "grad_norm": 6.84375,
      "learning_rate": 2.48715483424352e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1943120.0,
      "step": 4445
    },
    {
      "epoch": 0.3419736943312053,
      "grad_norm": 3.734375,
      "learning_rate": 2.487039458503192e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1943700.0,
      "step": 4446
    },
    {
      "epoch": 0.34205061149142374,
      "grad_norm": 3.5,
      "learning_rate": 2.4869240827628647e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 1944264.0,
      "step": 4447
    },
    {
      "epoch": 0.3421275286516422,
      "grad_norm": 5.03125,
      "learning_rate": 2.486808707022537e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 1944584.0,
      "step": 4448
    },
    {
      "epoch": 0.3422044458118606,
      "grad_norm": 4.8125,
      "learning_rate": 2.486693331282209e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1944930.0,
      "step": 4449
    },
    {
      "epoch": 0.34228136297207906,
      "grad_norm": 3.21875,
      "learning_rate": 2.4865779555418815e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 1945470.0,
      "step": 4450
    },
    {
      "epoch": 0.3423582801322975,
      "grad_norm": 5.65625,
      "learning_rate": 2.4864625798015538e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 1945796.0,
      "step": 4451
    },
    {
      "epoch": 0.342435197292516,
      "grad_norm": 4.1875,
      "learning_rate": 2.486347204061226e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 1946130.0,
      "step": 4452
    },
    {
      "epoch": 0.3425121144527344,
      "grad_norm": 2.84375,
      "learning_rate": 2.4862318283208987e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 1946682.0,
      "step": 4453
    },
    {
      "epoch": 0.34258903161295284,
      "grad_norm": 7.09375,
      "learning_rate": 2.4861164525805705e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 1947036.0,
      "step": 4454
    },
    {
      "epoch": 0.3426659487731713,
      "grad_norm": 3.296875,
      "learning_rate": 2.4860010768402432e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 1947624.0,
      "step": 4455
    },
    {
      "epoch": 0.34274286593338976,
      "grad_norm": 3.171875,
      "learning_rate": 2.4858857010999155e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 1948214.0,
      "step": 4456
    },
    {
      "epoch": 0.34281978309360817,
      "grad_norm": 4.65625,
      "learning_rate": 2.4857703253595877e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9915966391563416,
      "num_tokens": 1948810.0,
      "step": 4457
    },
    {
      "epoch": 0.3428967002538266,
      "grad_norm": 5.125,
      "learning_rate": 2.4856549496192604e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1949380.0,
      "step": 4458
    },
    {
      "epoch": 0.3429736174140451,
      "grad_norm": 2.921875,
      "learning_rate": 2.4855395738789327e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 1949960.0,
      "step": 4459
    },
    {
      "epoch": 0.34305053457426354,
      "grad_norm": 4.78125,
      "learning_rate": 2.4854241981386045e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.990328848361969,
      "num_tokens": 1950478.0,
      "step": 4460
    },
    {
      "epoch": 0.34312745173448195,
      "grad_norm": 2.359375,
      "learning_rate": 2.485308822398277e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 1950828.0,
      "step": 4461
    },
    {
      "epoch": 0.3432043688947004,
      "grad_norm": 6.125,
      "learning_rate": 2.4851934466579495e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 1951116.0,
      "step": 4462
    },
    {
      "epoch": 0.34328128605491887,
      "grad_norm": 2.375,
      "learning_rate": 2.4850780709176217e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 1951914.0,
      "step": 4463
    },
    {
      "epoch": 0.34335820321513727,
      "grad_norm": 4.3125,
      "learning_rate": 2.4849626951772944e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 1952252.0,
      "step": 4464
    },
    {
      "epoch": 0.34343512037535573,
      "grad_norm": 3.671875,
      "learning_rate": 2.4848473194369662e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1952824.0,
      "step": 4465
    },
    {
      "epoch": 0.3435120375355742,
      "grad_norm": 7.625,
      "learning_rate": 2.484731943696639e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9931034445762634,
      "num_tokens": 1953116.0,
      "step": 4466
    },
    {
      "epoch": 0.34358895469579265,
      "grad_norm": 4.90625,
      "learning_rate": 2.484616567956311e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 1953444.0,
      "step": 4467
    },
    {
      "epoch": 0.34366587185601105,
      "grad_norm": 9.125,
      "learning_rate": 2.4845011922159834e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1953778.0,
      "step": 4468
    },
    {
      "epoch": 0.3437427890162295,
      "grad_norm": 5.15625,
      "learning_rate": 2.4843858164756557e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1954112.0,
      "step": 4469
    },
    {
      "epoch": 0.34381970617644797,
      "grad_norm": 5.78125,
      "learning_rate": 2.4842704407353284e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 1954446.0,
      "step": 4470
    },
    {
      "epoch": 0.34389662333666643,
      "grad_norm": 4.65625,
      "learning_rate": 2.484155064995e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.996835470199585,
      "num_tokens": 1954764.0,
      "step": 4471
    },
    {
      "epoch": 0.34397354049688483,
      "grad_norm": 3.3125,
      "learning_rate": 2.484039689254673e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1955334.0,
      "step": 4472
    },
    {
      "epoch": 0.3440504576571033,
      "grad_norm": 2.984375,
      "learning_rate": 2.483924313514345e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 1956154.0,
      "step": 4473
    },
    {
      "epoch": 0.34412737481732175,
      "grad_norm": 4.03125,
      "learning_rate": 2.4838089377740174e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 1956696.0,
      "step": 4474
    },
    {
      "epoch": 0.3442042919775402,
      "grad_norm": 5.9375,
      "learning_rate": 2.48369356203369e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1956980.0,
      "step": 4475
    },
    {
      "epoch": 0.3442812091377586,
      "grad_norm": 3.265625,
      "learning_rate": 2.483578186293362e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1957566.0,
      "step": 4476
    },
    {
      "epoch": 0.3443581262979771,
      "grad_norm": 3.71875,
      "learning_rate": 2.483462810553034e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1958098.0,
      "step": 4477
    },
    {
      "epoch": 0.34443504345819553,
      "grad_norm": 6.9375,
      "learning_rate": 2.483347434812707e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1958398.0,
      "step": 4478
    },
    {
      "epoch": 0.344511960618414,
      "grad_norm": 8.9375,
      "learning_rate": 2.483232059072379e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1958714.0,
      "step": 4479
    },
    {
      "epoch": 0.3445888777786324,
      "grad_norm": 3.484375,
      "learning_rate": 2.4831166833320514e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 1959246.0,
      "step": 4480
    },
    {
      "epoch": 0.34466579493885086,
      "grad_norm": 6.78125,
      "learning_rate": 2.4830013075917236e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 1959538.0,
      "step": 4481
    },
    {
      "epoch": 0.3447427120990693,
      "grad_norm": 2.171875,
      "learning_rate": 2.482885931851396e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 1960112.0,
      "step": 4482
    },
    {
      "epoch": 0.3448196292592878,
      "grad_norm": 9.1875,
      "learning_rate": 2.4827705561110686e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 1960372.0,
      "step": 4483
    },
    {
      "epoch": 0.3448965464195062,
      "grad_norm": 2.78125,
      "learning_rate": 2.482655180370741e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 1960916.0,
      "step": 4484
    },
    {
      "epoch": 0.34497346357972464,
      "grad_norm": 2.0,
      "learning_rate": 2.482539804630413e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 1961674.0,
      "step": 4485
    },
    {
      "epoch": 0.3450503807399431,
      "grad_norm": 6.34375,
      "learning_rate": 2.4824244288900858e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1961956.0,
      "step": 4486
    },
    {
      "epoch": 0.3451272979001615,
      "grad_norm": 4.6875,
      "learning_rate": 2.4823090531497576e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 1962272.0,
      "step": 4487
    },
    {
      "epoch": 0.34520421506037996,
      "grad_norm": 2.3125,
      "learning_rate": 2.48219367740943e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 1962796.0,
      "step": 4488
    },
    {
      "epoch": 0.3452811322205984,
      "grad_norm": 8.0625,
      "learning_rate": 2.4820783016691025e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1963152.0,
      "step": 4489
    },
    {
      "epoch": 0.3453580493808169,
      "grad_norm": 5.375,
      "learning_rate": 2.481962925928775e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1963460.0,
      "step": 4490
    },
    {
      "epoch": 0.3454349665410353,
      "grad_norm": 6.28125,
      "learning_rate": 2.481847550188447e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 1963810.0,
      "step": 4491
    },
    {
      "epoch": 0.34551188370125374,
      "grad_norm": 8.0,
      "learning_rate": 2.4817321744481193e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1964138.0,
      "step": 4492
    },
    {
      "epoch": 0.3455888008614722,
      "grad_norm": 7.21875,
      "learning_rate": 2.4816167987077916e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1964456.0,
      "step": 4493
    },
    {
      "epoch": 0.34566571802169066,
      "grad_norm": 6.96875,
      "learning_rate": 2.4815014229674643e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.992094874382019,
      "num_tokens": 1964710.0,
      "step": 4494
    },
    {
      "epoch": 0.34574263518190906,
      "grad_norm": 3.046875,
      "learning_rate": 2.4813860472271365e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 1965272.0,
      "step": 4495
    },
    {
      "epoch": 0.3458195523421275,
      "grad_norm": 4.625,
      "learning_rate": 2.4812706714868088e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 1965572.0,
      "step": 4496
    },
    {
      "epoch": 0.345896469502346,
      "grad_norm": 9.9375,
      "learning_rate": 2.4811552957464815e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1965874.0,
      "step": 4497
    },
    {
      "epoch": 0.34597338666256444,
      "grad_norm": 10.0,
      "learning_rate": 2.4810399200061533e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 1966106.0,
      "step": 4498
    },
    {
      "epoch": 0.34605030382278285,
      "grad_norm": 6.5625,
      "learning_rate": 2.4809245442658255e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 1966462.0,
      "step": 4499
    },
    {
      "epoch": 0.3461272209830013,
      "grad_norm": 2.859375,
      "learning_rate": 2.4808091685254982e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 1967048.0,
      "step": 4500
    },
    {
      "epoch": 0.34620413814321976,
      "grad_norm": 7.75,
      "learning_rate": 2.4806937927851705e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 1967376.0,
      "step": 4501
    },
    {
      "epoch": 0.3462810553034382,
      "grad_norm": 8.125,
      "learning_rate": 2.4805784170448427e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 1967680.0,
      "step": 4502
    },
    {
      "epoch": 0.3463579724636566,
      "grad_norm": 4.28125,
      "learning_rate": 2.480463041304515e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9904030561447144,
      "num_tokens": 1968202.0,
      "step": 4503
    },
    {
      "epoch": 0.3464348896238751,
      "grad_norm": 5.0,
      "learning_rate": 2.4803476655641873e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1968486.0,
      "step": 4504
    },
    {
      "epoch": 0.34651180678409355,
      "grad_norm": 10.75,
      "learning_rate": 2.48023228982386e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 1968824.0,
      "step": 4505
    },
    {
      "epoch": 0.34658872394431195,
      "grad_norm": 10.9375,
      "learning_rate": 2.480116914083532e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 1969162.0,
      "step": 4506
    },
    {
      "epoch": 0.3466656411045304,
      "grad_norm": 7.9375,
      "learning_rate": 2.4800015383432045e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 1969486.0,
      "step": 4507
    },
    {
      "epoch": 0.34674255826474887,
      "grad_norm": 7.4375,
      "learning_rate": 2.479886162602877e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9969419240951538,
      "num_tokens": 1969814.0,
      "step": 4508
    },
    {
      "epoch": 0.3468194754249673,
      "grad_norm": 2.546875,
      "learning_rate": 2.479770786862549e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1970376.0,
      "step": 4509
    },
    {
      "epoch": 0.34689639258518573,
      "grad_norm": 2.71875,
      "learning_rate": 2.4796554111222212e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1970922.0,
      "step": 4510
    },
    {
      "epoch": 0.3469733097454042,
      "grad_norm": 6.78125,
      "learning_rate": 2.479540035381894e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 1971184.0,
      "step": 4511
    },
    {
      "epoch": 0.34705022690562265,
      "grad_norm": 2.046875,
      "learning_rate": 2.479424659641566e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1971730.0,
      "step": 4512
    },
    {
      "epoch": 0.3471271440658411,
      "grad_norm": 5.40625,
      "learning_rate": 2.4793092839012384e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 1972036.0,
      "step": 4513
    },
    {
      "epoch": 0.3472040612260595,
      "grad_norm": 8.25,
      "learning_rate": 2.4791939081609107e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 1972352.0,
      "step": 4514
    },
    {
      "epoch": 0.34728097838627797,
      "grad_norm": 6.15625,
      "learning_rate": 2.479078532420583e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 1972712.0,
      "step": 4515
    },
    {
      "epoch": 0.34735789554649643,
      "grad_norm": 5.75,
      "learning_rate": 2.4789631566802556e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 1973050.0,
      "step": 4516
    },
    {
      "epoch": 0.3474348127067149,
      "grad_norm": 3.734375,
      "learning_rate": 2.478847780939928e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1973372.0,
      "step": 4517
    },
    {
      "epoch": 0.3475117298669333,
      "grad_norm": 7.625,
      "learning_rate": 2.4787324051996e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 1973704.0,
      "step": 4518
    },
    {
      "epoch": 0.34758864702715175,
      "grad_norm": 4.59375,
      "learning_rate": 2.4786170294592724e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 1974012.0,
      "step": 4519
    },
    {
      "epoch": 0.3476655641873702,
      "grad_norm": 5.375,
      "learning_rate": 2.4785016537189447e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 1974320.0,
      "step": 4520
    },
    {
      "epoch": 0.34774248134758867,
      "grad_norm": 5.78125,
      "learning_rate": 2.478386277978617e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 1974638.0,
      "step": 4521
    },
    {
      "epoch": 0.3478193985078071,
      "grad_norm": 2.828125,
      "learning_rate": 2.4782709022382896e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 1975184.0,
      "step": 4522
    },
    {
      "epoch": 0.34789631566802554,
      "grad_norm": 7.0,
      "learning_rate": 2.478155526497962e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 1975466.0,
      "step": 4523
    },
    {
      "epoch": 0.347973232828244,
      "grad_norm": 3.5625,
      "learning_rate": 2.478040150757634e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 1976036.0,
      "step": 4524
    },
    {
      "epoch": 0.3480501499884624,
      "grad_norm": 10.125,
      "learning_rate": 2.4779247750173064e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 1976338.0,
      "step": 4525
    },
    {
      "epoch": 0.34812706714868086,
      "grad_norm": 8.375,
      "learning_rate": 2.4778093992769786e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 1976650.0,
      "step": 4526
    },
    {
      "epoch": 0.3482039843088993,
      "grad_norm": 5.21875,
      "learning_rate": 2.477694023536651e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 1976968.0,
      "step": 4527
    },
    {
      "epoch": 0.3482809014691178,
      "grad_norm": 4.9375,
      "learning_rate": 2.4775786477963236e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 1977266.0,
      "step": 4528
    },
    {
      "epoch": 0.3483578186293362,
      "grad_norm": 6.5,
      "learning_rate": 2.477463272055996e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 1977596.0,
      "step": 4529
    },
    {
      "epoch": 0.34843473578955464,
      "grad_norm": 2.640625,
      "learning_rate": 2.477347896315668e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 1978416.0,
      "step": 4530
    },
    {
      "epoch": 0.3485116529497731,
      "grad_norm": 7.375,
      "learning_rate": 2.4772325205753403e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9930070042610168,
      "num_tokens": 1978704.0,
      "step": 4531
    },
    {
      "epoch": 0.34858857010999156,
      "grad_norm": 5.4375,
      "learning_rate": 2.4771171448350126e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 1978990.0,
      "step": 4532
    },
    {
      "epoch": 0.34866548727020996,
      "grad_norm": 7.4375,
      "learning_rate": 2.4770017690946853e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1979326.0,
      "step": 4533
    },
    {
      "epoch": 0.3487424044304284,
      "grad_norm": 3.921875,
      "learning_rate": 2.4768863933543575e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1979898.0,
      "step": 4534
    },
    {
      "epoch": 0.3488193215906469,
      "grad_norm": 4.90625,
      "learning_rate": 2.47677101761403e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 1980220.0,
      "step": 4535
    },
    {
      "epoch": 0.34889623875086534,
      "grad_norm": 3.796875,
      "learning_rate": 2.476655641873702e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 1980816.0,
      "step": 4536
    },
    {
      "epoch": 0.34897315591108374,
      "grad_norm": 3.078125,
      "learning_rate": 2.4765402661333743e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 1981400.0,
      "step": 4537
    },
    {
      "epoch": 0.3490500730713022,
      "grad_norm": 3.0,
      "learning_rate": 2.4764248903930466e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 1981984.0,
      "step": 4538
    },
    {
      "epoch": 0.34912699023152066,
      "grad_norm": 5.84375,
      "learning_rate": 2.4763095146527193e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9963099360466003,
      "num_tokens": 1982256.0,
      "step": 4539
    },
    {
      "epoch": 0.3492039073917391,
      "grad_norm": 3.125,
      "learning_rate": 2.4761941389123915e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 1982836.0,
      "step": 4540
    },
    {
      "epoch": 0.3492808245519575,
      "grad_norm": 4.9375,
      "learning_rate": 2.4760787631720638e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 1983180.0,
      "step": 4541
    },
    {
      "epoch": 0.349357741712176,
      "grad_norm": 7.40625,
      "learning_rate": 2.475963387431736e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9826989769935608,
      "num_tokens": 1983470.0,
      "step": 4542
    },
    {
      "epoch": 0.34943465887239444,
      "grad_norm": 6.34375,
      "learning_rate": 2.4758480116914083e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 1983796.0,
      "step": 4543
    },
    {
      "epoch": 0.3495115760326129,
      "grad_norm": 3.390625,
      "learning_rate": 2.475732635951081e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 1984398.0,
      "step": 4544
    },
    {
      "epoch": 0.3495884931928313,
      "grad_norm": 6.03125,
      "learning_rate": 2.4756172602107532e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 1984738.0,
      "step": 4545
    },
    {
      "epoch": 0.34966541035304977,
      "grad_norm": 2.296875,
      "learning_rate": 2.4755018844704255e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 1985564.0,
      "step": 4546
    },
    {
      "epoch": 0.3497423275132682,
      "grad_norm": 6.3125,
      "learning_rate": 2.4753865087300977e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 1985920.0,
      "step": 4547
    },
    {
      "epoch": 0.34981924467348663,
      "grad_norm": 3.984375,
      "learning_rate": 2.47527113298977e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 1986464.0,
      "step": 4548
    },
    {
      "epoch": 0.3498961618337051,
      "grad_norm": 6.34375,
      "learning_rate": 2.4751557572494423e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9858356714248657,
      "num_tokens": 1986818.0,
      "step": 4549
    },
    {
      "epoch": 0.34997307899392355,
      "grad_norm": 4.5,
      "learning_rate": 2.475040381509115e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 1987338.0,
      "step": 4550
    },
    {
      "epoch": 0.350049996154142,
      "grad_norm": 8.8125,
      "learning_rate": 2.474925005768787e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 1987646.0,
      "step": 4551
    },
    {
      "epoch": 0.3501269133143604,
      "grad_norm": 3.0625,
      "learning_rate": 2.4748096300284595e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 1988182.0,
      "step": 4552
    },
    {
      "epoch": 0.35020383047457887,
      "grad_norm": 3.703125,
      "learning_rate": 2.4746942542881317e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 1988708.0,
      "step": 4553
    },
    {
      "epoch": 0.35028074763479733,
      "grad_norm": 4.0625,
      "learning_rate": 2.474578878547804e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 1989054.0,
      "step": 4554
    },
    {
      "epoch": 0.3503576647950158,
      "grad_norm": 3.46875,
      "learning_rate": 2.4744635028074767e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 1989376.0,
      "step": 4555
    },
    {
      "epoch": 0.3504345819552342,
      "grad_norm": 5.8125,
      "learning_rate": 2.474348127067149e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 1989738.0,
      "step": 4556
    },
    {
      "epoch": 0.35051149911545265,
      "grad_norm": 7.3125,
      "learning_rate": 2.4742327513268207e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 1990028.0,
      "step": 4557
    },
    {
      "epoch": 0.3505884162756711,
      "grad_norm": 7.4375,
      "learning_rate": 2.4741173755864934e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 1990356.0,
      "step": 4558
    },
    {
      "epoch": 0.35066533343588957,
      "grad_norm": 5.84375,
      "learning_rate": 2.4740019998461657e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 1990640.0,
      "step": 4559
    },
    {
      "epoch": 0.350742250596108,
      "grad_norm": 3.859375,
      "learning_rate": 2.473886624105838e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 1991204.0,
      "step": 4560
    },
    {
      "epoch": 0.35081916775632643,
      "grad_norm": 4.3125,
      "learning_rate": 2.4737712483655106e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 1991806.0,
      "step": 4561
    },
    {
      "epoch": 0.3508960849165449,
      "grad_norm": 5.21875,
      "learning_rate": 2.473655872625183e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 1992112.0,
      "step": 4562
    },
    {
      "epoch": 0.35097300207676335,
      "grad_norm": 2.09375,
      "learning_rate": 2.473540496884855e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 1992894.0,
      "step": 4563
    },
    {
      "epoch": 0.35104991923698176,
      "grad_norm": 6.78125,
      "learning_rate": 2.4734251211445274e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 1993176.0,
      "step": 4564
    },
    {
      "epoch": 0.3511268363972002,
      "grad_norm": 3.453125,
      "learning_rate": 2.4733097454041997e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 1993488.0,
      "step": 4565
    },
    {
      "epoch": 0.3512037535574187,
      "grad_norm": 2.296875,
      "learning_rate": 2.4731943696638723e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 1994062.0,
      "step": 4566
    },
    {
      "epoch": 0.3512806707176371,
      "grad_norm": 5.78125,
      "learning_rate": 2.4730789939235446e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 1994376.0,
      "step": 4567
    },
    {
      "epoch": 0.35135758787785554,
      "grad_norm": 2.234375,
      "learning_rate": 2.4729636181832164e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 1995156.0,
      "step": 4568
    },
    {
      "epoch": 0.351434505038074,
      "grad_norm": 6.21875,
      "learning_rate": 2.472848242442889e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 1995492.0,
      "step": 4569
    },
    {
      "epoch": 0.35151142219829246,
      "grad_norm": 2.734375,
      "learning_rate": 2.4727328667025614e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 1996010.0,
      "step": 4570
    },
    {
      "epoch": 0.35158833935851086,
      "grad_norm": 3.71875,
      "learning_rate": 2.4726174909622336e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 1996506.0,
      "step": 4571
    },
    {
      "epoch": 0.3516652565187293,
      "grad_norm": 4.96875,
      "learning_rate": 2.4725021152219063e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 1997078.0,
      "step": 4572
    },
    {
      "epoch": 0.3517421736789478,
      "grad_norm": 5.375,
      "learning_rate": 2.4723867394815786e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 1997418.0,
      "step": 4573
    },
    {
      "epoch": 0.35181909083916624,
      "grad_norm": 3.203125,
      "learning_rate": 2.472271363741251e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 1997978.0,
      "step": 4574
    },
    {
      "epoch": 0.35189600799938464,
      "grad_norm": 4.65625,
      "learning_rate": 2.472155988000923e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 1998296.0,
      "step": 4575
    },
    {
      "epoch": 0.3519729251596031,
      "grad_norm": 5.8125,
      "learning_rate": 2.4720406122605953e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 1998632.0,
      "step": 4576
    },
    {
      "epoch": 0.35204984231982156,
      "grad_norm": 3.21875,
      "learning_rate": 2.4719252365202676e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 1999194.0,
      "step": 4577
    },
    {
      "epoch": 0.35212675948004,
      "grad_norm": 3.0625,
      "learning_rate": 2.4718098607799403e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 1999754.0,
      "step": 4578
    },
    {
      "epoch": 0.3522036766402584,
      "grad_norm": 5.53125,
      "learning_rate": 2.471694485039612e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 2000074.0,
      "step": 4579
    },
    {
      "epoch": 0.3522805938004769,
      "grad_norm": 10.0,
      "learning_rate": 2.471579109299285e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2000328.0,
      "step": 4580
    },
    {
      "epoch": 0.35235751096069534,
      "grad_norm": 7.53125,
      "learning_rate": 2.471463733558957e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2000610.0,
      "step": 4581
    },
    {
      "epoch": 0.3524344281209138,
      "grad_norm": 2.0625,
      "learning_rate": 2.4713483578186293e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2001156.0,
      "step": 4582
    },
    {
      "epoch": 0.3525113452811322,
      "grad_norm": 9.4375,
      "learning_rate": 2.471232982078302e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 2001432.0,
      "step": 4583
    },
    {
      "epoch": 0.35258826244135066,
      "grad_norm": 6.875,
      "learning_rate": 2.471117606337974e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 2001740.0,
      "step": 4584
    },
    {
      "epoch": 0.3526651796015691,
      "grad_norm": 6.3125,
      "learning_rate": 2.471002230597646e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 2002018.0,
      "step": 4585
    },
    {
      "epoch": 0.3527420967617876,
      "grad_norm": 4.84375,
      "learning_rate": 2.4708868548573188e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2002336.0,
      "step": 4586
    },
    {
      "epoch": 0.352819013922006,
      "grad_norm": 6.65625,
      "learning_rate": 2.470771479116991e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2002632.0,
      "step": 4587
    },
    {
      "epoch": 0.35289593108222445,
      "grad_norm": 3.78125,
      "learning_rate": 2.4706561033766633e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2002916.0,
      "step": 4588
    },
    {
      "epoch": 0.3529728482424429,
      "grad_norm": 6.3125,
      "learning_rate": 2.470540727636336e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2003240.0,
      "step": 4589
    },
    {
      "epoch": 0.3530497654026613,
      "grad_norm": 6.59375,
      "learning_rate": 2.470425351896008e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 2003590.0,
      "step": 4590
    },
    {
      "epoch": 0.35312668256287977,
      "grad_norm": 4.59375,
      "learning_rate": 2.4703099761556805e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2003910.0,
      "step": 4591
    },
    {
      "epoch": 0.3532035997230982,
      "grad_norm": 5.125,
      "learning_rate": 2.4701946004153527e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2004226.0,
      "step": 4592
    },
    {
      "epoch": 0.3532805168833167,
      "grad_norm": 2.4375,
      "learning_rate": 2.470079224675025e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9962335228919983,
      "num_tokens": 2004758.0,
      "step": 4593
    },
    {
      "epoch": 0.3533574340435351,
      "grad_norm": 4.6875,
      "learning_rate": 2.4699638489346977e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 2005080.0,
      "step": 4594
    },
    {
      "epoch": 0.35343435120375355,
      "grad_norm": 3.359375,
      "learning_rate": 2.4698484731943695e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2005396.0,
      "step": 4595
    },
    {
      "epoch": 0.353511268363972,
      "grad_norm": 7.46875,
      "learning_rate": 2.4697330974540418e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 2005674.0,
      "step": 4596
    },
    {
      "epoch": 0.35358818552419047,
      "grad_norm": 3.234375,
      "learning_rate": 2.4696177217137145e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2006192.0,
      "step": 4597
    },
    {
      "epoch": 0.35366510268440887,
      "grad_norm": 6.0625,
      "learning_rate": 2.4695023459733867e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 2006742.0,
      "step": 4598
    },
    {
      "epoch": 0.35374201984462733,
      "grad_norm": 13.1875,
      "learning_rate": 2.469386970233059e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9789789915084839,
      "num_tokens": 2007078.0,
      "step": 4599
    },
    {
      "epoch": 0.3538189370048458,
      "grad_norm": 4.59375,
      "learning_rate": 2.4692715944927317e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 2007680.0,
      "step": 4600
    },
    {
      "epoch": 0.35389585416506425,
      "grad_norm": 4.875,
      "learning_rate": 2.4691562187524035e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 2008272.0,
      "step": 4601
    },
    {
      "epoch": 0.35397277132528265,
      "grad_norm": 5.46875,
      "learning_rate": 2.469040843012076e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2008602.0,
      "step": 4602
    },
    {
      "epoch": 0.3540496884855011,
      "grad_norm": 4.3125,
      "learning_rate": 2.4689254672717484e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2008896.0,
      "step": 4603
    },
    {
      "epoch": 0.35412660564571957,
      "grad_norm": 4.75,
      "learning_rate": 2.4688100915314207e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2009230.0,
      "step": 4604
    },
    {
      "epoch": 0.35420352280593803,
      "grad_norm": 9.875,
      "learning_rate": 2.4686947157910934e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 2009538.0,
      "step": 4605
    },
    {
      "epoch": 0.35428043996615644,
      "grad_norm": 4.84375,
      "learning_rate": 2.468579340050765e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2009864.0,
      "step": 4606
    },
    {
      "epoch": 0.3543573571263749,
      "grad_norm": 2.5,
      "learning_rate": 2.4684639643104375e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 2010448.0,
      "step": 4607
    },
    {
      "epoch": 0.35443427428659335,
      "grad_norm": 2.703125,
      "learning_rate": 2.46834858857011e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2010994.0,
      "step": 4608
    },
    {
      "epoch": 0.35451119144681176,
      "grad_norm": 3.296875,
      "learning_rate": 2.4682332128297824e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2011512.0,
      "step": 4609
    },
    {
      "epoch": 0.3545881086070302,
      "grad_norm": 3.109375,
      "learning_rate": 2.4681178370894547e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2012086.0,
      "step": 4610
    },
    {
      "epoch": 0.3546650257672487,
      "grad_norm": 3.3125,
      "learning_rate": 2.4680024613491273e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2012630.0,
      "step": 4611
    },
    {
      "epoch": 0.35474194292746714,
      "grad_norm": 3.890625,
      "learning_rate": 2.467887085608799e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2013204.0,
      "step": 4612
    },
    {
      "epoch": 0.35481886008768554,
      "grad_norm": 2.421875,
      "learning_rate": 2.467771709868472e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 2013984.0,
      "step": 4613
    },
    {
      "epoch": 0.354895777247904,
      "grad_norm": 2.9375,
      "learning_rate": 2.467656334128144e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 2014788.0,
      "step": 4614
    },
    {
      "epoch": 0.35497269440812246,
      "grad_norm": 4.6875,
      "learning_rate": 2.4675409583878164e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2015364.0,
      "step": 4615
    },
    {
      "epoch": 0.3550496115683409,
      "grad_norm": 3.734375,
      "learning_rate": 2.467425582647489e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2015916.0,
      "step": 4616
    },
    {
      "epoch": 0.3551265287285593,
      "grad_norm": 1.953125,
      "learning_rate": 2.467310206907161e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2016500.0,
      "step": 4617
    },
    {
      "epoch": 0.3552034458887778,
      "grad_norm": 5.8125,
      "learning_rate": 2.467194831166833e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2016856.0,
      "step": 4618
    },
    {
      "epoch": 0.35528036304899624,
      "grad_norm": 2.109375,
      "learning_rate": 2.467079455426506e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2017614.0,
      "step": 4619
    },
    {
      "epoch": 0.3553572802092147,
      "grad_norm": 1.890625,
      "learning_rate": 2.466964079686178e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9974554777145386,
      "num_tokens": 2018402.0,
      "step": 4620
    },
    {
      "epoch": 0.3554341973694331,
      "grad_norm": 1.6953125,
      "learning_rate": 2.4668487039458503e-06,
      "loss": 0.0062,
      "mean_token_accuracy": 0.999018669128418,
      "num_tokens": 2019424.0,
      "step": 4621
    },
    {
      "epoch": 0.35551111452965156,
      "grad_norm": 7.0,
      "learning_rate": 2.4667333282055226e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 2019738.0,
      "step": 4622
    },
    {
      "epoch": 0.35558803168987,
      "grad_norm": 2.34375,
      "learning_rate": 2.466617952465195e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2020294.0,
      "step": 4623
    },
    {
      "epoch": 0.3556649488500885,
      "grad_norm": 10.4375,
      "learning_rate": 2.4665025767248675e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2020576.0,
      "step": 4624
    },
    {
      "epoch": 0.3557418660103069,
      "grad_norm": 3.859375,
      "learning_rate": 2.46638720098454e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 2021098.0,
      "step": 4625
    },
    {
      "epoch": 0.35581878317052534,
      "grad_norm": 3.046875,
      "learning_rate": 2.466271825244212e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2021630.0,
      "step": 4626
    },
    {
      "epoch": 0.3558957003307438,
      "grad_norm": 2.9375,
      "learning_rate": 2.4661564495038843e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2022216.0,
      "step": 4627
    },
    {
      "epoch": 0.3559726174909622,
      "grad_norm": 2.359375,
      "learning_rate": 2.4660410737635566e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 2022750.0,
      "step": 4628
    },
    {
      "epoch": 0.35604953465118067,
      "grad_norm": 6.90625,
      "learning_rate": 2.465925698023229e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2023034.0,
      "step": 4629
    },
    {
      "epoch": 0.3561264518113991,
      "grad_norm": 7.875,
      "learning_rate": 2.4658103222829015e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9810725450515747,
      "num_tokens": 2023352.0,
      "step": 4630
    },
    {
      "epoch": 0.3562033689716176,
      "grad_norm": 5.78125,
      "learning_rate": 2.4656949465425738e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2023648.0,
      "step": 4631
    },
    {
      "epoch": 0.356280286131836,
      "grad_norm": 8.6875,
      "learning_rate": 2.465579570802246e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9820359349250793,
      "num_tokens": 2023984.0,
      "step": 4632
    },
    {
      "epoch": 0.35635720329205445,
      "grad_norm": 3.5625,
      "learning_rate": 2.4654641950619183e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2024548.0,
      "step": 4633
    },
    {
      "epoch": 0.3564341204522729,
      "grad_norm": 5.46875,
      "learning_rate": 2.4653488193215906e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2024894.0,
      "step": 4634
    },
    {
      "epoch": 0.35651103761249137,
      "grad_norm": 6.6875,
      "learning_rate": 2.465233443581263e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2025174.0,
      "step": 4635
    },
    {
      "epoch": 0.35658795477270977,
      "grad_norm": 4.53125,
      "learning_rate": 2.4651180678409355e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2025486.0,
      "step": 4636
    },
    {
      "epoch": 0.35666487193292823,
      "grad_norm": 3.453125,
      "learning_rate": 2.4650026921006077e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9907235503196716,
      "num_tokens": 2026026.0,
      "step": 4637
    },
    {
      "epoch": 0.3567417890931467,
      "grad_norm": 6.125,
      "learning_rate": 2.46488731636028e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 2026364.0,
      "step": 4638
    },
    {
      "epoch": 0.35681870625336515,
      "grad_norm": 7.875,
      "learning_rate": 2.4647719406199523e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2026666.0,
      "step": 4639
    },
    {
      "epoch": 0.35689562341358355,
      "grad_norm": 3.046875,
      "learning_rate": 2.4646565648796245e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2027250.0,
      "step": 4640
    },
    {
      "epoch": 0.356972540573802,
      "grad_norm": 2.09375,
      "learning_rate": 2.464541189139297e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2027796.0,
      "step": 4641
    },
    {
      "epoch": 0.35704945773402047,
      "grad_norm": 5.5625,
      "learning_rate": 2.4644258133989695e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2028096.0,
      "step": 4642
    },
    {
      "epoch": 0.35712637489423893,
      "grad_norm": 7.5625,
      "learning_rate": 2.4643104376586417e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 2028452.0,
      "step": 4643
    },
    {
      "epoch": 0.35720329205445733,
      "grad_norm": 6.65625,
      "learning_rate": 2.464195061918314e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2028780.0,
      "step": 4644
    },
    {
      "epoch": 0.3572802092146758,
      "grad_norm": 3.90625,
      "learning_rate": 2.4640796861779862e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2029086.0,
      "step": 4645
    },
    {
      "epoch": 0.35735712637489425,
      "grad_norm": 5.03125,
      "learning_rate": 2.4639643104376585e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2029380.0,
      "step": 4646
    },
    {
      "epoch": 0.3574340435351127,
      "grad_norm": 2.609375,
      "learning_rate": 2.463848934697331e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 2029938.0,
      "step": 4647
    },
    {
      "epoch": 0.3575109606953311,
      "grad_norm": 4.8125,
      "learning_rate": 2.4637335589570034e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2030250.0,
      "step": 4648
    },
    {
      "epoch": 0.3575878778555496,
      "grad_norm": 10.0,
      "learning_rate": 2.4636181832166757e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2030506.0,
      "step": 4649
    },
    {
      "epoch": 0.35766479501576803,
      "grad_norm": 5.25,
      "learning_rate": 2.463502807476348e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2030800.0,
      "step": 4650
    },
    {
      "epoch": 0.35774171217598644,
      "grad_norm": 7.40625,
      "learning_rate": 2.46338743173602e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2031116.0,
      "step": 4651
    },
    {
      "epoch": 0.3578186293362049,
      "grad_norm": 2.765625,
      "learning_rate": 2.463272055995693e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 2031668.0,
      "step": 4652
    },
    {
      "epoch": 0.35789554649642336,
      "grad_norm": 2.375,
      "learning_rate": 2.463156680255365e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2032242.0,
      "step": 4653
    },
    {
      "epoch": 0.3579724636566418,
      "grad_norm": 4.125,
      "learning_rate": 2.4630413045150374e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2032762.0,
      "step": 4654
    },
    {
      "epoch": 0.3580493808168602,
      "grad_norm": 4.09375,
      "learning_rate": 2.4629259287747097e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2033278.0,
      "step": 4655
    },
    {
      "epoch": 0.3581262979770787,
      "grad_norm": 7.15625,
      "learning_rate": 2.462810553034382e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2033634.0,
      "step": 4656
    },
    {
      "epoch": 0.35820321513729714,
      "grad_norm": 3.875,
      "learning_rate": 2.462695177294054e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2034198.0,
      "step": 4657
    },
    {
      "epoch": 0.3582801322975156,
      "grad_norm": 3.53125,
      "learning_rate": 2.462579801553727e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2034718.0,
      "step": 4658
    },
    {
      "epoch": 0.358357049457734,
      "grad_norm": 7.84375,
      "learning_rate": 2.462464425813399e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2034974.0,
      "step": 4659
    },
    {
      "epoch": 0.35843396661795246,
      "grad_norm": 1.859375,
      "learning_rate": 2.4623490500730714e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2035522.0,
      "step": 4660
    },
    {
      "epoch": 0.3585108837781709,
      "grad_norm": 6.375,
      "learning_rate": 2.4622336743327436e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 2036134.0,
      "step": 4661
    },
    {
      "epoch": 0.3585878009383894,
      "grad_norm": 5.40625,
      "learning_rate": 2.462118298592416e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2036462.0,
      "step": 4662
    },
    {
      "epoch": 0.3586647180986078,
      "grad_norm": 3.03125,
      "learning_rate": 2.4620029228520886e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9961904883384705,
      "num_tokens": 2036988.0,
      "step": 4663
    },
    {
      "epoch": 0.35874163525882624,
      "grad_norm": 5.46875,
      "learning_rate": 2.461887547111761e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2037272.0,
      "step": 4664
    },
    {
      "epoch": 0.3588185524190447,
      "grad_norm": 5.46875,
      "learning_rate": 2.461772171371433e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 2037622.0,
      "step": 4665
    },
    {
      "epoch": 0.35889546957926316,
      "grad_norm": 2.40625,
      "learning_rate": 2.4616567956311054e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2038180.0,
      "step": 4666
    },
    {
      "epoch": 0.35897238673948156,
      "grad_norm": 6.25,
      "learning_rate": 2.4615414198907776e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 2038792.0,
      "step": 4667
    },
    {
      "epoch": 0.3590493038997,
      "grad_norm": 2.3125,
      "learning_rate": 2.46142604415045e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2039344.0,
      "step": 4668
    },
    {
      "epoch": 0.3591262210599185,
      "grad_norm": 2.921875,
      "learning_rate": 2.4613106684101225e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 2039894.0,
      "step": 4669
    },
    {
      "epoch": 0.3592031382201369,
      "grad_norm": 7.0,
      "learning_rate": 2.461195292669795e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2040226.0,
      "step": 4670
    },
    {
      "epoch": 0.35928005538035535,
      "grad_norm": 2.640625,
      "learning_rate": 2.461079916929467e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2040782.0,
      "step": 4671
    },
    {
      "epoch": 0.3593569725405738,
      "grad_norm": 1.8984375,
      "learning_rate": 2.4609645411891393e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9960317611694336,
      "num_tokens": 2041540.0,
      "step": 4672
    },
    {
      "epoch": 0.35943388970079226,
      "grad_norm": 3.359375,
      "learning_rate": 2.4608491654488116e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 2042112.0,
      "step": 4673
    },
    {
      "epoch": 0.35951080686101067,
      "grad_norm": 2.203125,
      "learning_rate": 2.4607337897084843e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2042668.0,
      "step": 4674
    },
    {
      "epoch": 0.3595877240212291,
      "grad_norm": 4.25,
      "learning_rate": 2.4606184139681565e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2042994.0,
      "step": 4675
    },
    {
      "epoch": 0.3596646411814476,
      "grad_norm": 8.4375,
      "learning_rate": 2.4605030382278288e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 2043352.0,
      "step": 4676
    },
    {
      "epoch": 0.35974155834166605,
      "grad_norm": 5.65625,
      "learning_rate": 2.460387662487501e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2043654.0,
      "step": 4677
    },
    {
      "epoch": 0.35981847550188445,
      "grad_norm": 5.09375,
      "learning_rate": 2.4602722867471733e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2043948.0,
      "step": 4678
    },
    {
      "epoch": 0.3598953926621029,
      "grad_norm": 5.15625,
      "learning_rate": 2.4601569110068456e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2044258.0,
      "step": 4679
    },
    {
      "epoch": 0.35997230982232137,
      "grad_norm": 6.1875,
      "learning_rate": 2.4600415352665182e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 2044572.0,
      "step": 4680
    },
    {
      "epoch": 0.3600492269825398,
      "grad_norm": 2.1875,
      "learning_rate": 2.4599261595261905e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9987863898277283,
      "num_tokens": 2045398.0,
      "step": 4681
    },
    {
      "epoch": 0.36012614414275823,
      "grad_norm": 5.875,
      "learning_rate": 2.4598107837858628e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 2045692.0,
      "step": 4682
    },
    {
      "epoch": 0.3602030613029767,
      "grad_norm": 3.5,
      "learning_rate": 2.459695408045535e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2046240.0,
      "step": 4683
    },
    {
      "epoch": 0.36027997846319515,
      "grad_norm": 5.6875,
      "learning_rate": 2.4595800323052073e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2046550.0,
      "step": 4684
    },
    {
      "epoch": 0.3603568956234136,
      "grad_norm": 6.21875,
      "learning_rate": 2.4594646565648795e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2046906.0,
      "step": 4685
    },
    {
      "epoch": 0.360433812783632,
      "grad_norm": 5.8125,
      "learning_rate": 2.459349280824552e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2047218.0,
      "step": 4686
    },
    {
      "epoch": 0.36051072994385047,
      "grad_norm": 3.296875,
      "learning_rate": 2.459233905084224e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2047540.0,
      "step": 4687
    },
    {
      "epoch": 0.36058764710406893,
      "grad_norm": 3.109375,
      "learning_rate": 2.4591185293438967e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2048126.0,
      "step": 4688
    },
    {
      "epoch": 0.36066456426428734,
      "grad_norm": 6.96875,
      "learning_rate": 2.459003153603569e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2048412.0,
      "step": 4689
    },
    {
      "epoch": 0.3607414814245058,
      "grad_norm": 9.25,
      "learning_rate": 2.4588877778632412e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2048672.0,
      "step": 4690
    },
    {
      "epoch": 0.36081839858472425,
      "grad_norm": 3.265625,
      "learning_rate": 2.458772402122914e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2049224.0,
      "step": 4691
    },
    {
      "epoch": 0.3608953157449427,
      "grad_norm": 4.6875,
      "learning_rate": 2.458657026382586e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 2049568.0,
      "step": 4692
    },
    {
      "epoch": 0.3609722329051611,
      "grad_norm": 8.4375,
      "learning_rate": 2.458541650642258e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 2049940.0,
      "step": 4693
    },
    {
      "epoch": 0.3610491500653796,
      "grad_norm": 8.3125,
      "learning_rate": 2.4584262749019307e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9804560542106628,
      "num_tokens": 2050248.0,
      "step": 4694
    },
    {
      "epoch": 0.36112606722559804,
      "grad_norm": 9.375,
      "learning_rate": 2.458310899161603e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 2050548.0,
      "step": 4695
    },
    {
      "epoch": 0.3612029843858165,
      "grad_norm": 5.21875,
      "learning_rate": 2.4581955234212752e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2050854.0,
      "step": 4696
    },
    {
      "epoch": 0.3612799015460349,
      "grad_norm": 9.0625,
      "learning_rate": 2.458080147680948e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 2051154.0,
      "step": 4697
    },
    {
      "epoch": 0.36135681870625336,
      "grad_norm": 5.28125,
      "learning_rate": 2.4579647719406197e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 2051484.0,
      "step": 4698
    },
    {
      "epoch": 0.3614337358664718,
      "grad_norm": 6.5625,
      "learning_rate": 2.4578493962002924e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9858757257461548,
      "num_tokens": 2051840.0,
      "step": 4699
    },
    {
      "epoch": 0.3615106530266903,
      "grad_norm": 5.375,
      "learning_rate": 2.4577340204599647e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2052168.0,
      "step": 4700
    },
    {
      "epoch": 0.3615875701869087,
      "grad_norm": 2.140625,
      "learning_rate": 2.457618644719637e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2052976.0,
      "step": 4701
    },
    {
      "epoch": 0.36166448734712714,
      "grad_norm": 6.53125,
      "learning_rate": 2.4575032689793096e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2053284.0,
      "step": 4702
    },
    {
      "epoch": 0.3617414045073456,
      "grad_norm": 5.875,
      "learning_rate": 2.457387893238982e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2053566.0,
      "step": 4703
    },
    {
      "epoch": 0.36181832166756406,
      "grad_norm": 1.9609375,
      "learning_rate": 2.4572725174986537e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2054136.0,
      "step": 4704
    },
    {
      "epoch": 0.36189523882778246,
      "grad_norm": 7.75,
      "learning_rate": 2.4571571417583264e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 2054500.0,
      "step": 4705
    },
    {
      "epoch": 0.3619721559880009,
      "grad_norm": 4.15625,
      "learning_rate": 2.4570417660179986e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.990439772605896,
      "num_tokens": 2055024.0,
      "step": 4706
    },
    {
      "epoch": 0.3620490731482194,
      "grad_norm": 2.53125,
      "learning_rate": 2.456926390277671e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2055604.0,
      "step": 4707
    },
    {
      "epoch": 0.36212599030843784,
      "grad_norm": 2.84375,
      "learning_rate": 2.4568110145373436e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2056152.0,
      "step": 4708
    },
    {
      "epoch": 0.36220290746865624,
      "grad_norm": 4.8125,
      "learning_rate": 2.4566956387970154e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2056446.0,
      "step": 4709
    },
    {
      "epoch": 0.3622798246288747,
      "grad_norm": 3.09375,
      "learning_rate": 2.456580263056688e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2057032.0,
      "step": 4710
    },
    {
      "epoch": 0.36235674178909316,
      "grad_norm": 2.671875,
      "learning_rate": 2.4564648873163604e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2057594.0,
      "step": 4711
    },
    {
      "epoch": 0.36243365894931157,
      "grad_norm": 2.65625,
      "learning_rate": 2.4563495115760326e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2058420.0,
      "step": 4712
    },
    {
      "epoch": 0.36251057610953,
      "grad_norm": 3.53125,
      "learning_rate": 2.4562341358357053e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 2058998.0,
      "step": 4713
    },
    {
      "epoch": 0.3625874932697485,
      "grad_norm": 3.40625,
      "learning_rate": 2.4561187600953776e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2059544.0,
      "step": 4714
    },
    {
      "epoch": 0.36266441042996694,
      "grad_norm": 2.515625,
      "learning_rate": 2.4560033843550494e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2060302.0,
      "step": 4715
    },
    {
      "epoch": 0.36274132759018535,
      "grad_norm": 7.71875,
      "learning_rate": 2.455888008614722e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2060608.0,
      "step": 4716
    },
    {
      "epoch": 0.3628182447504038,
      "grad_norm": 6.3125,
      "learning_rate": 2.4557726328743943e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2060892.0,
      "step": 4717
    },
    {
      "epoch": 0.36289516191062227,
      "grad_norm": 2.421875,
      "learning_rate": 2.4556572571340666e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2061482.0,
      "step": 4718
    },
    {
      "epoch": 0.3629720790708407,
      "grad_norm": 3.84375,
      "learning_rate": 2.4555418813937393e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2062046.0,
      "step": 4719
    },
    {
      "epoch": 0.36304899623105913,
      "grad_norm": 1.6953125,
      "learning_rate": 2.455426505653411e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2062856.0,
      "step": 4720
    },
    {
      "epoch": 0.3631259133912776,
      "grad_norm": 4.875,
      "learning_rate": 2.4553111299130838e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2063176.0,
      "step": 4721
    },
    {
      "epoch": 0.36320283055149605,
      "grad_norm": 1.6953125,
      "learning_rate": 2.455195754172756e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2063738.0,
      "step": 4722
    },
    {
      "epoch": 0.3632797477117145,
      "grad_norm": 4.375,
      "learning_rate": 2.4550803784324283e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2064050.0,
      "step": 4723
    },
    {
      "epoch": 0.3633566648719329,
      "grad_norm": 9.5,
      "learning_rate": 2.454965002692101e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2064306.0,
      "step": 4724
    },
    {
      "epoch": 0.36343358203215137,
      "grad_norm": 1.71875,
      "learning_rate": 2.454849626951773e-06,
      "loss": 0.0069,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 2065326.0,
      "step": 4725
    },
    {
      "epoch": 0.36351049919236983,
      "grad_norm": 4.46875,
      "learning_rate": 2.454734251211445e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 2065918.0,
      "step": 4726
    },
    {
      "epoch": 0.3635874163525883,
      "grad_norm": 5.8125,
      "learning_rate": 2.4546188754711178e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2066202.0,
      "step": 4727
    },
    {
      "epoch": 0.3636643335128067,
      "grad_norm": 1.7578125,
      "learning_rate": 2.45450349973079e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 2067010.0,
      "step": 4728
    },
    {
      "epoch": 0.36374125067302515,
      "grad_norm": 5.125,
      "learning_rate": 2.4543881239904623e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2067326.0,
      "step": 4729
    },
    {
      "epoch": 0.3638181678332436,
      "grad_norm": 5.3125,
      "learning_rate": 2.454272748250135e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2067626.0,
      "step": 4730
    },
    {
      "epoch": 0.363895084993462,
      "grad_norm": 2.859375,
      "learning_rate": 2.4541573725098068e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2068206.0,
      "step": 4731
    },
    {
      "epoch": 0.3639720021536805,
      "grad_norm": 5.15625,
      "learning_rate": 2.4540419967694795e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2068506.0,
      "step": 4732
    },
    {
      "epoch": 0.36404891931389893,
      "grad_norm": 3.171875,
      "learning_rate": 2.4539266210291517e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2069070.0,
      "step": 4733
    },
    {
      "epoch": 0.3641258364741174,
      "grad_norm": 1.5625,
      "learning_rate": 2.453811245288824e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9973545074462891,
      "num_tokens": 2069828.0,
      "step": 4734
    },
    {
      "epoch": 0.3642027536343358,
      "grad_norm": 4.875,
      "learning_rate": 2.4536958695484962e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2070152.0,
      "step": 4735
    },
    {
      "epoch": 0.36427967079455426,
      "grad_norm": 4.71875,
      "learning_rate": 2.4535804938081685e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2070478.0,
      "step": 4736
    },
    {
      "epoch": 0.3643565879547727,
      "grad_norm": 2.078125,
      "learning_rate": 2.4534651180678408e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 2071038.0,
      "step": 4737
    },
    {
      "epoch": 0.3644335051149912,
      "grad_norm": 1.78125,
      "learning_rate": 2.4533497423275134e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 2071834.0,
      "step": 4738
    },
    {
      "epoch": 0.3645104222752096,
      "grad_norm": 4.8125,
      "learning_rate": 2.4532343665871857e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2072128.0,
      "step": 4739
    },
    {
      "epoch": 0.36458733943542804,
      "grad_norm": 2.109375,
      "learning_rate": 2.453118990846858e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2072674.0,
      "step": 4740
    },
    {
      "epoch": 0.3646642565956465,
      "grad_norm": 5.28125,
      "learning_rate": 2.4530036151065306e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2073008.0,
      "step": 4741
    },
    {
      "epoch": 0.36474117375586496,
      "grad_norm": 5.21875,
      "learning_rate": 2.4528882393662025e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2073354.0,
      "step": 4742
    },
    {
      "epoch": 0.36481809091608336,
      "grad_norm": 5.84375,
      "learning_rate": 2.4527728636258747e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 2073688.0,
      "step": 4743
    },
    {
      "epoch": 0.3648950080763018,
      "grad_norm": 3.109375,
      "learning_rate": 2.4526574878855474e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2074230.0,
      "step": 4744
    },
    {
      "epoch": 0.3649719252365203,
      "grad_norm": 6.125,
      "learning_rate": 2.4525421121452197e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2074544.0,
      "step": 4745
    },
    {
      "epoch": 0.36504884239673874,
      "grad_norm": 6.28125,
      "learning_rate": 2.452426736404892e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2074896.0,
      "step": 4746
    },
    {
      "epoch": 0.36512575955695714,
      "grad_norm": 2.109375,
      "learning_rate": 2.452311360664564e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2075480.0,
      "step": 4747
    },
    {
      "epoch": 0.3652026767171756,
      "grad_norm": 12.125,
      "learning_rate": 2.4521959849242364e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 2075732.0,
      "step": 4748
    },
    {
      "epoch": 0.36527959387739406,
      "grad_norm": 4.40625,
      "learning_rate": 2.452080609183909e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2076092.0,
      "step": 4749
    },
    {
      "epoch": 0.36535651103761246,
      "grad_norm": 4.5,
      "learning_rate": 2.4519652334435814e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.990439772605896,
      "num_tokens": 2076616.0,
      "step": 4750
    },
    {
      "epoch": 0.3654334281978309,
      "grad_norm": 3.875,
      "learning_rate": 2.4518498577032536e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 2077188.0,
      "step": 4751
    },
    {
      "epoch": 0.3655103453580494,
      "grad_norm": 8.3125,
      "learning_rate": 2.4517344819629263e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2077488.0,
      "step": 4752
    },
    {
      "epoch": 0.36558726251826784,
      "grad_norm": 8.125,
      "learning_rate": 2.451619106222598e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2077748.0,
      "step": 4753
    },
    {
      "epoch": 0.36566417967848625,
      "grad_norm": 2.53125,
      "learning_rate": 2.4515037304822704e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2078280.0,
      "step": 4754
    },
    {
      "epoch": 0.3657410968387047,
      "grad_norm": 3.296875,
      "learning_rate": 2.451388354741943e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2078848.0,
      "step": 4755
    },
    {
      "epoch": 0.36581801399892316,
      "grad_norm": 5.625,
      "learning_rate": 2.4512729790016154e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 2079176.0,
      "step": 4756
    },
    {
      "epoch": 0.3658949311591416,
      "grad_norm": 6.375,
      "learning_rate": 2.4511576032612876e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2079522.0,
      "step": 4757
    },
    {
      "epoch": 0.36597184831936,
      "grad_norm": 4.1875,
      "learning_rate": 2.45104222752096e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2080074.0,
      "step": 4758
    },
    {
      "epoch": 0.3660487654795785,
      "grad_norm": 2.5625,
      "learning_rate": 2.450926851780632e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2080622.0,
      "step": 4759
    },
    {
      "epoch": 0.36612568263979695,
      "grad_norm": 3.59375,
      "learning_rate": 2.450811476040305e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2081148.0,
      "step": 4760
    },
    {
      "epoch": 0.3662025998000154,
      "grad_norm": 2.1875,
      "learning_rate": 2.450696100299977e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2081692.0,
      "step": 4761
    },
    {
      "epoch": 0.3662795169602338,
      "grad_norm": 4.25,
      "learning_rate": 2.4505807245596493e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 2082292.0,
      "step": 4762
    },
    {
      "epoch": 0.36635643412045227,
      "grad_norm": 3.5,
      "learning_rate": 2.4504653488193216e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2082856.0,
      "step": 4763
    },
    {
      "epoch": 0.3664333512806707,
      "grad_norm": 1.90625,
      "learning_rate": 2.450349973078994e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2083402.0,
      "step": 4764
    },
    {
      "epoch": 0.3665102684408892,
      "grad_norm": 7.71875,
      "learning_rate": 2.450234597338666e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2083686.0,
      "step": 4765
    },
    {
      "epoch": 0.3665871856011076,
      "grad_norm": 6.34375,
      "learning_rate": 2.4501192215983388e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2083994.0,
      "step": 4766
    },
    {
      "epoch": 0.36666410276132605,
      "grad_norm": 3.84375,
      "learning_rate": 2.450003845858011e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2084574.0,
      "step": 4767
    },
    {
      "epoch": 0.3667410199215445,
      "grad_norm": 1.9765625,
      "learning_rate": 2.4498884701176833e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 2085358.0,
      "step": 4768
    },
    {
      "epoch": 0.36681793708176297,
      "grad_norm": 2.4375,
      "learning_rate": 2.4497730943773556e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 2086418.0,
      "step": 4769
    },
    {
      "epoch": 0.36689485424198137,
      "grad_norm": 8.1875,
      "learning_rate": 2.449657718637028e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2086748.0,
      "step": 4770
    },
    {
      "epoch": 0.36697177140219983,
      "grad_norm": 6.0,
      "learning_rate": 2.4495423428967005e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 2087066.0,
      "step": 4771
    },
    {
      "epoch": 0.3670486885624183,
      "grad_norm": 4.25,
      "learning_rate": 2.4494269671563728e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 2087392.0,
      "step": 4772
    },
    {
      "epoch": 0.3671256057226367,
      "grad_norm": 6.59375,
      "learning_rate": 2.449311591416045e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2087698.0,
      "step": 4773
    },
    {
      "epoch": 0.36720252288285515,
      "grad_norm": 2.640625,
      "learning_rate": 2.4491962156757173e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 2088258.0,
      "step": 4774
    },
    {
      "epoch": 0.3672794400430736,
      "grad_norm": 9.6875,
      "learning_rate": 2.4490808399353895e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 2088562.0,
      "step": 4775
    },
    {
      "epoch": 0.36735635720329207,
      "grad_norm": 8.0625,
      "learning_rate": 2.448965464195062e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2088854.0,
      "step": 4776
    },
    {
      "epoch": 0.3674332743635105,
      "grad_norm": 2.625,
      "learning_rate": 2.4488500884547345e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2089440.0,
      "step": 4777
    },
    {
      "epoch": 0.36751019152372894,
      "grad_norm": 9.125,
      "learning_rate": 2.4487347127144067e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 2089746.0,
      "step": 4778
    },
    {
      "epoch": 0.3675871086839474,
      "grad_norm": 8.8125,
      "learning_rate": 2.448619336974079e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2090028.0,
      "step": 4779
    },
    {
      "epoch": 0.36766402584416585,
      "grad_norm": 4.5,
      "learning_rate": 2.4485039612337512e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2090548.0,
      "step": 4780
    },
    {
      "epoch": 0.36774094300438426,
      "grad_norm": 3.390625,
      "learning_rate": 2.4483885854934235e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2091066.0,
      "step": 4781
    },
    {
      "epoch": 0.3678178601646027,
      "grad_norm": 7.625,
      "learning_rate": 2.448273209753096e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2091344.0,
      "step": 4782
    },
    {
      "epoch": 0.3678947773248212,
      "grad_norm": 2.9375,
      "learning_rate": 2.4481578340127684e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 2091864.0,
      "step": 4783
    },
    {
      "epoch": 0.36797169448503964,
      "grad_norm": 3.734375,
      "learning_rate": 2.4480424582724407e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2092450.0,
      "step": 4784
    },
    {
      "epoch": 0.36804861164525804,
      "grad_norm": 5.71875,
      "learning_rate": 2.447927082532113e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2092794.0,
      "step": 4785
    },
    {
      "epoch": 0.3681255288054765,
      "grad_norm": 5.6875,
      "learning_rate": 2.4478117067917852e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2093144.0,
      "step": 4786
    },
    {
      "epoch": 0.36820244596569496,
      "grad_norm": 2.328125,
      "learning_rate": 2.4476963310514575e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2093714.0,
      "step": 4787
    },
    {
      "epoch": 0.3682793631259134,
      "grad_norm": 6.125,
      "learning_rate": 2.44758095531113e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2093994.0,
      "step": 4788
    },
    {
      "epoch": 0.3683562802861318,
      "grad_norm": 2.140625,
      "learning_rate": 2.4474655795708024e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2094540.0,
      "step": 4789
    },
    {
      "epoch": 0.3684331974463503,
      "grad_norm": 3.703125,
      "learning_rate": 2.4473502038304747e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2095076.0,
      "step": 4790
    },
    {
      "epoch": 0.36851011460656874,
      "grad_norm": 6.0,
      "learning_rate": 2.447234828090147e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2095410.0,
      "step": 4791
    },
    {
      "epoch": 0.36858703176678714,
      "grad_norm": 8.1875,
      "learning_rate": 2.447119452349819e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2095762.0,
      "step": 4792
    },
    {
      "epoch": 0.3686639489270056,
      "grad_norm": 6.03125,
      "learning_rate": 2.4470040766094914e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2096098.0,
      "step": 4793
    },
    {
      "epoch": 0.36874086608722406,
      "grad_norm": 4.53125,
      "learning_rate": 2.446888700869164e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2096380.0,
      "step": 4794
    },
    {
      "epoch": 0.3688177832474425,
      "grad_norm": 2.84375,
      "learning_rate": 2.4467733251288364e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 2096930.0,
      "step": 4795
    },
    {
      "epoch": 0.3688947004076609,
      "grad_norm": 2.109375,
      "learning_rate": 2.4466579493885086e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2097504.0,
      "step": 4796
    },
    {
      "epoch": 0.3689716175678794,
      "grad_norm": 3.640625,
      "learning_rate": 2.446542573648181e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2098084.0,
      "step": 4797
    },
    {
      "epoch": 0.36904853472809784,
      "grad_norm": 2.28125,
      "learning_rate": 2.446427197907853e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2098632.0,
      "step": 4798
    },
    {
      "epoch": 0.3691254518883163,
      "grad_norm": 2.8125,
      "learning_rate": 2.446311822167526e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 2099196.0,
      "step": 4799
    },
    {
      "epoch": 0.3692023690485347,
      "grad_norm": 7.34375,
      "learning_rate": 2.446196446427198e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2099456.0,
      "step": 4800
    },
    {
      "epoch": 0.36927928620875317,
      "grad_norm": 2.578125,
      "learning_rate": 2.44608107068687e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2100040.0,
      "step": 4801
    },
    {
      "epoch": 0.3693562033689716,
      "grad_norm": 5.90625,
      "learning_rate": 2.4459656949465426e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2100364.0,
      "step": 4802
    },
    {
      "epoch": 0.3694331205291901,
      "grad_norm": 7.0625,
      "learning_rate": 2.445850319206215e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2100696.0,
      "step": 4803
    },
    {
      "epoch": 0.3695100376894085,
      "grad_norm": 2.625,
      "learning_rate": 2.445734943465887e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2101230.0,
      "step": 4804
    },
    {
      "epoch": 0.36958695484962695,
      "grad_norm": 1.8828125,
      "learning_rate": 2.44561956772556e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2102040.0,
      "step": 4805
    },
    {
      "epoch": 0.3696638720098454,
      "grad_norm": 5.9375,
      "learning_rate": 2.445504191985232e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2102322.0,
      "step": 4806
    },
    {
      "epoch": 0.36974078917006387,
      "grad_norm": 1.9375,
      "learning_rate": 2.4453888162449043e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 2103106.0,
      "step": 4807
    },
    {
      "epoch": 0.36981770633028227,
      "grad_norm": 3.5,
      "learning_rate": 2.4452734405045766e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2103622.0,
      "step": 4808
    },
    {
      "epoch": 0.36989462349050073,
      "grad_norm": 6.90625,
      "learning_rate": 2.445158064764249e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2103984.0,
      "step": 4809
    },
    {
      "epoch": 0.3699715406507192,
      "grad_norm": 2.890625,
      "learning_rate": 2.4450426890239215e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 2104534.0,
      "step": 4810
    },
    {
      "epoch": 0.37004845781093765,
      "grad_norm": 2.109375,
      "learning_rate": 2.4449273132835938e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 2105314.0,
      "step": 4811
    },
    {
      "epoch": 0.37012537497115605,
      "grad_norm": 6.625,
      "learning_rate": 2.4448119375432656e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 2105622.0,
      "step": 4812
    },
    {
      "epoch": 0.3702022921313745,
      "grad_norm": 5.65625,
      "learning_rate": 2.4446965618029383e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2105982.0,
      "step": 4813
    },
    {
      "epoch": 0.37027920929159297,
      "grad_norm": 5.75,
      "learning_rate": 2.4445811860626106e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2106318.0,
      "step": 4814
    },
    {
      "epoch": 0.3703561264518114,
      "grad_norm": 8.875,
      "learning_rate": 2.444465810322283e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2106576.0,
      "step": 4815
    },
    {
      "epoch": 0.37043304361202983,
      "grad_norm": 2.140625,
      "learning_rate": 2.4443504345819555e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2107110.0,
      "step": 4816
    },
    {
      "epoch": 0.3705099607722483,
      "grad_norm": 3.859375,
      "learning_rate": 2.4442350588416278e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2107678.0,
      "step": 4817
    },
    {
      "epoch": 0.37058687793246675,
      "grad_norm": 2.25,
      "learning_rate": 2.4441196831013e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 2108236.0,
      "step": 4818
    },
    {
      "epoch": 0.37066379509268516,
      "grad_norm": 3.59375,
      "learning_rate": 2.4440043073609723e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2108806.0,
      "step": 4819
    },
    {
      "epoch": 0.3707407122529036,
      "grad_norm": 6.78125,
      "learning_rate": 2.4438889316206445e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 2109078.0,
      "step": 4820
    },
    {
      "epoch": 0.3708176294131221,
      "grad_norm": 6.0,
      "learning_rate": 2.4437735558803172e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2109394.0,
      "step": 4821
    },
    {
      "epoch": 0.37089454657334053,
      "grad_norm": 7.46875,
      "learning_rate": 2.4436581801399895e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2109702.0,
      "step": 4822
    },
    {
      "epoch": 0.37097146373355894,
      "grad_norm": 1.921875,
      "learning_rate": 2.4435428043996613e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2110522.0,
      "step": 4823
    },
    {
      "epoch": 0.3710483808937774,
      "grad_norm": 5.5625,
      "learning_rate": 2.443427428659334e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2110862.0,
      "step": 4824
    },
    {
      "epoch": 0.37112529805399586,
      "grad_norm": 3.1875,
      "learning_rate": 2.4433120529190062e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 2111410.0,
      "step": 4825
    },
    {
      "epoch": 0.3712022152142143,
      "grad_norm": 3.40625,
      "learning_rate": 2.4431966771786785e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 2111950.0,
      "step": 4826
    },
    {
      "epoch": 0.3712791323744327,
      "grad_norm": 4.65625,
      "learning_rate": 2.443081301438351e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2112278.0,
      "step": 4827
    },
    {
      "epoch": 0.3713560495346512,
      "grad_norm": 5.78125,
      "learning_rate": 2.442965925698023e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2112594.0,
      "step": 4828
    },
    {
      "epoch": 0.37143296669486964,
      "grad_norm": 5.75,
      "learning_rate": 2.4428505499576957e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2112934.0,
      "step": 4829
    },
    {
      "epoch": 0.3715098838550881,
      "grad_norm": 8.625,
      "learning_rate": 2.442735174217368e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2113252.0,
      "step": 4830
    },
    {
      "epoch": 0.3715868010153065,
      "grad_norm": 6.25,
      "learning_rate": 2.4426197984770402e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 2113566.0,
      "step": 4831
    },
    {
      "epoch": 0.37166371817552496,
      "grad_norm": 4.0625,
      "learning_rate": 2.442504422736713e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 2113912.0,
      "step": 4832
    },
    {
      "epoch": 0.3717406353357434,
      "grad_norm": 3.609375,
      "learning_rate": 2.442389046996385e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2114234.0,
      "step": 4833
    },
    {
      "epoch": 0.3718175524959618,
      "grad_norm": 7.5625,
      "learning_rate": 2.442273671256057e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 2114542.0,
      "step": 4834
    },
    {
      "epoch": 0.3718944696561803,
      "grad_norm": 5.3125,
      "learning_rate": 2.4421582955157297e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.993630588054657,
      "num_tokens": 2114858.0,
      "step": 4835
    },
    {
      "epoch": 0.37197138681639874,
      "grad_norm": 5.625,
      "learning_rate": 2.442042919775402e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 2115220.0,
      "step": 4836
    },
    {
      "epoch": 0.3720483039766172,
      "grad_norm": 2.296875,
      "learning_rate": 2.441927544035074e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2115766.0,
      "step": 4837
    },
    {
      "epoch": 0.3721252211368356,
      "grad_norm": 6.1875,
      "learning_rate": 2.441812168294747e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9858356714248657,
      "num_tokens": 2116120.0,
      "step": 4838
    },
    {
      "epoch": 0.37220213829705406,
      "grad_norm": 2.703125,
      "learning_rate": 2.4416967925544187e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2116442.0,
      "step": 4839
    },
    {
      "epoch": 0.3722790554572725,
      "grad_norm": 3.015625,
      "learning_rate": 2.4415814168140914e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2117200.0,
      "step": 4840
    },
    {
      "epoch": 0.372355972617491,
      "grad_norm": 3.828125,
      "learning_rate": 2.4414660410737636e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 2117552.0,
      "step": 4841
    },
    {
      "epoch": 0.3724328897777094,
      "grad_norm": 4.8125,
      "learning_rate": 2.441350665333436e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 2118124.0,
      "step": 4842
    },
    {
      "epoch": 0.37250980693792785,
      "grad_norm": 5.5,
      "learning_rate": 2.4412352895931086e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2118434.0,
      "step": 4843
    },
    {
      "epoch": 0.3725867240981463,
      "grad_norm": 2.984375,
      "learning_rate": 2.441119913852781e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 2119018.0,
      "step": 4844
    },
    {
      "epoch": 0.37266364125836476,
      "grad_norm": 7.5625,
      "learning_rate": 2.4410045381124527e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2119326.0,
      "step": 4845
    },
    {
      "epoch": 0.37274055841858317,
      "grad_norm": 2.734375,
      "learning_rate": 2.4408891623721254e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2119896.0,
      "step": 4846
    },
    {
      "epoch": 0.3728174755788016,
      "grad_norm": 6.625,
      "learning_rate": 2.4407737866317976e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2120230.0,
      "step": 4847
    },
    {
      "epoch": 0.3728943927390201,
      "grad_norm": 8.5,
      "learning_rate": 2.44065841089147e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2120578.0,
      "step": 4848
    },
    {
      "epoch": 0.37297130989923855,
      "grad_norm": 12.625,
      "learning_rate": 2.4405430351511426e-06,
      "loss": 0.0451,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2120836.0,
      "step": 4849
    },
    {
      "epoch": 0.37304822705945695,
      "grad_norm": 3.109375,
      "learning_rate": 2.4404276594108144e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2121390.0,
      "step": 4850
    },
    {
      "epoch": 0.3731251442196754,
      "grad_norm": 2.9375,
      "learning_rate": 2.440312283670487e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 2121712.0,
      "step": 4851
    },
    {
      "epoch": 0.37320206137989387,
      "grad_norm": 4.8125,
      "learning_rate": 2.4401969079301593e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2122030.0,
      "step": 4852
    },
    {
      "epoch": 0.37327897854011227,
      "grad_norm": 4.375,
      "learning_rate": 2.4400815321898316e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 2122374.0,
      "step": 4853
    },
    {
      "epoch": 0.37335589570033073,
      "grad_norm": 4.75,
      "learning_rate": 2.439966156449504e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2122696.0,
      "step": 4854
    },
    {
      "epoch": 0.3734328128605492,
      "grad_norm": 7.125,
      "learning_rate": 2.4398507807091765e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 2123024.0,
      "step": 4855
    },
    {
      "epoch": 0.37350973002076765,
      "grad_norm": 3.484375,
      "learning_rate": 2.4397354049688484e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2123570.0,
      "step": 4856
    },
    {
      "epoch": 0.37358664718098605,
      "grad_norm": 3.65625,
      "learning_rate": 2.439620029228521e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2123888.0,
      "step": 4857
    },
    {
      "epoch": 0.3736635643412045,
      "grad_norm": 3.453125,
      "learning_rate": 2.4395046534881933e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2124206.0,
      "step": 4858
    },
    {
      "epoch": 0.37374048150142297,
      "grad_norm": 2.40625,
      "learning_rate": 2.4393892777478656e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2124748.0,
      "step": 4859
    },
    {
      "epoch": 0.37381739866164143,
      "grad_norm": 5.53125,
      "learning_rate": 2.4392739020075382e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2125098.0,
      "step": 4860
    },
    {
      "epoch": 0.37389431582185984,
      "grad_norm": 2.59375,
      "learning_rate": 2.43915852626721e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2125918.0,
      "step": 4861
    },
    {
      "epoch": 0.3739712329820783,
      "grad_norm": 3.25,
      "learning_rate": 2.4390431505268823e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2126462.0,
      "step": 4862
    },
    {
      "epoch": 0.37404815014229675,
      "grad_norm": 2.78125,
      "learning_rate": 2.438927774786555e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2127008.0,
      "step": 4863
    },
    {
      "epoch": 0.3741250673025152,
      "grad_norm": 7.15625,
      "learning_rate": 2.4388123990462273e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 2127342.0,
      "step": 4864
    },
    {
      "epoch": 0.3742019844627336,
      "grad_norm": 6.875,
      "learning_rate": 2.4386970233058995e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9937499761581421,
      "num_tokens": 2127666.0,
      "step": 4865
    },
    {
      "epoch": 0.3742789016229521,
      "grad_norm": 3.203125,
      "learning_rate": 2.438581647565572e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2128228.0,
      "step": 4866
    },
    {
      "epoch": 0.37435581878317054,
      "grad_norm": 6.875,
      "learning_rate": 2.438466271825244e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 2128576.0,
      "step": 4867
    },
    {
      "epoch": 0.374432735943389,
      "grad_norm": 5.78125,
      "learning_rate": 2.4383508960849167e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2128880.0,
      "step": 4868
    },
    {
      "epoch": 0.3745096531036074,
      "grad_norm": 4.34375,
      "learning_rate": 2.438235520344589e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 2129374.0,
      "step": 4869
    },
    {
      "epoch": 0.37458657026382586,
      "grad_norm": 2.640625,
      "learning_rate": 2.4381201446042613e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2129958.0,
      "step": 4870
    },
    {
      "epoch": 0.3746634874240443,
      "grad_norm": 4.21875,
      "learning_rate": 2.438004768863934e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2130520.0,
      "step": 4871
    },
    {
      "epoch": 0.3747404045842628,
      "grad_norm": 3.765625,
      "learning_rate": 2.4378893931236058e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2131110.0,
      "step": 4872
    },
    {
      "epoch": 0.3748173217444812,
      "grad_norm": 6.5,
      "learning_rate": 2.437774017383278e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 2131470.0,
      "step": 4873
    },
    {
      "epoch": 0.37489423890469964,
      "grad_norm": 1.8671875,
      "learning_rate": 2.4376586416429507e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2132054.0,
      "step": 4874
    },
    {
      "epoch": 0.3749711560649181,
      "grad_norm": 5.875,
      "learning_rate": 2.437543265902623e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 2132546.0,
      "step": 4875
    },
    {
      "epoch": 0.3750480732251365,
      "grad_norm": 3.203125,
      "learning_rate": 2.4374278901622952e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2133092.0,
      "step": 4876
    },
    {
      "epoch": 0.37512499038535496,
      "grad_norm": 2.484375,
      "learning_rate": 2.4373125144219675e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2133872.0,
      "step": 4877
    },
    {
      "epoch": 0.3752019075455734,
      "grad_norm": 1.7265625,
      "learning_rate": 2.4371971386816397e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2134456.0,
      "step": 4878
    },
    {
      "epoch": 0.3752788247057919,
      "grad_norm": 8.625,
      "learning_rate": 2.4370817629413124e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2134738.0,
      "step": 4879
    },
    {
      "epoch": 0.3753557418660103,
      "grad_norm": 7.375,
      "learning_rate": 2.4369663872009847e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2135066.0,
      "step": 4880
    },
    {
      "epoch": 0.37543265902622874,
      "grad_norm": 7.4375,
      "learning_rate": 2.436851011460657e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 2135338.0,
      "step": 4881
    },
    {
      "epoch": 0.3755095761864472,
      "grad_norm": 12.0,
      "learning_rate": 2.4367356357203296e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 2135594.0,
      "step": 4882
    },
    {
      "epoch": 0.37558649334666566,
      "grad_norm": 3.796875,
      "learning_rate": 2.4366202599800015e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2136134.0,
      "step": 4883
    },
    {
      "epoch": 0.37566341050688407,
      "grad_norm": 5.34375,
      "learning_rate": 2.4365048842396737e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2136466.0,
      "step": 4884
    },
    {
      "epoch": 0.3757403276671025,
      "grad_norm": 6.78125,
      "learning_rate": 2.4363895084993464e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2136778.0,
      "step": 4885
    },
    {
      "epoch": 0.375817244827321,
      "grad_norm": 2.578125,
      "learning_rate": 2.4362741327590187e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2137598.0,
      "step": 4886
    },
    {
      "epoch": 0.37589416198753944,
      "grad_norm": 6.34375,
      "learning_rate": 2.436158757018691e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 2138200.0,
      "step": 4887
    },
    {
      "epoch": 0.37597107914775785,
      "grad_norm": 8.6875,
      "learning_rate": 2.436043381278363e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 2138480.0,
      "step": 4888
    },
    {
      "epoch": 0.3760479963079763,
      "grad_norm": 4.125,
      "learning_rate": 2.4359280055380354e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2138786.0,
      "step": 4889
    },
    {
      "epoch": 0.37612491346819477,
      "grad_norm": 6.53125,
      "learning_rate": 2.435812629797708e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2139108.0,
      "step": 4890
    },
    {
      "epoch": 0.3762018306284132,
      "grad_norm": 7.59375,
      "learning_rate": 2.4356972540573804e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2139408.0,
      "step": 4891
    },
    {
      "epoch": 0.37627874778863163,
      "grad_norm": 7.03125,
      "learning_rate": 2.4355818783170526e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 2139698.0,
      "step": 4892
    },
    {
      "epoch": 0.3763556649488501,
      "grad_norm": 6.28125,
      "learning_rate": 2.435466502576725e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 2140014.0,
      "step": 4893
    },
    {
      "epoch": 0.37643258210906855,
      "grad_norm": 6.4375,
      "learning_rate": 2.435351126836397e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2140338.0,
      "step": 4894
    },
    {
      "epoch": 0.37650949926928695,
      "grad_norm": 4.6875,
      "learning_rate": 2.4352357510960694e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2140650.0,
      "step": 4895
    },
    {
      "epoch": 0.3765864164295054,
      "grad_norm": 5.6875,
      "learning_rate": 2.435120375355742e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2140950.0,
      "step": 4896
    },
    {
      "epoch": 0.37666333358972387,
      "grad_norm": 7.03125,
      "learning_rate": 2.4350049996154143e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2141286.0,
      "step": 4897
    },
    {
      "epoch": 0.37674025074994233,
      "grad_norm": 3.53125,
      "learning_rate": 2.4348896238750866e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2141854.0,
      "step": 4898
    },
    {
      "epoch": 0.37681716791016073,
      "grad_norm": 7.96875,
      "learning_rate": 2.434774248134759e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2142132.0,
      "step": 4899
    },
    {
      "epoch": 0.3768940850703792,
      "grad_norm": 2.109375,
      "learning_rate": 2.434658872394431e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 2142946.0,
      "step": 4900
    },
    {
      "epoch": 0.37697100223059765,
      "grad_norm": 2.484375,
      "learning_rate": 2.434543496654104e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2143510.0,
      "step": 4901
    },
    {
      "epoch": 0.3770479193908161,
      "grad_norm": 7.375,
      "learning_rate": 2.434428120913776e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2143792.0,
      "step": 4902
    },
    {
      "epoch": 0.3771248365510345,
      "grad_norm": 6.625,
      "learning_rate": 2.4343127451734483e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 2144128.0,
      "step": 4903
    },
    {
      "epoch": 0.377201753711253,
      "grad_norm": 2.828125,
      "learning_rate": 2.4341973694331206e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 2144708.0,
      "step": 4904
    },
    {
      "epoch": 0.37727867087147143,
      "grad_norm": 8.0625,
      "learning_rate": 2.434081993692793e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2144990.0,
      "step": 4905
    },
    {
      "epoch": 0.3773555880316899,
      "grad_norm": 4.90625,
      "learning_rate": 2.433966617952465e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 2145342.0,
      "step": 4906
    },
    {
      "epoch": 0.3774325051919083,
      "grad_norm": 2.921875,
      "learning_rate": 2.4338512422121378e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9948006868362427,
      "num_tokens": 2145920.0,
      "step": 4907
    },
    {
      "epoch": 0.37750942235212676,
      "grad_norm": 5.8125,
      "learning_rate": 2.43373586647181e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2146214.0,
      "step": 4908
    },
    {
      "epoch": 0.3775863395123452,
      "grad_norm": 7.875,
      "learning_rate": 2.4336204907314823e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2146496.0,
      "step": 4909
    },
    {
      "epoch": 0.3776632566725637,
      "grad_norm": 5.5625,
      "learning_rate": 2.4335051149911545e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2146790.0,
      "step": 4910
    },
    {
      "epoch": 0.3777401738327821,
      "grad_norm": 3.546875,
      "learning_rate": 2.433389739250827e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2147140.0,
      "step": 4911
    },
    {
      "epoch": 0.37781709099300054,
      "grad_norm": 5.1875,
      "learning_rate": 2.433274363510499e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2147424.0,
      "step": 4912
    },
    {
      "epoch": 0.377894008153219,
      "grad_norm": 9.25,
      "learning_rate": 2.4331589877701717e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 2147724.0,
      "step": 4913
    },
    {
      "epoch": 0.3779709253134374,
      "grad_norm": 5.875,
      "learning_rate": 2.433043612029844e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 2148010.0,
      "step": 4914
    },
    {
      "epoch": 0.37804784247365586,
      "grad_norm": 3.40625,
      "learning_rate": 2.4329282362895163e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2148546.0,
      "step": 4915
    },
    {
      "epoch": 0.3781247596338743,
      "grad_norm": 2.640625,
      "learning_rate": 2.4328128605491885e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 2149068.0,
      "step": 4916
    },
    {
      "epoch": 0.3782016767940928,
      "grad_norm": 5.75,
      "learning_rate": 2.4326974848088608e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2149350.0,
      "step": 4917
    },
    {
      "epoch": 0.3782785939543112,
      "grad_norm": 4.21875,
      "learning_rate": 2.4325821090685335e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2149670.0,
      "step": 4918
    },
    {
      "epoch": 0.37835551111452964,
      "grad_norm": 4.78125,
      "learning_rate": 2.4324667333282057e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2149998.0,
      "step": 4919
    },
    {
      "epoch": 0.3784324282747481,
      "grad_norm": 5.125,
      "learning_rate": 2.432351357587878e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2150322.0,
      "step": 4920
    },
    {
      "epoch": 0.37850934543496656,
      "grad_norm": 4.0625,
      "learning_rate": 2.4322359818475502e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2150606.0,
      "step": 4921
    },
    {
      "epoch": 0.37858626259518496,
      "grad_norm": 5.71875,
      "learning_rate": 2.4321206061072225e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2150892.0,
      "step": 4922
    },
    {
      "epoch": 0.3786631797554034,
      "grad_norm": 3.5,
      "learning_rate": 2.4320052303668947e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2151204.0,
      "step": 4923
    },
    {
      "epoch": 0.3787400969156219,
      "grad_norm": 2.265625,
      "learning_rate": 2.4318898546265674e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2151750.0,
      "step": 4924
    },
    {
      "epoch": 0.37881701407584034,
      "grad_norm": 6.9375,
      "learning_rate": 2.4317744788862397e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2152102.0,
      "step": 4925
    },
    {
      "epoch": 0.37889393123605875,
      "grad_norm": 3.921875,
      "learning_rate": 2.431659103145912e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2152682.0,
      "step": 4926
    },
    {
      "epoch": 0.3789708483962772,
      "grad_norm": 7.90625,
      "learning_rate": 2.431543727405584e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 2152990.0,
      "step": 4927
    },
    {
      "epoch": 0.37904776555649566,
      "grad_norm": 6.21875,
      "learning_rate": 2.4314283516652565e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2153270.0,
      "step": 4928
    },
    {
      "epoch": 0.3791246827167141,
      "grad_norm": 4.4375,
      "learning_rate": 2.431312975924929e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2153614.0,
      "step": 4929
    },
    {
      "epoch": 0.3792015998769325,
      "grad_norm": 2.71875,
      "learning_rate": 2.4311976001846014e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2154176.0,
      "step": 4930
    },
    {
      "epoch": 0.379278517037151,
      "grad_norm": 4.25,
      "learning_rate": 2.4310822244442732e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 2154748.0,
      "step": 4931
    },
    {
      "epoch": 0.37935543419736945,
      "grad_norm": 6.8125,
      "learning_rate": 2.430966848703946e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2155028.0,
      "step": 4932
    },
    {
      "epoch": 0.3794323513575879,
      "grad_norm": 3.421875,
      "learning_rate": 2.430851472963618e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 2155326.0,
      "step": 4933
    },
    {
      "epoch": 0.3795092685178063,
      "grad_norm": 2.28125,
      "learning_rate": 2.4307360972232904e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2155910.0,
      "step": 4934
    },
    {
      "epoch": 0.37958618567802477,
      "grad_norm": 5.46875,
      "learning_rate": 2.430620721482963e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 2156250.0,
      "step": 4935
    },
    {
      "epoch": 0.3796631028382432,
      "grad_norm": 4.5,
      "learning_rate": 2.4305053457426354e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2156598.0,
      "step": 4936
    },
    {
      "epoch": 0.37974001999846163,
      "grad_norm": 2.078125,
      "learning_rate": 2.4303899700023076e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2157178.0,
      "step": 4937
    },
    {
      "epoch": 0.3798169371586801,
      "grad_norm": 4.125,
      "learning_rate": 2.43027459426198e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2157500.0,
      "step": 4938
    },
    {
      "epoch": 0.37989385431889855,
      "grad_norm": 8.6875,
      "learning_rate": 2.430159218521652e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2157782.0,
      "step": 4939
    },
    {
      "epoch": 0.379970771479117,
      "grad_norm": 3.875,
      "learning_rate": 2.430043842781325e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 2158346.0,
      "step": 4940
    },
    {
      "epoch": 0.3800476886393354,
      "grad_norm": 8.5,
      "learning_rate": 2.429928467040997e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2158650.0,
      "step": 4941
    },
    {
      "epoch": 0.38012460579955387,
      "grad_norm": 5.875,
      "learning_rate": 2.429813091300669e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2158956.0,
      "step": 4942
    },
    {
      "epoch": 0.38020152295977233,
      "grad_norm": 3.90625,
      "learning_rate": 2.4296977155603416e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2159472.0,
      "step": 4943
    },
    {
      "epoch": 0.3802784401199908,
      "grad_norm": 2.234375,
      "learning_rate": 2.429582339820014e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2160292.0,
      "step": 4944
    },
    {
      "epoch": 0.3803553572802092,
      "grad_norm": 6.5,
      "learning_rate": 2.429466964079686e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2160618.0,
      "step": 4945
    },
    {
      "epoch": 0.38043227444042765,
      "grad_norm": 5.625,
      "learning_rate": 2.429351588339359e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2160900.0,
      "step": 4946
    },
    {
      "epoch": 0.3805091916006461,
      "grad_norm": 8.375,
      "learning_rate": 2.429236212599031e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 2161154.0,
      "step": 4947
    },
    {
      "epoch": 0.38058610876086457,
      "grad_norm": 5.3125,
      "learning_rate": 2.4291208368587033e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 2161494.0,
      "step": 4948
    },
    {
      "epoch": 0.380663025921083,
      "grad_norm": 2.8125,
      "learning_rate": 2.4290054611183756e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2162290.0,
      "step": 4949
    },
    {
      "epoch": 0.38073994308130144,
      "grad_norm": 6.65625,
      "learning_rate": 2.428890085378048e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2162606.0,
      "step": 4950
    },
    {
      "epoch": 0.3808168602415199,
      "grad_norm": 5.375,
      "learning_rate": 2.4287747096377205e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 2162944.0,
      "step": 4951
    },
    {
      "epoch": 0.38089377740173835,
      "grad_norm": 2.65625,
      "learning_rate": 2.4286593338973928e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 2163994.0,
      "step": 4952
    },
    {
      "epoch": 0.38097069456195676,
      "grad_norm": 5.5625,
      "learning_rate": 2.4285439581570646e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2164278.0,
      "step": 4953
    },
    {
      "epoch": 0.3810476117221752,
      "grad_norm": 5.15625,
      "learning_rate": 2.4284285824167373e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 2164878.0,
      "step": 4954
    },
    {
      "epoch": 0.3811245288823937,
      "grad_norm": 3.65625,
      "learning_rate": 2.4283132066764095e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2165222.0,
      "step": 4955
    },
    {
      "epoch": 0.3812014460426121,
      "grad_norm": 6.75,
      "learning_rate": 2.428197830936082e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2165532.0,
      "step": 4956
    },
    {
      "epoch": 0.38127836320283054,
      "grad_norm": 4.0625,
      "learning_rate": 2.4280824551957545e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 2166028.0,
      "step": 4957
    },
    {
      "epoch": 0.381355280363049,
      "grad_norm": 5.625,
      "learning_rate": 2.4279670794554267e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2166352.0,
      "step": 4958
    },
    {
      "epoch": 0.38143219752326746,
      "grad_norm": 2.5625,
      "learning_rate": 2.427851703715099e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2166870.0,
      "step": 4959
    },
    {
      "epoch": 0.38150911468348586,
      "grad_norm": 2.71875,
      "learning_rate": 2.4277363279747713e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 2167668.0,
      "step": 4960
    },
    {
      "epoch": 0.3815860318437043,
      "grad_norm": 5.125,
      "learning_rate": 2.4276209522344435e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2167996.0,
      "step": 4961
    },
    {
      "epoch": 0.3816629490039228,
      "grad_norm": 4.1875,
      "learning_rate": 2.4275055764941158e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 2168322.0,
      "step": 4962
    },
    {
      "epoch": 0.38173986616414124,
      "grad_norm": 2.46875,
      "learning_rate": 2.4273902007537885e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2168870.0,
      "step": 4963
    },
    {
      "epoch": 0.38181678332435964,
      "grad_norm": 4.96875,
      "learning_rate": 2.4272748250134603e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2169214.0,
      "step": 4964
    },
    {
      "epoch": 0.3818937004845781,
      "grad_norm": 6.53125,
      "learning_rate": 2.427159449273133e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2169554.0,
      "step": 4965
    },
    {
      "epoch": 0.38197061764479656,
      "grad_norm": 5.375,
      "learning_rate": 2.4270440735328052e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2169866.0,
      "step": 4966
    },
    {
      "epoch": 0.382047534805015,
      "grad_norm": 3.875,
      "learning_rate": 2.4269286977924775e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 2170390.0,
      "step": 4967
    },
    {
      "epoch": 0.3821244519652334,
      "grad_norm": 4.53125,
      "learning_rate": 2.42681332205215e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2170708.0,
      "step": 4968
    },
    {
      "epoch": 0.3822013691254519,
      "grad_norm": 6.46875,
      "learning_rate": 2.426697946311822e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2170992.0,
      "step": 4969
    },
    {
      "epoch": 0.38227828628567034,
      "grad_norm": 3.59375,
      "learning_rate": 2.4265825705714943e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 2171584.0,
      "step": 4970
    },
    {
      "epoch": 0.3823552034458888,
      "grad_norm": 3.421875,
      "learning_rate": 2.426467194831167e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2172102.0,
      "step": 4971
    },
    {
      "epoch": 0.3824321206061072,
      "grad_norm": 8.0,
      "learning_rate": 2.426351819090839e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 2172410.0,
      "step": 4972
    },
    {
      "epoch": 0.38250903776632567,
      "grad_norm": 1.890625,
      "learning_rate": 2.4262364433505115e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 2173208.0,
      "step": 4973
    },
    {
      "epoch": 0.3825859549265441,
      "grad_norm": 7.65625,
      "learning_rate": 2.426121067610184e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2173512.0,
      "step": 4974
    },
    {
      "epoch": 0.3826628720867626,
      "grad_norm": 7.4375,
      "learning_rate": 2.426005691869856e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2173790.0,
      "step": 4975
    },
    {
      "epoch": 0.382739789246981,
      "grad_norm": 1.625,
      "learning_rate": 2.4258903161295287e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2174342.0,
      "step": 4976
    },
    {
      "epoch": 0.38281670640719945,
      "grad_norm": 2.75,
      "learning_rate": 2.425774940389201e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2174860.0,
      "step": 4977
    },
    {
      "epoch": 0.3828936235674179,
      "grad_norm": 3.625,
      "learning_rate": 2.425659564648873e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2175430.0,
      "step": 4978
    },
    {
      "epoch": 0.3829705407276363,
      "grad_norm": 5.34375,
      "learning_rate": 2.425544188908546e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 2175740.0,
      "step": 4979
    },
    {
      "epoch": 0.38304745788785477,
      "grad_norm": 6.09375,
      "learning_rate": 2.4254288131682177e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2176086.0,
      "step": 4980
    },
    {
      "epoch": 0.38312437504807323,
      "grad_norm": 6.15625,
      "learning_rate": 2.42531343742789e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9825072884559631,
      "num_tokens": 2176430.0,
      "step": 4981
    },
    {
      "epoch": 0.3832012922082917,
      "grad_norm": 6.25,
      "learning_rate": 2.4251980616875626e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9817629456520081,
      "num_tokens": 2176760.0,
      "step": 4982
    },
    {
      "epoch": 0.3832782093685101,
      "grad_norm": 2.4375,
      "learning_rate": 2.425082685947235e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 2177328.0,
      "step": 4983
    },
    {
      "epoch": 0.38335512652872855,
      "grad_norm": 5.84375,
      "learning_rate": 2.424967310206907e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2177640.0,
      "step": 4984
    },
    {
      "epoch": 0.383432043688947,
      "grad_norm": 2.15625,
      "learning_rate": 2.42485193446658e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 2178428.0,
      "step": 4985
    },
    {
      "epoch": 0.38350896084916547,
      "grad_norm": 6.375,
      "learning_rate": 2.4247365587262517e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 2178698.0,
      "step": 4986
    },
    {
      "epoch": 0.3835858780093839,
      "grad_norm": 4.34375,
      "learning_rate": 2.4246211829859243e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2178986.0,
      "step": 4987
    },
    {
      "epoch": 0.38366279516960233,
      "grad_norm": 5.59375,
      "learning_rate": 2.4245058072455966e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 2179300.0,
      "step": 4988
    },
    {
      "epoch": 0.3837397123298208,
      "grad_norm": 2.125,
      "learning_rate": 2.424390431505269e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2179848.0,
      "step": 4989
    },
    {
      "epoch": 0.38381662949003925,
      "grad_norm": 2.546875,
      "learning_rate": 2.4242750557649415e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 2180652.0,
      "step": 4990
    },
    {
      "epoch": 0.38389354665025766,
      "grad_norm": 2.453125,
      "learning_rate": 2.4241596800246134e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 2181464.0,
      "step": 4991
    },
    {
      "epoch": 0.3839704638104761,
      "grad_norm": 5.0625,
      "learning_rate": 2.4240443042842856e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2182060.0,
      "step": 4992
    },
    {
      "epoch": 0.3840473809706946,
      "grad_norm": 7.53125,
      "learning_rate": 2.4239289285439583e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2182394.0,
      "step": 4993
    },
    {
      "epoch": 0.38412429813091303,
      "grad_norm": 4.625,
      "learning_rate": 2.4238135528036306e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.997032642364502,
      "num_tokens": 2182732.0,
      "step": 4994
    },
    {
      "epoch": 0.38420121529113144,
      "grad_norm": 6.09375,
      "learning_rate": 2.423698177063303e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9941520690917969,
      "num_tokens": 2183076.0,
      "step": 4995
    },
    {
      "epoch": 0.3842781324513499,
      "grad_norm": 8.1875,
      "learning_rate": 2.4235828013229755e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2183432.0,
      "step": 4996
    },
    {
      "epoch": 0.38435504961156836,
      "grad_norm": 3.046875,
      "learning_rate": 2.4234674255826473e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2183996.0,
      "step": 4997
    },
    {
      "epoch": 0.38443196677178676,
      "grad_norm": 6.75,
      "learning_rate": 2.42335204984232e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 2184282.0,
      "step": 4998
    },
    {
      "epoch": 0.3845088839320052,
      "grad_norm": 6.71875,
      "learning_rate": 2.4232366741019923e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2184616.0,
      "step": 4999
    },
    {
      "epoch": 0.3845858010922237,
      "grad_norm": 2.28125,
      "learning_rate": 2.4231212983616645e-06,
      "loss": 0.0154,
      "step": 5000
    },
    {
      "epoch": 0.3845858010922237,
      "eval_loss": 0.02112576551735401,
      "eval_mean_token_accuracy": 0.9924541926383972,
      "eval_num_tokens": 2185176.0,
      "eval_runtime": 21.0996,
      "eval_samples_per_second": 9.479,
      "eval_steps_per_second": 1.185,
      "step": 5000
    },
    {
      "epoch": 0.38466271825244214,
      "grad_norm": 2.859375,
      "learning_rate": 2.4230059226213372e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946804344654083,
      "num_tokens": 2185746.0,
      "step": 5001
    },
    {
      "epoch": 0.38473963541266054,
      "grad_norm": 5.34375,
      "learning_rate": 2.422890546881009e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2186106.0,
      "step": 5002
    },
    {
      "epoch": 0.384816552572879,
      "grad_norm": 7.625,
      "learning_rate": 2.4227751711406813e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 2186478.0,
      "step": 5003
    },
    {
      "epoch": 0.38489346973309746,
      "grad_norm": 4.84375,
      "learning_rate": 2.422659795400354e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 2186818.0,
      "step": 5004
    },
    {
      "epoch": 0.3849703868933159,
      "grad_norm": 13.375,
      "learning_rate": 2.4225444196600263e-06,
      "loss": 0.046,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 2187074.0,
      "step": 5005
    },
    {
      "epoch": 0.3850473040535343,
      "grad_norm": 4.1875,
      "learning_rate": 2.4224290439196985e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2187396.0,
      "step": 5006
    },
    {
      "epoch": 0.3851242212137528,
      "grad_norm": 2.8125,
      "learning_rate": 2.4223136681793708e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 2187988.0,
      "step": 5007
    },
    {
      "epoch": 0.38520113837397124,
      "grad_norm": 3.3125,
      "learning_rate": 2.422198292439043e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2188540.0,
      "step": 5008
    },
    {
      "epoch": 0.3852780555341897,
      "grad_norm": 6.125,
      "learning_rate": 2.4220829166987157e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2188862.0,
      "step": 5009
    },
    {
      "epoch": 0.3853549726944081,
      "grad_norm": 2.828125,
      "learning_rate": 2.421967540958388e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2189644.0,
      "step": 5010
    },
    {
      "epoch": 0.38543188985462656,
      "grad_norm": 5.625,
      "learning_rate": 2.4218521652180602e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2189964.0,
      "step": 5011
    },
    {
      "epoch": 0.385508807014845,
      "grad_norm": 4.5,
      "learning_rate": 2.4217367894777325e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9907578825950623,
      "num_tokens": 2190506.0,
      "step": 5012
    },
    {
      "epoch": 0.3855857241750635,
      "grad_norm": 3.8125,
      "learning_rate": 2.4216214137374047e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2190830.0,
      "step": 5013
    },
    {
      "epoch": 0.3856626413352819,
      "grad_norm": 5.9375,
      "learning_rate": 2.421506037997077e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9969040155410767,
      "num_tokens": 2191154.0,
      "step": 5014
    },
    {
      "epoch": 0.38573955849550035,
      "grad_norm": 3.28125,
      "learning_rate": 2.4213906622567497e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2191700.0,
      "step": 5015
    },
    {
      "epoch": 0.3858164756557188,
      "grad_norm": 2.9375,
      "learning_rate": 2.421275286516422e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2192246.0,
      "step": 5016
    },
    {
      "epoch": 0.3858933928159372,
      "grad_norm": 5.125,
      "learning_rate": 2.421159910776094e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 2192584.0,
      "step": 5017
    },
    {
      "epoch": 0.38597030997615567,
      "grad_norm": 2.921875,
      "learning_rate": 2.4210445350357665e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2193174.0,
      "step": 5018
    },
    {
      "epoch": 0.3860472271363741,
      "grad_norm": 4.25,
      "learning_rate": 2.4209291592954387e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2193482.0,
      "step": 5019
    },
    {
      "epoch": 0.3861241442965926,
      "grad_norm": 3.734375,
      "learning_rate": 2.420813783555111e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2194062.0,
      "step": 5020
    },
    {
      "epoch": 0.386201061456811,
      "grad_norm": 6.28125,
      "learning_rate": 2.4206984078147837e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2194342.0,
      "step": 5021
    },
    {
      "epoch": 0.38627797861702945,
      "grad_norm": 2.8125,
      "learning_rate": 2.420583032074456e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2194876.0,
      "step": 5022
    },
    {
      "epoch": 0.3863548957772479,
      "grad_norm": 4.96875,
      "learning_rate": 2.420467656334128e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2195220.0,
      "step": 5023
    },
    {
      "epoch": 0.38643181293746637,
      "grad_norm": 2.796875,
      "learning_rate": 2.4203522805938004e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2195754.0,
      "step": 5024
    },
    {
      "epoch": 0.38650873009768477,
      "grad_norm": 4.3125,
      "learning_rate": 2.4202369048534727e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2196086.0,
      "step": 5025
    },
    {
      "epoch": 0.38658564725790323,
      "grad_norm": 5.625,
      "learning_rate": 2.4201215291131454e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2196410.0,
      "step": 5026
    },
    {
      "epoch": 0.3866625644181217,
      "grad_norm": 3.921875,
      "learning_rate": 2.4200061533728176e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 2196708.0,
      "step": 5027
    },
    {
      "epoch": 0.38673948157834015,
      "grad_norm": 5.375,
      "learning_rate": 2.41989077763249e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2196994.0,
      "step": 5028
    },
    {
      "epoch": 0.38681639873855855,
      "grad_norm": 3.09375,
      "learning_rate": 2.419775401892162e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 2197514.0,
      "step": 5029
    },
    {
      "epoch": 0.386893315898777,
      "grad_norm": 4.9375,
      "learning_rate": 2.4196600261518344e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2197842.0,
      "step": 5030
    },
    {
      "epoch": 0.38697023305899547,
      "grad_norm": 2.265625,
      "learning_rate": 2.4195446504115067e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 2198392.0,
      "step": 5031
    },
    {
      "epoch": 0.38704715021921393,
      "grad_norm": 2.578125,
      "learning_rate": 2.4194292746711793e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2198962.0,
      "step": 5032
    },
    {
      "epoch": 0.38712406737943234,
      "grad_norm": 3.328125,
      "learning_rate": 2.4193138989308516e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2199498.0,
      "step": 5033
    },
    {
      "epoch": 0.3872009845396508,
      "grad_norm": 3.734375,
      "learning_rate": 2.419198523190524e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2199842.0,
      "step": 5034
    },
    {
      "epoch": 0.38727790169986925,
      "grad_norm": 2.328125,
      "learning_rate": 2.419083147450196e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2200388.0,
      "step": 5035
    },
    {
      "epoch": 0.3873548188600877,
      "grad_norm": 2.34375,
      "learning_rate": 2.4189677717098684e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2201208.0,
      "step": 5036
    },
    {
      "epoch": 0.3874317360203061,
      "grad_norm": 2.203125,
      "learning_rate": 2.418852395969541e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2201756.0,
      "step": 5037
    },
    {
      "epoch": 0.3875086531805246,
      "grad_norm": 2.453125,
      "learning_rate": 2.4187370202292133e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2202320.0,
      "step": 5038
    },
    {
      "epoch": 0.38758557034074304,
      "grad_norm": 3.421875,
      "learning_rate": 2.4186216444888856e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2202840.0,
      "step": 5039
    },
    {
      "epoch": 0.38766248750096144,
      "grad_norm": 2.015625,
      "learning_rate": 2.418506268748558e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2203424.0,
      "step": 5040
    },
    {
      "epoch": 0.3877394046611799,
      "grad_norm": 2.0625,
      "learning_rate": 2.41839089300823e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 2204232.0,
      "step": 5041
    },
    {
      "epoch": 0.38781632182139836,
      "grad_norm": 1.5859375,
      "learning_rate": 2.4182755172679023e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 2205028.0,
      "step": 5042
    },
    {
      "epoch": 0.3878932389816168,
      "grad_norm": 1.765625,
      "learning_rate": 2.418160141527575e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2205586.0,
      "step": 5043
    },
    {
      "epoch": 0.3879701561418352,
      "grad_norm": 6.375,
      "learning_rate": 2.4180447657872473e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2205902.0,
      "step": 5044
    },
    {
      "epoch": 0.3880470733020537,
      "grad_norm": 3.796875,
      "learning_rate": 2.4179293900469195e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2206230.0,
      "step": 5045
    },
    {
      "epoch": 0.38812399046227214,
      "grad_norm": 5.09375,
      "learning_rate": 2.417814014306592e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2206546.0,
      "step": 5046
    },
    {
      "epoch": 0.3882009076224906,
      "grad_norm": 2.453125,
      "learning_rate": 2.417698638566264e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2207116.0,
      "step": 5047
    },
    {
      "epoch": 0.388277824782709,
      "grad_norm": 8.4375,
      "learning_rate": 2.4175832628259367e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 2207448.0,
      "step": 5048
    },
    {
      "epoch": 0.38835474194292746,
      "grad_norm": 5.34375,
      "learning_rate": 2.417467887085609e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 2207742.0,
      "step": 5049
    },
    {
      "epoch": 0.3884316591031459,
      "grad_norm": 8.125,
      "learning_rate": 2.4173525113452813e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2208002.0,
      "step": 5050
    },
    {
      "epoch": 0.3885085762633644,
      "grad_norm": 2.59375,
      "learning_rate": 2.4172371356049535e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 2208538.0,
      "step": 5051
    },
    {
      "epoch": 0.3885854934235828,
      "grad_norm": 3.90625,
      "learning_rate": 2.4171217598646258e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 2208888.0,
      "step": 5052
    },
    {
      "epoch": 0.38866241058380124,
      "grad_norm": 1.7109375,
      "learning_rate": 2.417006384124298e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9975247383117676,
      "num_tokens": 2209698.0,
      "step": 5053
    },
    {
      "epoch": 0.3887393277440197,
      "grad_norm": 2.328125,
      "learning_rate": 2.4168910083839707e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2210244.0,
      "step": 5054
    },
    {
      "epoch": 0.38881624490423816,
      "grad_norm": 4.5625,
      "learning_rate": 2.416775632643643e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2210556.0,
      "step": 5055
    },
    {
      "epoch": 0.38889316206445657,
      "grad_norm": 7.53125,
      "learning_rate": 2.4166602569033152e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 2210864.0,
      "step": 5056
    },
    {
      "epoch": 0.388970079224675,
      "grad_norm": 6.28125,
      "learning_rate": 2.4165448811629875e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2211200.0,
      "step": 5057
    },
    {
      "epoch": 0.3890469963848935,
      "grad_norm": 3.265625,
      "learning_rate": 2.4164295054226597e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2211780.0,
      "step": 5058
    },
    {
      "epoch": 0.3891239135451119,
      "grad_norm": 6.78125,
      "learning_rate": 2.4163141296823324e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9928315281867981,
      "num_tokens": 2212060.0,
      "step": 5059
    },
    {
      "epoch": 0.38920083070533035,
      "grad_norm": 4.4375,
      "learning_rate": 2.4161987539420047e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2212382.0,
      "step": 5060
    },
    {
      "epoch": 0.3892777478655488,
      "grad_norm": 5.1875,
      "learning_rate": 2.416083378201677e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2212712.0,
      "step": 5061
    },
    {
      "epoch": 0.38935466502576727,
      "grad_norm": 3.03125,
      "learning_rate": 2.415968002461349e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 2213236.0,
      "step": 5062
    },
    {
      "epoch": 0.38943158218598567,
      "grad_norm": 5.34375,
      "learning_rate": 2.4158526267210215e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2213552.0,
      "step": 5063
    },
    {
      "epoch": 0.38950849934620413,
      "grad_norm": 3.921875,
      "learning_rate": 2.4157372509806937e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2213896.0,
      "step": 5064
    },
    {
      "epoch": 0.3895854165064226,
      "grad_norm": 5.46875,
      "learning_rate": 2.4156218752403664e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 2214214.0,
      "step": 5065
    },
    {
      "epoch": 0.38966233366664105,
      "grad_norm": 3.015625,
      "learning_rate": 2.4155064995000387e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 2214778.0,
      "step": 5066
    },
    {
      "epoch": 0.38973925082685945,
      "grad_norm": 5.40625,
      "learning_rate": 2.415391123759711e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2215084.0,
      "step": 5067
    },
    {
      "epoch": 0.3898161679870779,
      "grad_norm": 4.0,
      "learning_rate": 2.415275748019383e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 2215676.0,
      "step": 5068
    },
    {
      "epoch": 0.38989308514729637,
      "grad_norm": 3.296875,
      "learning_rate": 2.4151603722790554e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2216230.0,
      "step": 5069
    },
    {
      "epoch": 0.38997000230751483,
      "grad_norm": 3.21875,
      "learning_rate": 2.4150449965387277e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 2216798.0,
      "step": 5070
    },
    {
      "epoch": 0.39004691946773323,
      "grad_norm": 4.0,
      "learning_rate": 2.4149296207984004e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2217142.0,
      "step": 5071
    },
    {
      "epoch": 0.3901238366279517,
      "grad_norm": 7.5625,
      "learning_rate": 2.414814245058072e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2217396.0,
      "step": 5072
    },
    {
      "epoch": 0.39020075378817015,
      "grad_norm": 3.71875,
      "learning_rate": 2.414698869317745e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 2217730.0,
      "step": 5073
    },
    {
      "epoch": 0.3902776709483886,
      "grad_norm": 5.5625,
      "learning_rate": 2.414583493577417e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2218028.0,
      "step": 5074
    },
    {
      "epoch": 0.390354588108607,
      "grad_norm": 4.40625,
      "learning_rate": 2.4144681178370894e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2218350.0,
      "step": 5075
    },
    {
      "epoch": 0.3904315052688255,
      "grad_norm": 2.0625,
      "learning_rate": 2.414352742096762e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 2219164.0,
      "step": 5076
    },
    {
      "epoch": 0.39050842242904393,
      "grad_norm": 4.3125,
      "learning_rate": 2.4142373663564343e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2219760.0,
      "step": 5077
    },
    {
      "epoch": 0.39058533958926234,
      "grad_norm": 2.703125,
      "learning_rate": 2.414121990616106e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 2220320.0,
      "step": 5078
    },
    {
      "epoch": 0.3906622567494808,
      "grad_norm": 2.734375,
      "learning_rate": 2.414006614875779e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2220860.0,
      "step": 5079
    },
    {
      "epoch": 0.39073917390969926,
      "grad_norm": 2.203125,
      "learning_rate": 2.413891239135451e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2221642.0,
      "step": 5080
    },
    {
      "epoch": 0.3908160910699177,
      "grad_norm": 4.03125,
      "learning_rate": 2.4137758633951234e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 2222244.0,
      "step": 5081
    },
    {
      "epoch": 0.3908930082301361,
      "grad_norm": 4.5,
      "learning_rate": 2.413660487654796e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2222544.0,
      "step": 5082
    },
    {
      "epoch": 0.3909699253903546,
      "grad_norm": 5.40625,
      "learning_rate": 2.413545111914468e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 2222832.0,
      "step": 5083
    },
    {
      "epoch": 0.39104684255057304,
      "grad_norm": 6.875,
      "learning_rate": 2.4134297361741406e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2223138.0,
      "step": 5084
    },
    {
      "epoch": 0.3911237597107915,
      "grad_norm": 3.296875,
      "learning_rate": 2.413314360433813e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2223464.0,
      "step": 5085
    },
    {
      "epoch": 0.3912006768710099,
      "grad_norm": 5.28125,
      "learning_rate": 2.413198984693485e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2223778.0,
      "step": 5086
    },
    {
      "epoch": 0.39127759403122836,
      "grad_norm": 3.453125,
      "learning_rate": 2.4130836089531578e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2224304.0,
      "step": 5087
    },
    {
      "epoch": 0.3913545111914468,
      "grad_norm": 2.609375,
      "learning_rate": 2.41296823321283e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2224872.0,
      "step": 5088
    },
    {
      "epoch": 0.3914314283516653,
      "grad_norm": 3.578125,
      "learning_rate": 2.412852857472502e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 2225194.0,
      "step": 5089
    },
    {
      "epoch": 0.3915083455118837,
      "grad_norm": 2.15625,
      "learning_rate": 2.4127374817321746e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2225774.0,
      "step": 5090
    },
    {
      "epoch": 0.39158526267210214,
      "grad_norm": 4.875,
      "learning_rate": 2.412622105991847e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9943661689758301,
      "num_tokens": 2226130.0,
      "step": 5091
    },
    {
      "epoch": 0.3916621798323206,
      "grad_norm": 3.65625,
      "learning_rate": 2.412506730251519e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2226452.0,
      "step": 5092
    },
    {
      "epoch": 0.39173909699253906,
      "grad_norm": 2.390625,
      "learning_rate": 2.4123913545111917e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9948892593383789,
      "num_tokens": 2227040.0,
      "step": 5093
    },
    {
      "epoch": 0.39181601415275746,
      "grad_norm": 3.328125,
      "learning_rate": 2.4122759787708636e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 2227612.0,
      "step": 5094
    },
    {
      "epoch": 0.3918929313129759,
      "grad_norm": 2.984375,
      "learning_rate": 2.4121606030305363e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2227934.0,
      "step": 5095
    },
    {
      "epoch": 0.3919698484731944,
      "grad_norm": 4.5625,
      "learning_rate": 2.4120452272902085e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 2228272.0,
      "step": 5096
    },
    {
      "epoch": 0.39204676563341284,
      "grad_norm": 1.9296875,
      "learning_rate": 2.4119298515498808e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2228858.0,
      "step": 5097
    },
    {
      "epoch": 0.39212368279363125,
      "grad_norm": 5.15625,
      "learning_rate": 2.4118144758095535e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2229174.0,
      "step": 5098
    },
    {
      "epoch": 0.3922005999538497,
      "grad_norm": 7.3125,
      "learning_rate": 2.4116991000692257e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9820359349250793,
      "num_tokens": 2229510.0,
      "step": 5099
    },
    {
      "epoch": 0.39227751711406816,
      "grad_norm": 3.34375,
      "learning_rate": 2.4115837243288976e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2229834.0,
      "step": 5100
    },
    {
      "epoch": 0.39235443427428657,
      "grad_norm": 4.5,
      "learning_rate": 2.4114683485885702e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2230374.0,
      "step": 5101
    },
    {
      "epoch": 0.392431351434505,
      "grad_norm": 4.71875,
      "learning_rate": 2.4113529728482425e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 2230688.0,
      "step": 5102
    },
    {
      "epoch": 0.3925082685947235,
      "grad_norm": 5.5,
      "learning_rate": 2.4112375971079148e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2230984.0,
      "step": 5103
    },
    {
      "epoch": 0.39258518575494195,
      "grad_norm": 4.84375,
      "learning_rate": 2.4111222213675874e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2231320.0,
      "step": 5104
    },
    {
      "epoch": 0.39266210291516035,
      "grad_norm": 6.21875,
      "learning_rate": 2.4110068456272593e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9828178882598877,
      "num_tokens": 2231612.0,
      "step": 5105
    },
    {
      "epoch": 0.3927390200753788,
      "grad_norm": 7.28125,
      "learning_rate": 2.410891469886932e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2231916.0,
      "step": 5106
    },
    {
      "epoch": 0.39281593723559727,
      "grad_norm": 4.0,
      "learning_rate": 2.410776094146604e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2232432.0,
      "step": 5107
    },
    {
      "epoch": 0.3928928543958157,
      "grad_norm": 7.1875,
      "learning_rate": 2.4106607184062765e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2232760.0,
      "step": 5108
    },
    {
      "epoch": 0.39296977155603413,
      "grad_norm": 3.078125,
      "learning_rate": 2.410545342665949e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2233336.0,
      "step": 5109
    },
    {
      "epoch": 0.3930466887162526,
      "grad_norm": 5.28125,
      "learning_rate": 2.410429966925621e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 2233948.0,
      "step": 5110
    },
    {
      "epoch": 0.39312360587647105,
      "grad_norm": 4.625,
      "learning_rate": 2.4103145911852932e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 2234262.0,
      "step": 5111
    },
    {
      "epoch": 0.3932005230366895,
      "grad_norm": 4.65625,
      "learning_rate": 2.410199215444966e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2234586.0,
      "step": 5112
    },
    {
      "epoch": 0.3932774401969079,
      "grad_norm": 2.28125,
      "learning_rate": 2.410083839704638e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2235382.0,
      "step": 5113
    },
    {
      "epoch": 0.39335435735712637,
      "grad_norm": 5.5625,
      "learning_rate": 2.4099684639643104e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 2235708.0,
      "step": 5114
    },
    {
      "epoch": 0.39343127451734483,
      "grad_norm": 2.46875,
      "learning_rate": 2.409853088223983e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2236252.0,
      "step": 5115
    },
    {
      "epoch": 0.3935081916775633,
      "grad_norm": 4.53125,
      "learning_rate": 2.409737712483655e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2236556.0,
      "step": 5116
    },
    {
      "epoch": 0.3935851088377817,
      "grad_norm": 5.78125,
      "learning_rate": 2.4096223367433276e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2236908.0,
      "step": 5117
    },
    {
      "epoch": 0.39366202599800015,
      "grad_norm": 3.03125,
      "learning_rate": 2.409506961003e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2237460.0,
      "step": 5118
    },
    {
      "epoch": 0.3937389431582186,
      "grad_norm": 7.84375,
      "learning_rate": 2.409391585262672e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2237780.0,
      "step": 5119
    },
    {
      "epoch": 0.393815860318437,
      "grad_norm": 5.53125,
      "learning_rate": 2.4092762095223444e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2238124.0,
      "step": 5120
    },
    {
      "epoch": 0.3938927774786555,
      "grad_norm": 5.5625,
      "learning_rate": 2.4091608337820167e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2238456.0,
      "step": 5121
    },
    {
      "epoch": 0.39396969463887394,
      "grad_norm": 3.5,
      "learning_rate": 2.409045458041689e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2238998.0,
      "step": 5122
    },
    {
      "epoch": 0.3940466117990924,
      "grad_norm": 4.9375,
      "learning_rate": 2.4089300823013616e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2239282.0,
      "step": 5123
    },
    {
      "epoch": 0.3941235289593108,
      "grad_norm": 3.109375,
      "learning_rate": 2.408814706561034e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2239602.0,
      "step": 5124
    },
    {
      "epoch": 0.39420044611952926,
      "grad_norm": 4.21875,
      "learning_rate": 2.408699330820706e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2239900.0,
      "step": 5125
    },
    {
      "epoch": 0.3942773632797477,
      "grad_norm": 1.921875,
      "learning_rate": 2.408583955080379e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 2240450.0,
      "step": 5126
    },
    {
      "epoch": 0.3943542804399662,
      "grad_norm": 4.5,
      "learning_rate": 2.4084685793400506e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2240774.0,
      "step": 5127
    },
    {
      "epoch": 0.3944311976001846,
      "grad_norm": 3.21875,
      "learning_rate": 2.408353203599723e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 2241294.0,
      "step": 5128
    },
    {
      "epoch": 0.39450811476040304,
      "grad_norm": 2.140625,
      "learning_rate": 2.4082378278593956e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2241850.0,
      "step": 5129
    },
    {
      "epoch": 0.3945850319206215,
      "grad_norm": 3.71875,
      "learning_rate": 2.408122452119068e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2242446.0,
      "step": 5130
    },
    {
      "epoch": 0.39466194908083996,
      "grad_norm": 7.46875,
      "learning_rate": 2.40800707637874e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2242746.0,
      "step": 5131
    },
    {
      "epoch": 0.39473886624105836,
      "grad_norm": 5.5,
      "learning_rate": 2.4078917006384124e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2243070.0,
      "step": 5132
    },
    {
      "epoch": 0.3948157834012768,
      "grad_norm": 4.5,
      "learning_rate": 2.4077763248980846e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2243382.0,
      "step": 5133
    },
    {
      "epoch": 0.3948927005614953,
      "grad_norm": 9.5625,
      "learning_rate": 2.4076609491577573e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2243684.0,
      "step": 5134
    },
    {
      "epoch": 0.39496961772171374,
      "grad_norm": 3.75,
      "learning_rate": 2.4075455734174296e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2244000.0,
      "step": 5135
    },
    {
      "epoch": 0.39504653488193214,
      "grad_norm": 6.34375,
      "learning_rate": 2.407430197677102e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2244302.0,
      "step": 5136
    },
    {
      "epoch": 0.3951234520421506,
      "grad_norm": 10.4375,
      "learning_rate": 2.407314821936774e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 2244534.0,
      "step": 5137
    },
    {
      "epoch": 0.39520036920236906,
      "grad_norm": 5.40625,
      "learning_rate": 2.4071994461964463e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2244816.0,
      "step": 5138
    },
    {
      "epoch": 0.39527728636258747,
      "grad_norm": 8.75,
      "learning_rate": 2.4070840704561186e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 2245090.0,
      "step": 5139
    },
    {
      "epoch": 0.3953542035228059,
      "grad_norm": 3.046875,
      "learning_rate": 2.4069686947157913e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2245680.0,
      "step": 5140
    },
    {
      "epoch": 0.3954311206830244,
      "grad_norm": 4.625,
      "learning_rate": 2.4068533189754635e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2246030.0,
      "step": 5141
    },
    {
      "epoch": 0.39550803784324284,
      "grad_norm": 5.625,
      "learning_rate": 2.4067379432351358e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2246342.0,
      "step": 5142
    },
    {
      "epoch": 0.39558495500346125,
      "grad_norm": 2.703125,
      "learning_rate": 2.406622567494808e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2246916.0,
      "step": 5143
    },
    {
      "epoch": 0.3956618721636797,
      "grad_norm": 6.125,
      "learning_rate": 2.4065071917544803e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 2247256.0,
      "step": 5144
    },
    {
      "epoch": 0.39573878932389817,
      "grad_norm": 5.8125,
      "learning_rate": 2.406391816014153e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 2247596.0,
      "step": 5145
    },
    {
      "epoch": 0.3958157064841166,
      "grad_norm": 5.5,
      "learning_rate": 2.4062764402738252e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2247908.0,
      "step": 5146
    },
    {
      "epoch": 0.39589262364433503,
      "grad_norm": 2.671875,
      "learning_rate": 2.4061610645334975e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 2248452.0,
      "step": 5147
    },
    {
      "epoch": 0.3959695408045535,
      "grad_norm": 5.875,
      "learning_rate": 2.4060456887931698e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2248742.0,
      "step": 5148
    },
    {
      "epoch": 0.39604645796477195,
      "grad_norm": 7.0,
      "learning_rate": 2.405930313052842e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2249074.0,
      "step": 5149
    },
    {
      "epoch": 0.3961233751249904,
      "grad_norm": 4.90625,
      "learning_rate": 2.4058149373125143e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 2249390.0,
      "step": 5150
    },
    {
      "epoch": 0.3962002922852088,
      "grad_norm": 5.375,
      "learning_rate": 2.405699561572187e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2249672.0,
      "step": 5151
    },
    {
      "epoch": 0.39627720944542727,
      "grad_norm": 2.578125,
      "learning_rate": 2.405584185831859e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 2250470.0,
      "step": 5152
    },
    {
      "epoch": 0.39635412660564573,
      "grad_norm": 6.03125,
      "learning_rate": 2.4054688100915315e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2250756.0,
      "step": 5153
    },
    {
      "epoch": 0.3964310437658642,
      "grad_norm": 5.78125,
      "learning_rate": 2.4053534343512037e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2251116.0,
      "step": 5154
    },
    {
      "epoch": 0.3965079609260826,
      "grad_norm": 2.03125,
      "learning_rate": 2.405238058610876e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 2251926.0,
      "step": 5155
    },
    {
      "epoch": 0.39658487808630105,
      "grad_norm": 3.71875,
      "learning_rate": 2.4051226828705487e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 2252214.0,
      "step": 5156
    },
    {
      "epoch": 0.3966617952465195,
      "grad_norm": 2.921875,
      "learning_rate": 2.405007307130221e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2252794.0,
      "step": 5157
    },
    {
      "epoch": 0.39673871240673797,
      "grad_norm": 3.75,
      "learning_rate": 2.404891931389893e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9966158866882324,
      "num_tokens": 2253386.0,
      "step": 5158
    },
    {
      "epoch": 0.3968156295669564,
      "grad_norm": 6.78125,
      "learning_rate": 2.4047765556495654e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2253714.0,
      "step": 5159
    },
    {
      "epoch": 0.39689254672717483,
      "grad_norm": 1.9375,
      "learning_rate": 2.4046611799092377e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2254288.0,
      "step": 5160
    },
    {
      "epoch": 0.3969694638873933,
      "grad_norm": 4.15625,
      "learning_rate": 2.40454580416891e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2254634.0,
      "step": 5161
    },
    {
      "epoch": 0.3970463810476117,
      "grad_norm": 2.71875,
      "learning_rate": 2.4044304284285826e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2255186.0,
      "step": 5162
    },
    {
      "epoch": 0.39712329820783016,
      "grad_norm": 2.578125,
      "learning_rate": 2.404315052688255e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2255770.0,
      "step": 5163
    },
    {
      "epoch": 0.3972002153680486,
      "grad_norm": 7.46875,
      "learning_rate": 2.404199676947927e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9771241545677185,
      "num_tokens": 2256078.0,
      "step": 5164
    },
    {
      "epoch": 0.3972771325282671,
      "grad_norm": 3.703125,
      "learning_rate": 2.4040843012075994e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 2256412.0,
      "step": 5165
    },
    {
      "epoch": 0.3973540496884855,
      "grad_norm": 7.96875,
      "learning_rate": 2.4039689254672717e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2256744.0,
      "step": 5166
    },
    {
      "epoch": 0.39743096684870394,
      "grad_norm": 5.28125,
      "learning_rate": 2.4038535497269444e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2257038.0,
      "step": 5167
    },
    {
      "epoch": 0.3975078840089224,
      "grad_norm": 2.96875,
      "learning_rate": 2.4037381739866166e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 2257560.0,
      "step": 5168
    },
    {
      "epoch": 0.39758480116914086,
      "grad_norm": 2.65625,
      "learning_rate": 2.403622798246289e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2258106.0,
      "step": 5169
    },
    {
      "epoch": 0.39766171832935926,
      "grad_norm": 2.890625,
      "learning_rate": 2.403507422505961e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 2258674.0,
      "step": 5170
    },
    {
      "epoch": 0.3977386354895777,
      "grad_norm": 6.84375,
      "learning_rate": 2.4033920467656334e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 2258944.0,
      "step": 5171
    },
    {
      "epoch": 0.3978155526497962,
      "grad_norm": 5.96875,
      "learning_rate": 2.4032766710253056e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2259244.0,
      "step": 5172
    },
    {
      "epoch": 0.39789246981001464,
      "grad_norm": 2.53125,
      "learning_rate": 2.4031612952849783e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2259788.0,
      "step": 5173
    },
    {
      "epoch": 0.39796938697023304,
      "grad_norm": 2.015625,
      "learning_rate": 2.4030459195446506e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2260334.0,
      "step": 5174
    },
    {
      "epoch": 0.3980463041304515,
      "grad_norm": 1.90625,
      "learning_rate": 2.402930543804323e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2261114.0,
      "step": 5175
    },
    {
      "epoch": 0.39812322129066996,
      "grad_norm": 4.3125,
      "learning_rate": 2.402815168063995e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2261430.0,
      "step": 5176
    },
    {
      "epoch": 0.3982001384508884,
      "grad_norm": 4.03125,
      "learning_rate": 2.4026997923236674e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2261736.0,
      "step": 5177
    },
    {
      "epoch": 0.3982770556111068,
      "grad_norm": 5.5,
      "learning_rate": 2.4025844165833396e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2262058.0,
      "step": 5178
    },
    {
      "epoch": 0.3983539727713253,
      "grad_norm": 5.09375,
      "learning_rate": 2.4024690408430123e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2262342.0,
      "step": 5179
    },
    {
      "epoch": 0.39843088993154374,
      "grad_norm": 8.125,
      "learning_rate": 2.4023536651026846e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2262672.0,
      "step": 5180
    },
    {
      "epoch": 0.39850780709176215,
      "grad_norm": 3.71875,
      "learning_rate": 2.402238289362357e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 2263164.0,
      "step": 5181
    },
    {
      "epoch": 0.3985847242519806,
      "grad_norm": 3.671875,
      "learning_rate": 2.402122913622029e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2263690.0,
      "step": 5182
    },
    {
      "epoch": 0.39866164141219906,
      "grad_norm": 3.359375,
      "learning_rate": 2.4020075378817013e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2264232.0,
      "step": 5183
    },
    {
      "epoch": 0.3987385585724175,
      "grad_norm": 2.65625,
      "learning_rate": 2.401892162141374e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9959431886672974,
      "num_tokens": 2264726.0,
      "step": 5184
    },
    {
      "epoch": 0.3988154757326359,
      "grad_norm": 3.46875,
      "learning_rate": 2.4017767864010463e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2265042.0,
      "step": 5185
    },
    {
      "epoch": 0.3988923928928544,
      "grad_norm": 10.5625,
      "learning_rate": 2.401661410660718e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9827089309692383,
      "num_tokens": 2265390.0,
      "step": 5186
    },
    {
      "epoch": 0.39896931005307285,
      "grad_norm": 4.6875,
      "learning_rate": 2.4015460349203908e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9925650358200073,
      "num_tokens": 2265660.0,
      "step": 5187
    },
    {
      "epoch": 0.3990462272132913,
      "grad_norm": 3.234375,
      "learning_rate": 2.401430659180063e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2266156.0,
      "step": 5188
    },
    {
      "epoch": 0.3991231443735097,
      "grad_norm": 6.40625,
      "learning_rate": 2.4013152834397353e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 2266734.0,
      "step": 5189
    },
    {
      "epoch": 0.39920006153372817,
      "grad_norm": 6.1875,
      "learning_rate": 2.401199907699408e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 2267282.0,
      "step": 5190
    },
    {
      "epoch": 0.3992769786939466,
      "grad_norm": 2.75,
      "learning_rate": 2.4010845319590802e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 2268090.0,
      "step": 5191
    },
    {
      "epoch": 0.3993538958541651,
      "grad_norm": 4.4375,
      "learning_rate": 2.4009691562187525e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 2268372.0,
      "step": 5192
    },
    {
      "epoch": 0.3994308130143835,
      "grad_norm": 3.625,
      "learning_rate": 2.4008537804784248e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2268936.0,
      "step": 5193
    },
    {
      "epoch": 0.39950773017460195,
      "grad_norm": 10.1875,
      "learning_rate": 2.400738404738097e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9834254384040833,
      "num_tokens": 2269300.0,
      "step": 5194
    },
    {
      "epoch": 0.3995846473348204,
      "grad_norm": 9.5625,
      "learning_rate": 2.4006230289977697e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 2269608.0,
      "step": 5195
    },
    {
      "epoch": 0.39966156449503887,
      "grad_norm": 8.0625,
      "learning_rate": 2.400507653257442e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 2269936.0,
      "step": 5196
    },
    {
      "epoch": 0.39973848165525727,
      "grad_norm": 4.3125,
      "learning_rate": 2.400392277517114e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 2270236.0,
      "step": 5197
    },
    {
      "epoch": 0.39981539881547573,
      "grad_norm": 3.03125,
      "learning_rate": 2.4002769017767865e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2270732.0,
      "step": 5198
    },
    {
      "epoch": 0.3998923159756942,
      "grad_norm": 4.65625,
      "learning_rate": 2.4001615260364587e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2271056.0,
      "step": 5199
    },
    {
      "epoch": 0.39996923313591265,
      "grad_norm": 4.5625,
      "learning_rate": 2.400046150296131e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 2271366.0,
      "step": 5200
    },
    {
      "epoch": 0.40004615029613105,
      "grad_norm": 7.21875,
      "learning_rate": 2.3999307745558037e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2271646.0,
      "step": 5201
    },
    {
      "epoch": 0.4001230674563495,
      "grad_norm": 2.8125,
      "learning_rate": 2.399815398815476e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2272226.0,
      "step": 5202
    },
    {
      "epoch": 0.40019998461656797,
      "grad_norm": 7.21875,
      "learning_rate": 2.399700023075148e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.992094874382019,
      "num_tokens": 2272480.0,
      "step": 5203
    },
    {
      "epoch": 0.4002769017767864,
      "grad_norm": 7.5,
      "learning_rate": 2.3995846473348204e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 2272782.0,
      "step": 5204
    },
    {
      "epoch": 0.40035381893700484,
      "grad_norm": 4.46875,
      "learning_rate": 2.3994692715944927e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 2273078.0,
      "step": 5205
    },
    {
      "epoch": 0.4004307360972233,
      "grad_norm": 6.1875,
      "learning_rate": 2.3993538958541654e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 2273416.0,
      "step": 5206
    },
    {
      "epoch": 0.40050765325744175,
      "grad_norm": 5.46875,
      "learning_rate": 2.3992385201138376e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9895833134651184,
      "num_tokens": 2273706.0,
      "step": 5207
    },
    {
      "epoch": 0.40058457041766016,
      "grad_norm": 6.03125,
      "learning_rate": 2.3991231443735095e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2273984.0,
      "step": 5208
    },
    {
      "epoch": 0.4006614875778786,
      "grad_norm": 5.3125,
      "learning_rate": 2.399007768633182e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2274330.0,
      "step": 5209
    },
    {
      "epoch": 0.4007384047380971,
      "grad_norm": 3.171875,
      "learning_rate": 2.3988923928928544e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 2274902.0,
      "step": 5210
    },
    {
      "epoch": 0.40081532189831554,
      "grad_norm": 2.171875,
      "learning_rate": 2.3987770171525267e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2275448.0,
      "step": 5211
    },
    {
      "epoch": 0.40089223905853394,
      "grad_norm": 6.59375,
      "learning_rate": 2.3986616414121994e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 2275708.0,
      "step": 5212
    },
    {
      "epoch": 0.4009691562187524,
      "grad_norm": 7.5,
      "learning_rate": 2.398546265671871e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2276070.0,
      "step": 5213
    },
    {
      "epoch": 0.40104607337897086,
      "grad_norm": 5.28125,
      "learning_rate": 2.398430889931544e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2276384.0,
      "step": 5214
    },
    {
      "epoch": 0.4011229905391893,
      "grad_norm": 6.1875,
      "learning_rate": 2.398315514191216e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 2276724.0,
      "step": 5215
    },
    {
      "epoch": 0.4011999076994077,
      "grad_norm": 7.59375,
      "learning_rate": 2.3982001384508884e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2277080.0,
      "step": 5216
    },
    {
      "epoch": 0.4012768248596262,
      "grad_norm": 3.53125,
      "learning_rate": 2.398084762710561e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2277656.0,
      "step": 5217
    },
    {
      "epoch": 0.40135374201984464,
      "grad_norm": 3.5,
      "learning_rate": 2.3979693869702333e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2278176.0,
      "step": 5218
    },
    {
      "epoch": 0.4014306591800631,
      "grad_norm": 7.125,
      "learning_rate": 2.397854011229905e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 2278476.0,
      "step": 5219
    },
    {
      "epoch": 0.4015075763402815,
      "grad_norm": 5.75,
      "learning_rate": 2.397738635489578e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2278828.0,
      "step": 5220
    },
    {
      "epoch": 0.40158449350049996,
      "grad_norm": 7.3125,
      "learning_rate": 2.39762325974925e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 2279128.0,
      "step": 5221
    },
    {
      "epoch": 0.4016614106607184,
      "grad_norm": 2.453125,
      "learning_rate": 2.3975078840089224e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2279948.0,
      "step": 5222
    },
    {
      "epoch": 0.4017383278209368,
      "grad_norm": 2.375,
      "learning_rate": 2.397392508268595e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2280506.0,
      "step": 5223
    },
    {
      "epoch": 0.4018152449811553,
      "grad_norm": 2.296875,
      "learning_rate": 2.397277132528267e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2281064.0,
      "step": 5224
    },
    {
      "epoch": 0.40189216214137374,
      "grad_norm": 3.890625,
      "learning_rate": 2.3971617567879396e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2281660.0,
      "step": 5225
    },
    {
      "epoch": 0.4019690793015922,
      "grad_norm": 3.515625,
      "learning_rate": 2.397046381047612e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2282234.0,
      "step": 5226
    },
    {
      "epoch": 0.4020459964618106,
      "grad_norm": 2.859375,
      "learning_rate": 2.396931005307284e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 2283070.0,
      "step": 5227
    },
    {
      "epoch": 0.40212291362202907,
      "grad_norm": 5.40625,
      "learning_rate": 2.3968156295669568e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2283406.0,
      "step": 5228
    },
    {
      "epoch": 0.4021998307822475,
      "grad_norm": 1.984375,
      "learning_rate": 2.396700253826629e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 2284002.0,
      "step": 5229
    },
    {
      "epoch": 0.402276747942466,
      "grad_norm": 6.03125,
      "learning_rate": 2.396584878086301e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2284318.0,
      "step": 5230
    },
    {
      "epoch": 0.4023536651026844,
      "grad_norm": 3.1875,
      "learning_rate": 2.3964695023459735e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 2284680.0,
      "step": 5231
    },
    {
      "epoch": 0.40243058226290285,
      "grad_norm": 9.9375,
      "learning_rate": 2.396354126605646e-06,
      "loss": 0.0398,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2284964.0,
      "step": 5232
    },
    {
      "epoch": 0.4025074994231213,
      "grad_norm": 2.09375,
      "learning_rate": 2.396238750865318e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 2285760.0,
      "step": 5233
    },
    {
      "epoch": 0.40258441658333977,
      "grad_norm": 3.046875,
      "learning_rate": 2.3961233751249907e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 2286548.0,
      "step": 5234
    },
    {
      "epoch": 0.40266133374355817,
      "grad_norm": 5.15625,
      "learning_rate": 2.3960079993846626e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2287090.0,
      "step": 5235
    },
    {
      "epoch": 0.40273825090377663,
      "grad_norm": 13.4375,
      "learning_rate": 2.3958926236443352e-06,
      "loss": 0.0509,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2287344.0,
      "step": 5236
    },
    {
      "epoch": 0.4028151680639951,
      "grad_norm": 10.6875,
      "learning_rate": 2.3957772479040075e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2287644.0,
      "step": 5237
    },
    {
      "epoch": 0.40289208522421355,
      "grad_norm": 2.75,
      "learning_rate": 2.3956618721636798e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2288402.0,
      "step": 5238
    },
    {
      "epoch": 0.40296900238443195,
      "grad_norm": 4.25,
      "learning_rate": 2.395546496423352e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2288922.0,
      "step": 5239
    },
    {
      "epoch": 0.4030459195446504,
      "grad_norm": 6.3125,
      "learning_rate": 2.3954311206830243e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 2289256.0,
      "step": 5240
    },
    {
      "epoch": 0.40312283670486887,
      "grad_norm": 4.75,
      "learning_rate": 2.3953157449426965e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2289556.0,
      "step": 5241
    },
    {
      "epoch": 0.4031997538650873,
      "grad_norm": 6.0625,
      "learning_rate": 2.3952003692023692e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2289876.0,
      "step": 5242
    },
    {
      "epoch": 0.40327667102530573,
      "grad_norm": 4.3125,
      "learning_rate": 2.3950849934620415e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2290200.0,
      "step": 5243
    },
    {
      "epoch": 0.4033535881855242,
      "grad_norm": 3.921875,
      "learning_rate": 2.3949696177217137e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2290522.0,
      "step": 5244
    },
    {
      "epoch": 0.40343050534574265,
      "grad_norm": 8.1875,
      "learning_rate": 2.3948542419813864e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 2290860.0,
      "step": 5245
    },
    {
      "epoch": 0.40350742250596106,
      "grad_norm": 5.6875,
      "learning_rate": 2.3947388662410582e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2291178.0,
      "step": 5246
    },
    {
      "epoch": 0.4035843396661795,
      "grad_norm": 5.21875,
      "learning_rate": 2.3946234905007305e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2291488.0,
      "step": 5247
    },
    {
      "epoch": 0.403661256826398,
      "grad_norm": 2.140625,
      "learning_rate": 2.394508114760403e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2292044.0,
      "step": 5248
    },
    {
      "epoch": 0.40373817398661643,
      "grad_norm": 2.1875,
      "learning_rate": 2.3943927390200754e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2292600.0,
      "step": 5249
    },
    {
      "epoch": 0.40381509114683484,
      "grad_norm": 4.0625,
      "learning_rate": 2.3942773632797477e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2293174.0,
      "step": 5250
    },
    {
      "epoch": 0.4038920083070533,
      "grad_norm": 5.65625,
      "learning_rate": 2.39416198753942e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2293488.0,
      "step": 5251
    },
    {
      "epoch": 0.40396892546727176,
      "grad_norm": 2.78125,
      "learning_rate": 2.3940466117990922e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2294074.0,
      "step": 5252
    },
    {
      "epoch": 0.4040458426274902,
      "grad_norm": 3.875,
      "learning_rate": 2.393931236058765e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2294390.0,
      "step": 5253
    },
    {
      "epoch": 0.4041227597877086,
      "grad_norm": 3.8125,
      "learning_rate": 2.393815860318437e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 2294736.0,
      "step": 5254
    },
    {
      "epoch": 0.4041996769479271,
      "grad_norm": 1.6953125,
      "learning_rate": 2.3937004845781094e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2295292.0,
      "step": 5255
    },
    {
      "epoch": 0.40427659410814554,
      "grad_norm": 1.984375,
      "learning_rate": 2.393585108837782e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2295854.0,
      "step": 5256
    },
    {
      "epoch": 0.404353511268364,
      "grad_norm": 7.09375,
      "learning_rate": 2.393469733097454e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 2296182.0,
      "step": 5257
    },
    {
      "epoch": 0.4044304284285824,
      "grad_norm": 3.265625,
      "learning_rate": 2.393354357357126e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2296762.0,
      "step": 5258
    },
    {
      "epoch": 0.40450734558880086,
      "grad_norm": 4.46875,
      "learning_rate": 2.393238981616799e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2297114.0,
      "step": 5259
    },
    {
      "epoch": 0.4045842627490193,
      "grad_norm": 3.390625,
      "learning_rate": 2.393123605876471e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 2297658.0,
      "step": 5260
    },
    {
      "epoch": 0.4046611799092378,
      "grad_norm": 2.390625,
      "learning_rate": 2.3930082301361434e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2298440.0,
      "step": 5261
    },
    {
      "epoch": 0.4047380970694562,
      "grad_norm": 7.21875,
      "learning_rate": 2.3928928543958156e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2298718.0,
      "step": 5262
    },
    {
      "epoch": 0.40481501422967464,
      "grad_norm": 6.28125,
      "learning_rate": 2.392777478655488e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 2299006.0,
      "step": 5263
    },
    {
      "epoch": 0.4048919313898931,
      "grad_norm": 5.375,
      "learning_rate": 2.3926621029151606e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2299300.0,
      "step": 5264
    },
    {
      "epoch": 0.4049688485501115,
      "grad_norm": 4.75,
      "learning_rate": 2.392546727174833e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 2299636.0,
      "step": 5265
    },
    {
      "epoch": 0.40504576571032996,
      "grad_norm": 6.4375,
      "learning_rate": 2.392431351434505e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 2299964.0,
      "step": 5266
    },
    {
      "epoch": 0.4051226828705484,
      "grad_norm": 2.109375,
      "learning_rate": 2.3923159756941778e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2300520.0,
      "step": 5267
    },
    {
      "epoch": 0.4051996000307669,
      "grad_norm": 4.1875,
      "learning_rate": 2.3922005999538496e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2300882.0,
      "step": 5268
    },
    {
      "epoch": 0.4052765171909853,
      "grad_norm": 6.53125,
      "learning_rate": 2.392085224213522e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9969419240951538,
      "num_tokens": 2301210.0,
      "step": 5269
    },
    {
      "epoch": 0.40535343435120375,
      "grad_norm": 3.046875,
      "learning_rate": 2.3919698484731946e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2301790.0,
      "step": 5270
    },
    {
      "epoch": 0.4054303515114222,
      "grad_norm": 2.140625,
      "learning_rate": 2.391854472732867e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 2302600.0,
      "step": 5271
    },
    {
      "epoch": 0.40550726867164066,
      "grad_norm": 1.8671875,
      "learning_rate": 2.391739096992539e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 2303420.0,
      "step": 5272
    },
    {
      "epoch": 0.40558418583185907,
      "grad_norm": 5.09375,
      "learning_rate": 2.3916237212522113e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 2303756.0,
      "step": 5273
    },
    {
      "epoch": 0.4056611029920775,
      "grad_norm": 4.78125,
      "learning_rate": 2.3915083455118836e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2304092.0,
      "step": 5274
    },
    {
      "epoch": 0.405738020152296,
      "grad_norm": 5.9375,
      "learning_rate": 2.3913929697715563e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2304418.0,
      "step": 5275
    },
    {
      "epoch": 0.40581493731251445,
      "grad_norm": 3.65625,
      "learning_rate": 2.3912775940312285e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2304970.0,
      "step": 5276
    },
    {
      "epoch": 0.40589185447273285,
      "grad_norm": 9.5,
      "learning_rate": 2.391162218290901e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2305226.0,
      "step": 5277
    },
    {
      "epoch": 0.4059687716329513,
      "grad_norm": 3.765625,
      "learning_rate": 2.391046842550573e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 2305818.0,
      "step": 5278
    },
    {
      "epoch": 0.40604568879316977,
      "grad_norm": 3.03125,
      "learning_rate": 2.3909314668102453e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2306140.0,
      "step": 5279
    },
    {
      "epoch": 0.4061226059533882,
      "grad_norm": 7.1875,
      "learning_rate": 2.3908160910699176e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 2306426.0,
      "step": 5280
    },
    {
      "epoch": 0.40619952311360663,
      "grad_norm": 5.3125,
      "learning_rate": 2.3907007153295902e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2306736.0,
      "step": 5281
    },
    {
      "epoch": 0.4062764402738251,
      "grad_norm": 3.578125,
      "learning_rate": 2.3905853395892625e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2307272.0,
      "step": 5282
    },
    {
      "epoch": 0.40635335743404355,
      "grad_norm": 6.25,
      "learning_rate": 2.3904699638489348e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 2307572.0,
      "step": 5283
    },
    {
      "epoch": 0.40643027459426195,
      "grad_norm": 7.8125,
      "learning_rate": 2.390354588108607e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 2307834.0,
      "step": 5284
    },
    {
      "epoch": 0.4065071917544804,
      "grad_norm": 1.9453125,
      "learning_rate": 2.3902392123682793e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 2308414.0,
      "step": 5285
    },
    {
      "epoch": 0.40658410891469887,
      "grad_norm": 8.5,
      "learning_rate": 2.390123836627952e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2308670.0,
      "step": 5286
    },
    {
      "epoch": 0.40666102607491733,
      "grad_norm": 9.375,
      "learning_rate": 2.3900084608876242e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 2308934.0,
      "step": 5287
    },
    {
      "epoch": 0.40673794323513573,
      "grad_norm": 6.03125,
      "learning_rate": 2.3898930851472965e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2309216.0,
      "step": 5288
    },
    {
      "epoch": 0.4068148603953542,
      "grad_norm": 1.8984375,
      "learning_rate": 2.3897777094069687e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 2310012.0,
      "step": 5289
    },
    {
      "epoch": 0.40689177755557265,
      "grad_norm": 4.5625,
      "learning_rate": 2.389662333666641e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 2310282.0,
      "step": 5290
    },
    {
      "epoch": 0.4069686947157911,
      "grad_norm": 3.75,
      "learning_rate": 2.3895469579263133e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 2310608.0,
      "step": 5291
    },
    {
      "epoch": 0.4070456118760095,
      "grad_norm": 3.59375,
      "learning_rate": 2.389431582185986e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9913644194602966,
      "num_tokens": 2311188.0,
      "step": 5292
    },
    {
      "epoch": 0.407122529036228,
      "grad_norm": 3.828125,
      "learning_rate": 2.389316206445658e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2311510.0,
      "step": 5293
    },
    {
      "epoch": 0.40719944619644644,
      "grad_norm": 4.5625,
      "learning_rate": 2.3892008307053304e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 2311820.0,
      "step": 5294
    },
    {
      "epoch": 0.4072763633566649,
      "grad_norm": 2.234375,
      "learning_rate": 2.3890854549650027e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2312368.0,
      "step": 5295
    },
    {
      "epoch": 0.4073532805168833,
      "grad_norm": 3.8125,
      "learning_rate": 2.388970079224675e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2312652.0,
      "step": 5296
    },
    {
      "epoch": 0.40743019767710176,
      "grad_norm": 5.25,
      "learning_rate": 2.3888547034843472e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2312968.0,
      "step": 5297
    },
    {
      "epoch": 0.4075071148373202,
      "grad_norm": 3.609375,
      "learning_rate": 2.38873932774402e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2313252.0,
      "step": 5298
    },
    {
      "epoch": 0.4075840319975387,
      "grad_norm": 5.84375,
      "learning_rate": 2.388623952003692e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9792284965515137,
      "num_tokens": 2313590.0,
      "step": 5299
    },
    {
      "epoch": 0.4076609491577571,
      "grad_norm": 4.75,
      "learning_rate": 2.3885085762633644e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2313896.0,
      "step": 5300
    },
    {
      "epoch": 0.40773786631797554,
      "grad_norm": 2.65625,
      "learning_rate": 2.3883932005230367e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9961758852005005,
      "num_tokens": 2314420.0,
      "step": 5301
    },
    {
      "epoch": 0.407814783478194,
      "grad_norm": 5.15625,
      "learning_rate": 2.388277824782709e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2314712.0,
      "step": 5302
    },
    {
      "epoch": 0.4078917006384124,
      "grad_norm": 3.546875,
      "learning_rate": 2.3881624490423816e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2315282.0,
      "step": 5303
    },
    {
      "epoch": 0.40796861779863086,
      "grad_norm": 2.421875,
      "learning_rate": 2.388047073302054e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 2316106.0,
      "step": 5304
    },
    {
      "epoch": 0.4080455349588493,
      "grad_norm": 2.671875,
      "learning_rate": 2.387931697561726e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2316690.0,
      "step": 5305
    },
    {
      "epoch": 0.4081224521190678,
      "grad_norm": 2.1875,
      "learning_rate": 2.3878163218213984e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2317236.0,
      "step": 5306
    },
    {
      "epoch": 0.4081993692792862,
      "grad_norm": 5.5,
      "learning_rate": 2.3877009460810707e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 2317576.0,
      "step": 5307
    },
    {
      "epoch": 0.40827628643950464,
      "grad_norm": 3.25,
      "learning_rate": 2.387585570340743e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2317864.0,
      "step": 5308
    },
    {
      "epoch": 0.4083532035997231,
      "grad_norm": 3.078125,
      "learning_rate": 2.3874701946004156e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 2318406.0,
      "step": 5309
    },
    {
      "epoch": 0.40843012075994156,
      "grad_norm": 4.9375,
      "learning_rate": 2.387354818860088e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2318754.0,
      "step": 5310
    },
    {
      "epoch": 0.40850703792015997,
      "grad_norm": 6.3125,
      "learning_rate": 2.38723944311976e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2319012.0,
      "step": 5311
    },
    {
      "epoch": 0.4085839550803784,
      "grad_norm": 4.65625,
      "learning_rate": 2.3871240673794324e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 2319304.0,
      "step": 5312
    },
    {
      "epoch": 0.4086608722405969,
      "grad_norm": 2.734375,
      "learning_rate": 2.3870086916391046e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 2319872.0,
      "step": 5313
    },
    {
      "epoch": 0.40873778940081534,
      "grad_norm": 2.359375,
      "learning_rate": 2.3868933158987773e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2320424.0,
      "step": 5314
    },
    {
      "epoch": 0.40881470656103375,
      "grad_norm": 4.875,
      "learning_rate": 2.3867779401584496e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 2320764.0,
      "step": 5315
    },
    {
      "epoch": 0.4088916237212522,
      "grad_norm": 4.84375,
      "learning_rate": 2.3866625644181214e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 2321092.0,
      "step": 5316
    },
    {
      "epoch": 0.40896854088147067,
      "grad_norm": 4.03125,
      "learning_rate": 2.386547188677794e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2321414.0,
      "step": 5317
    },
    {
      "epoch": 0.4090454580416891,
      "grad_norm": 2.484375,
      "learning_rate": 2.3864318129374663e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2321960.0,
      "step": 5318
    },
    {
      "epoch": 0.40912237520190753,
      "grad_norm": 2.4375,
      "learning_rate": 2.3863164371971386e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 2322770.0,
      "step": 5319
    },
    {
      "epoch": 0.409199292362126,
      "grad_norm": 1.859375,
      "learning_rate": 2.3862010614568113e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2323550.0,
      "step": 5320
    },
    {
      "epoch": 0.40927620952234445,
      "grad_norm": 3.4375,
      "learning_rate": 2.3860856857164835e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2324130.0,
      "step": 5321
    },
    {
      "epoch": 0.4093531266825629,
      "grad_norm": 8.6875,
      "learning_rate": 2.385970309976156e-06,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 2324410.0,
      "step": 5322
    },
    {
      "epoch": 0.4094300438427813,
      "grad_norm": 5.6875,
      "learning_rate": 2.385854934235828e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2324704.0,
      "step": 5323
    },
    {
      "epoch": 0.40950696100299977,
      "grad_norm": 2.6875,
      "learning_rate": 2.3857395584955003e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2325252.0,
      "step": 5324
    },
    {
      "epoch": 0.40958387816321823,
      "grad_norm": 4.71875,
      "learning_rate": 2.385624182755173e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2325550.0,
      "step": 5325
    },
    {
      "epoch": 0.40966079532343663,
      "grad_norm": 1.8203125,
      "learning_rate": 2.3855088070148452e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2326094.0,
      "step": 5326
    },
    {
      "epoch": 0.4097377124836551,
      "grad_norm": 3.734375,
      "learning_rate": 2.385393431274517e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2326416.0,
      "step": 5327
    },
    {
      "epoch": 0.40981462964387355,
      "grad_norm": 2.046875,
      "learning_rate": 2.3852780555341898e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2326964.0,
      "step": 5328
    },
    {
      "epoch": 0.409891546804092,
      "grad_norm": 7.125,
      "learning_rate": 2.385162679793862e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 2327226.0,
      "step": 5329
    },
    {
      "epoch": 0.4099684639643104,
      "grad_norm": 7.3125,
      "learning_rate": 2.3850473040535343e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2327530.0,
      "step": 5330
    },
    {
      "epoch": 0.4100453811245289,
      "grad_norm": 2.140625,
      "learning_rate": 2.384931928313207e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2328072.0,
      "step": 5331
    },
    {
      "epoch": 0.41012229828474733,
      "grad_norm": 4.5,
      "learning_rate": 2.3848165525728792e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 2328360.0,
      "step": 5332
    },
    {
      "epoch": 0.4101992154449658,
      "grad_norm": 3.421875,
      "learning_rate": 2.3847011768325515e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2328940.0,
      "step": 5333
    },
    {
      "epoch": 0.4102761326051842,
      "grad_norm": 1.9375,
      "learning_rate": 2.3845858010922237e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2329486.0,
      "step": 5334
    },
    {
      "epoch": 0.41035304976540266,
      "grad_norm": 4.0625,
      "learning_rate": 2.384470425351896e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2329802.0,
      "step": 5335
    },
    {
      "epoch": 0.4104299669256211,
      "grad_norm": 2.21875,
      "learning_rate": 2.3843550496115687e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2330358.0,
      "step": 5336
    },
    {
      "epoch": 0.4105068840858396,
      "grad_norm": 6.09375,
      "learning_rate": 2.384239673871241e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2330710.0,
      "step": 5337
    },
    {
      "epoch": 0.410583801246058,
      "grad_norm": 6.71875,
      "learning_rate": 2.3841242981309128e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 2330966.0,
      "step": 5338
    },
    {
      "epoch": 0.41066071840627644,
      "grad_norm": 2.265625,
      "learning_rate": 2.3840089223905855e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2331484.0,
      "step": 5339
    },
    {
      "epoch": 0.4107376355664949,
      "grad_norm": 2.3125,
      "learning_rate": 2.3838935466502577e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2332068.0,
      "step": 5340
    },
    {
      "epoch": 0.41081455272671336,
      "grad_norm": 2.8125,
      "learning_rate": 2.38377817090993e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 2332608.0,
      "step": 5341
    },
    {
      "epoch": 0.41089146988693176,
      "grad_norm": 1.953125,
      "learning_rate": 2.3836627951696026e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2333418.0,
      "step": 5342
    },
    {
      "epoch": 0.4109683870471502,
      "grad_norm": 5.875,
      "learning_rate": 2.3835474194292745e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 2333754.0,
      "step": 5343
    },
    {
      "epoch": 0.4110453042073687,
      "grad_norm": 4.5,
      "learning_rate": 2.383432043688947e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 2334052.0,
      "step": 5344
    },
    {
      "epoch": 0.4111222213675871,
      "grad_norm": 11.4375,
      "learning_rate": 2.3833166679486194e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.976190447807312,
      "num_tokens": 2334392.0,
      "step": 5345
    },
    {
      "epoch": 0.41119913852780554,
      "grad_norm": 7.0625,
      "learning_rate": 2.3832012922082917e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 2334700.0,
      "step": 5346
    },
    {
      "epoch": 0.411276055688024,
      "grad_norm": 6.4375,
      "learning_rate": 2.383085916467964e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 2334956.0,
      "step": 5347
    },
    {
      "epoch": 0.41135297284824246,
      "grad_norm": 8.6875,
      "learning_rate": 2.3829705407276366e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2335306.0,
      "step": 5348
    },
    {
      "epoch": 0.41142989000846086,
      "grad_norm": 4.875,
      "learning_rate": 2.3828551649873085e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 2335646.0,
      "step": 5349
    },
    {
      "epoch": 0.4115068071686793,
      "grad_norm": 2.1875,
      "learning_rate": 2.382739789246981e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2336220.0,
      "step": 5350
    },
    {
      "epoch": 0.4115837243288978,
      "grad_norm": 2.0625,
      "learning_rate": 2.3826244135066534e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 2337006.0,
      "step": 5351
    },
    {
      "epoch": 0.41166064148911624,
      "grad_norm": 7.15625,
      "learning_rate": 2.3825090377663257e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2337336.0,
      "step": 5352
    },
    {
      "epoch": 0.41173755864933465,
      "grad_norm": 6.4375,
      "learning_rate": 2.3823936620259983e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2337638.0,
      "step": 5353
    },
    {
      "epoch": 0.4118144758095531,
      "grad_norm": 4.0,
      "learning_rate": 2.38227828628567e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 2337982.0,
      "step": 5354
    },
    {
      "epoch": 0.41189139296977156,
      "grad_norm": 6.09375,
      "learning_rate": 2.3821629105453424e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2338344.0,
      "step": 5355
    },
    {
      "epoch": 0.41196831012999,
      "grad_norm": 3.921875,
      "learning_rate": 2.382047534805015e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 2338936.0,
      "step": 5356
    },
    {
      "epoch": 0.4120452272902084,
      "grad_norm": 2.140625,
      "learning_rate": 2.3819321590646874e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2339484.0,
      "step": 5357
    },
    {
      "epoch": 0.4121221444504269,
      "grad_norm": 6.34375,
      "learning_rate": 2.3818167833243596e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 2339824.0,
      "step": 5358
    },
    {
      "epoch": 0.41219906161064535,
      "grad_norm": 7.21875,
      "learning_rate": 2.3817014075840323e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2340128.0,
      "step": 5359
    },
    {
      "epoch": 0.4122759787708638,
      "grad_norm": 3.171875,
      "learning_rate": 2.381586031843704e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2340674.0,
      "step": 5360
    },
    {
      "epoch": 0.4123528959310822,
      "grad_norm": 7.09375,
      "learning_rate": 2.381470656103377e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2340956.0,
      "step": 5361
    },
    {
      "epoch": 0.41242981309130067,
      "grad_norm": 3.359375,
      "learning_rate": 2.381355280363049e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2341302.0,
      "step": 5362
    },
    {
      "epoch": 0.4125067302515191,
      "grad_norm": 5.8125,
      "learning_rate": 2.3812399046227213e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 2341598.0,
      "step": 5363
    },
    {
      "epoch": 0.4125836474117376,
      "grad_norm": 4.96875,
      "learning_rate": 2.381124528882394e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2341914.0,
      "step": 5364
    },
    {
      "epoch": 0.412660564571956,
      "grad_norm": 2.15625,
      "learning_rate": 2.381009153142066e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 2342474.0,
      "step": 5365
    },
    {
      "epoch": 0.41273748173217445,
      "grad_norm": 1.96875,
      "learning_rate": 2.380893777401738e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2343026.0,
      "step": 5366
    },
    {
      "epoch": 0.4128143988923929,
      "grad_norm": 6.34375,
      "learning_rate": 2.380778401661411e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2343338.0,
      "step": 5367
    },
    {
      "epoch": 0.4128913160526113,
      "grad_norm": 5.125,
      "learning_rate": 2.380663025921083e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 2343638.0,
      "step": 5368
    },
    {
      "epoch": 0.41296823321282977,
      "grad_norm": 5.875,
      "learning_rate": 2.3805476501807553e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 2343948.0,
      "step": 5369
    },
    {
      "epoch": 0.41304515037304823,
      "grad_norm": 3.96875,
      "learning_rate": 2.380432274440428e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2344270.0,
      "step": 5370
    },
    {
      "epoch": 0.4131220675332667,
      "grad_norm": 5.5,
      "learning_rate": 2.3803168987001e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2344622.0,
      "step": 5371
    },
    {
      "epoch": 0.4131989846934851,
      "grad_norm": 2.671875,
      "learning_rate": 2.3802015229597725e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2345190.0,
      "step": 5372
    },
    {
      "epoch": 0.41327590185370355,
      "grad_norm": 2.65625,
      "learning_rate": 2.3800861472194448e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2345708.0,
      "step": 5373
    },
    {
      "epoch": 0.413352819013922,
      "grad_norm": 3.65625,
      "learning_rate": 2.379970771479117e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2346058.0,
      "step": 5374
    },
    {
      "epoch": 0.41342973617414047,
      "grad_norm": 1.9765625,
      "learning_rate": 2.3798553957387897e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2346604.0,
      "step": 5375
    },
    {
      "epoch": 0.4135066533343589,
      "grad_norm": 5.15625,
      "learning_rate": 2.3797400199984615e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2346918.0,
      "step": 5376
    },
    {
      "epoch": 0.41358357049457734,
      "grad_norm": 7.5,
      "learning_rate": 2.379624644258134e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 2347260.0,
      "step": 5377
    },
    {
      "epoch": 0.4136604876547958,
      "grad_norm": 4.21875,
      "learning_rate": 2.3795092685178065e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9943342804908752,
      "num_tokens": 2347614.0,
      "step": 5378
    },
    {
      "epoch": 0.41373740481501425,
      "grad_norm": 2.3125,
      "learning_rate": 2.3793938927774787e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 2348438.0,
      "step": 5379
    },
    {
      "epoch": 0.41381432197523266,
      "grad_norm": 4.03125,
      "learning_rate": 2.379278517037151e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2348738.0,
      "step": 5380
    },
    {
      "epoch": 0.4138912391354511,
      "grad_norm": 5.28125,
      "learning_rate": 2.3791631412968233e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2349090.0,
      "step": 5381
    },
    {
      "epoch": 0.4139681562956696,
      "grad_norm": 4.46875,
      "learning_rate": 2.3790477655564955e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 2349404.0,
      "step": 5382
    },
    {
      "epoch": 0.41404507345588804,
      "grad_norm": 3.78125,
      "learning_rate": 2.378932389816168e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2349924.0,
      "step": 5383
    },
    {
      "epoch": 0.41412199061610644,
      "grad_norm": 1.6484375,
      "learning_rate": 2.3788170140758405e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2350498.0,
      "step": 5384
    },
    {
      "epoch": 0.4141989077763249,
      "grad_norm": 2.953125,
      "learning_rate": 2.3787016383355127e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9913644194602966,
      "num_tokens": 2351078.0,
      "step": 5385
    },
    {
      "epoch": 0.41427582493654336,
      "grad_norm": 3.40625,
      "learning_rate": 2.3785862625951854e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2351658.0,
      "step": 5386
    },
    {
      "epoch": 0.41435274209676176,
      "grad_norm": 8.9375,
      "learning_rate": 2.3784708868548572e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 2351986.0,
      "step": 5387
    },
    {
      "epoch": 0.4144296592569802,
      "grad_norm": 4.96875,
      "learning_rate": 2.3783555111145295e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2352312.0,
      "step": 5388
    },
    {
      "epoch": 0.4145065764171987,
      "grad_norm": 2.609375,
      "learning_rate": 2.378240135374202e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2352860.0,
      "step": 5389
    },
    {
      "epoch": 0.41458349357741714,
      "grad_norm": 4.53125,
      "learning_rate": 2.3781247596338744e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2353176.0,
      "step": 5390
    },
    {
      "epoch": 0.41466041073763554,
      "grad_norm": 6.34375,
      "learning_rate": 2.3780093838935467e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2353458.0,
      "step": 5391
    },
    {
      "epoch": 0.414737327897854,
      "grad_norm": 7.3125,
      "learning_rate": 2.377894008153219e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2353762.0,
      "step": 5392
    },
    {
      "epoch": 0.41481424505807246,
      "grad_norm": 2.609375,
      "learning_rate": 2.377778632412891e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 2354338.0,
      "step": 5393
    },
    {
      "epoch": 0.4148911622182909,
      "grad_norm": 2.4375,
      "learning_rate": 2.377663256672564e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 2354862.0,
      "step": 5394
    },
    {
      "epoch": 0.4149680793785093,
      "grad_norm": 6.125,
      "learning_rate": 2.377547880932236e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2355172.0,
      "step": 5395
    },
    {
      "epoch": 0.4150449965387278,
      "grad_norm": 3.0625,
      "learning_rate": 2.3774325051919084e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9925187230110168,
      "num_tokens": 2355976.0,
      "step": 5396
    },
    {
      "epoch": 0.41512191369894624,
      "grad_norm": 2.890625,
      "learning_rate": 2.3773171294515807e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2356562.0,
      "step": 5397
    },
    {
      "epoch": 0.4151988308591647,
      "grad_norm": 4.5625,
      "learning_rate": 2.377201753711253e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 2356868.0,
      "step": 5398
    },
    {
      "epoch": 0.4152757480193831,
      "grad_norm": 5.40625,
      "learning_rate": 2.377086377970925e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2357174.0,
      "step": 5399
    },
    {
      "epoch": 0.41535266517960157,
      "grad_norm": 4.53125,
      "learning_rate": 2.376971002230598e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2357474.0,
      "step": 5400
    },
    {
      "epoch": 0.41542958233982,
      "grad_norm": 2.84375,
      "learning_rate": 2.37685562649027e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 2358062.0,
      "step": 5401
    },
    {
      "epoch": 0.4155064995000385,
      "grad_norm": 3.53125,
      "learning_rate": 2.3767402507499424e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2358658.0,
      "step": 5402
    },
    {
      "epoch": 0.4155834166602569,
      "grad_norm": 1.9921875,
      "learning_rate": 2.3766248750096146e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 2359466.0,
      "step": 5403
    },
    {
      "epoch": 0.41566033382047535,
      "grad_norm": 2.046875,
      "learning_rate": 2.376509499269287e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 2360040.0,
      "step": 5404
    },
    {
      "epoch": 0.4157372509806938,
      "grad_norm": 2.234375,
      "learning_rate": 2.376394123528959e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2360582.0,
      "step": 5405
    },
    {
      "epoch": 0.4158141681409122,
      "grad_norm": 3.984375,
      "learning_rate": 2.376278747788632e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 2360926.0,
      "step": 5406
    },
    {
      "epoch": 0.41589108530113067,
      "grad_norm": 2.34375,
      "learning_rate": 2.376163372048304e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2361460.0,
      "step": 5407
    },
    {
      "epoch": 0.41596800246134913,
      "grad_norm": 2.046875,
      "learning_rate": 2.3760479963079763e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2362256.0,
      "step": 5408
    },
    {
      "epoch": 0.4160449196215676,
      "grad_norm": 5.5625,
      "learning_rate": 2.3759326205676486e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2362562.0,
      "step": 5409
    },
    {
      "epoch": 0.416121836781786,
      "grad_norm": 3.265625,
      "learning_rate": 2.375817244827321e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2363142.0,
      "step": 5410
    },
    {
      "epoch": 0.41619875394200445,
      "grad_norm": 7.0625,
      "learning_rate": 2.3757018690869935e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2363428.0,
      "step": 5411
    },
    {
      "epoch": 0.4162756711022229,
      "grad_norm": 4.625,
      "learning_rate": 2.375586493346666e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2363736.0,
      "step": 5412
    },
    {
      "epoch": 0.41635258826244137,
      "grad_norm": 3.5625,
      "learning_rate": 2.375471117606338e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2364312.0,
      "step": 5413
    },
    {
      "epoch": 0.4164295054226598,
      "grad_norm": 2.125,
      "learning_rate": 2.3753557418660103e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2364868.0,
      "step": 5414
    },
    {
      "epoch": 0.41650642258287823,
      "grad_norm": 5.625,
      "learning_rate": 2.3752403661256826e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2365152.0,
      "step": 5415
    },
    {
      "epoch": 0.4165833397430967,
      "grad_norm": 4.875,
      "learning_rate": 2.375124990385355e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2365482.0,
      "step": 5416
    },
    {
      "epoch": 0.41666025690331515,
      "grad_norm": 6.40625,
      "learning_rate": 2.3750096146450275e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 2365810.0,
      "step": 5417
    },
    {
      "epoch": 0.41673717406353356,
      "grad_norm": 2.421875,
      "learning_rate": 2.3748942389046998e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2366328.0,
      "step": 5418
    },
    {
      "epoch": 0.416814091223752,
      "grad_norm": 2.625,
      "learning_rate": 2.374778863164372e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2366874.0,
      "step": 5419
    },
    {
      "epoch": 0.4168910083839705,
      "grad_norm": 6.0625,
      "learning_rate": 2.3746634874240443e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 2367170.0,
      "step": 5420
    },
    {
      "epoch": 0.41696792554418893,
      "grad_norm": 4.1875,
      "learning_rate": 2.3745481116837165e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 2367766.0,
      "step": 5421
    },
    {
      "epoch": 0.41704484270440734,
      "grad_norm": 6.0625,
      "learning_rate": 2.3744327359433892e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 2368100.0,
      "step": 5422
    },
    {
      "epoch": 0.4171217598646258,
      "grad_norm": 7.5625,
      "learning_rate": 2.3743173602030615e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2368402.0,
      "step": 5423
    },
    {
      "epoch": 0.41719867702484426,
      "grad_norm": 4.6875,
      "learning_rate": 2.3742019844627337e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 2368680.0,
      "step": 5424
    },
    {
      "epoch": 0.4172755941850627,
      "grad_norm": 5.15625,
      "learning_rate": 2.374086608722406e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2368986.0,
      "step": 5425
    },
    {
      "epoch": 0.4173525113452811,
      "grad_norm": 4.40625,
      "learning_rate": 2.3739712329820783e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2369566.0,
      "step": 5426
    },
    {
      "epoch": 0.4174294285054996,
      "grad_norm": 4.1875,
      "learning_rate": 2.3738558572417505e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 2370162.0,
      "step": 5427
    },
    {
      "epoch": 0.41750634566571804,
      "grad_norm": 6.21875,
      "learning_rate": 2.373740481501423e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2370446.0,
      "step": 5428
    },
    {
      "epoch": 0.41758326282593644,
      "grad_norm": 6.25,
      "learning_rate": 2.3736251057610955e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2370732.0,
      "step": 5429
    },
    {
      "epoch": 0.4176601799861549,
      "grad_norm": 7.09375,
      "learning_rate": 2.3735097300207677e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2371084.0,
      "step": 5430
    },
    {
      "epoch": 0.41773709714637336,
      "grad_norm": 3.40625,
      "learning_rate": 2.37339435428044e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 2371652.0,
      "step": 5431
    },
    {
      "epoch": 0.4178140143065918,
      "grad_norm": 5.84375,
      "learning_rate": 2.3732789785401122e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2372008.0,
      "step": 5432
    },
    {
      "epoch": 0.4178909314668102,
      "grad_norm": 4.0625,
      "learning_rate": 2.373163602799785e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2372588.0,
      "step": 5433
    },
    {
      "epoch": 0.4179678486270287,
      "grad_norm": 3.0625,
      "learning_rate": 2.373048227059457e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 2373104.0,
      "step": 5434
    },
    {
      "epoch": 0.41804476578724714,
      "grad_norm": 2.21875,
      "learning_rate": 2.3729328513191294e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2373622.0,
      "step": 5435
    },
    {
      "epoch": 0.4181216829474656,
      "grad_norm": 3.5,
      "learning_rate": 2.3728174755788017e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2374202.0,
      "step": 5436
    },
    {
      "epoch": 0.418198600107684,
      "grad_norm": 5.5625,
      "learning_rate": 2.372702099838474e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2374548.0,
      "step": 5437
    },
    {
      "epoch": 0.41827551726790246,
      "grad_norm": 8.125,
      "learning_rate": 2.372586724098146e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2374852.0,
      "step": 5438
    },
    {
      "epoch": 0.4183524344281209,
      "grad_norm": 8.0625,
      "learning_rate": 2.372471348357819e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2375202.0,
      "step": 5439
    },
    {
      "epoch": 0.4184293515883394,
      "grad_norm": 4.40625,
      "learning_rate": 2.372355972617491e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 2375542.0,
      "step": 5440
    },
    {
      "epoch": 0.4185062687485578,
      "grad_norm": 4.21875,
      "learning_rate": 2.3722405968771634e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2375854.0,
      "step": 5441
    },
    {
      "epoch": 0.41858318590877625,
      "grad_norm": 2.140625,
      "learning_rate": 2.3721252211368357e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2376412.0,
      "step": 5442
    },
    {
      "epoch": 0.4186601030689947,
      "grad_norm": 4.3125,
      "learning_rate": 2.372009845396508e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2376728.0,
      "step": 5443
    },
    {
      "epoch": 0.41873702022921316,
      "grad_norm": 2.28125,
      "learning_rate": 2.3718944696561806e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2377302.0,
      "step": 5444
    },
    {
      "epoch": 0.41881393738943157,
      "grad_norm": 2.65625,
      "learning_rate": 2.371779093915853e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2377854.0,
      "step": 5445
    },
    {
      "epoch": 0.41889085454965,
      "grad_norm": 4.53125,
      "learning_rate": 2.3716637181755247e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 2378186.0,
      "step": 5446
    },
    {
      "epoch": 0.4189677717098685,
      "grad_norm": 7.34375,
      "learning_rate": 2.3715483424351974e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2378492.0,
      "step": 5447
    },
    {
      "epoch": 0.4190446888700869,
      "grad_norm": 6.78125,
      "learning_rate": 2.3714329666948696e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2378804.0,
      "step": 5448
    },
    {
      "epoch": 0.41912160603030535,
      "grad_norm": 2.5625,
      "learning_rate": 2.371317590954542e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 2379368.0,
      "step": 5449
    },
    {
      "epoch": 0.4191985231905238,
      "grad_norm": 6.40625,
      "learning_rate": 2.3712022152142146e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2379674.0,
      "step": 5450
    },
    {
      "epoch": 0.41927544035074227,
      "grad_norm": 4.53125,
      "learning_rate": 2.371086839473887e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2380002.0,
      "step": 5451
    },
    {
      "epoch": 0.41935235751096067,
      "grad_norm": 5.5,
      "learning_rate": 2.370971463733559e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2380320.0,
      "step": 5452
    },
    {
      "epoch": 0.41942927467117913,
      "grad_norm": 6.0625,
      "learning_rate": 2.3708560879932313e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 2380590.0,
      "step": 5453
    },
    {
      "epoch": 0.4195061918313976,
      "grad_norm": 3.25,
      "learning_rate": 2.3707407122529036e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 2381186.0,
      "step": 5454
    },
    {
      "epoch": 0.41958310899161605,
      "grad_norm": 4.96875,
      "learning_rate": 2.370625336512576e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2381548.0,
      "step": 5455
    },
    {
      "epoch": 0.41966002615183445,
      "grad_norm": 6.75,
      "learning_rate": 2.3705099607722485e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2381864.0,
      "step": 5456
    },
    {
      "epoch": 0.4197369433120529,
      "grad_norm": 4.5,
      "learning_rate": 2.3703945850319204e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 2382198.0,
      "step": 5457
    },
    {
      "epoch": 0.41981386047227137,
      "grad_norm": 3.65625,
      "learning_rate": 2.370279209291593e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 2382776.0,
      "step": 5458
    },
    {
      "epoch": 0.41989077763248983,
      "grad_norm": 2.015625,
      "learning_rate": 2.3701638335512653e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2383360.0,
      "step": 5459
    },
    {
      "epoch": 0.41996769479270823,
      "grad_norm": 3.890625,
      "learning_rate": 2.3700484578109376e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2383682.0,
      "step": 5460
    },
    {
      "epoch": 0.4200446119529267,
      "grad_norm": 2.0625,
      "learning_rate": 2.3699330820706103e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2384502.0,
      "step": 5461
    },
    {
      "epoch": 0.42012152911314515,
      "grad_norm": 9.3125,
      "learning_rate": 2.3698177063302825e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9788732528686523,
      "num_tokens": 2384788.0,
      "step": 5462
    },
    {
      "epoch": 0.4201984462733636,
      "grad_norm": 4.9375,
      "learning_rate": 2.3697023305899544e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 2385134.0,
      "step": 5463
    },
    {
      "epoch": 0.420275363433582,
      "grad_norm": 2.671875,
      "learning_rate": 2.369586954849627e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 2385708.0,
      "step": 5464
    },
    {
      "epoch": 0.4203522805938005,
      "grad_norm": 7.0,
      "learning_rate": 2.3694715791092993e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2386020.0,
      "step": 5465
    },
    {
      "epoch": 0.42042919775401894,
      "grad_norm": 6.65625,
      "learning_rate": 2.3693562033689715e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2386328.0,
      "step": 5466
    },
    {
      "epoch": 0.42050611491423734,
      "grad_norm": 4.03125,
      "learning_rate": 2.3692408276286442e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2386904.0,
      "step": 5467
    },
    {
      "epoch": 0.4205830320744558,
      "grad_norm": 3.109375,
      "learning_rate": 2.369125451888316e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 2387426.0,
      "step": 5468
    },
    {
      "epoch": 0.42065994923467426,
      "grad_norm": 5.375,
      "learning_rate": 2.3690100761479887e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2387744.0,
      "step": 5469
    },
    {
      "epoch": 0.4207368663948927,
      "grad_norm": 2.09375,
      "learning_rate": 2.368894700407661e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 2388304.0,
      "step": 5470
    },
    {
      "epoch": 0.4208137835551111,
      "grad_norm": 6.09375,
      "learning_rate": 2.3687793246673333e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2388656.0,
      "step": 5471
    },
    {
      "epoch": 0.4208907007153296,
      "grad_norm": 5.59375,
      "learning_rate": 2.368663948927006e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 2388964.0,
      "step": 5472
    },
    {
      "epoch": 0.42096761787554804,
      "grad_norm": 2.390625,
      "learning_rate": 2.368548573186678e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2389760.0,
      "step": 5473
    },
    {
      "epoch": 0.4210445350357665,
      "grad_norm": 2.1875,
      "learning_rate": 2.36843319744635e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2390308.0,
      "step": 5474
    },
    {
      "epoch": 0.4211214521959849,
      "grad_norm": 4.375,
      "learning_rate": 2.3683178217060227e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2390608.0,
      "step": 5475
    },
    {
      "epoch": 0.42119836935620336,
      "grad_norm": 3.546875,
      "learning_rate": 2.368202445965695e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2391126.0,
      "step": 5476
    },
    {
      "epoch": 0.4212752865164218,
      "grad_norm": 11.5,
      "learning_rate": 2.3680870702253672e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2391428.0,
      "step": 5477
    },
    {
      "epoch": 0.4213522036766403,
      "grad_norm": 2.453125,
      "learning_rate": 2.36797169448504e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 2392014.0,
      "step": 5478
    },
    {
      "epoch": 0.4214291208368587,
      "grad_norm": 9.3125,
      "learning_rate": 2.3678563187447118e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2392306.0,
      "step": 5479
    },
    {
      "epoch": 0.42150603799707714,
      "grad_norm": 8.25,
      "learning_rate": 2.3677409430043844e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2392602.0,
      "step": 5480
    },
    {
      "epoch": 0.4215829551572956,
      "grad_norm": 3.015625,
      "learning_rate": 2.3676255672640567e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2393188.0,
      "step": 5481
    },
    {
      "epoch": 0.42165987231751406,
      "grad_norm": 6.28125,
      "learning_rate": 2.367510191523729e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2393470.0,
      "step": 5482
    },
    {
      "epoch": 0.42173678947773247,
      "grad_norm": 8.4375,
      "learning_rate": 2.3673948157834016e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2393802.0,
      "step": 5483
    },
    {
      "epoch": 0.4218137066379509,
      "grad_norm": 5.875,
      "learning_rate": 2.3672794400430735e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2394100.0,
      "step": 5484
    },
    {
      "epoch": 0.4218906237981694,
      "grad_norm": 3.984375,
      "learning_rate": 2.3671640643027457e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2394426.0,
      "step": 5485
    },
    {
      "epoch": 0.42196754095838784,
      "grad_norm": 4.8125,
      "learning_rate": 2.3670486885624184e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2394742.0,
      "step": 5486
    },
    {
      "epoch": 0.42204445811860625,
      "grad_norm": 2.546875,
      "learning_rate": 2.3669333128220907e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2395552.0,
      "step": 5487
    },
    {
      "epoch": 0.4221213752788247,
      "grad_norm": 2.578125,
      "learning_rate": 2.366817937081763e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2396348.0,
      "step": 5488
    },
    {
      "epoch": 0.42219829243904317,
      "grad_norm": 3.5,
      "learning_rate": 2.3667025613414356e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2396938.0,
      "step": 5489
    },
    {
      "epoch": 0.42227520959926157,
      "grad_norm": 2.4375,
      "learning_rate": 2.3665871856011074e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2397484.0,
      "step": 5490
    },
    {
      "epoch": 0.42235212675948003,
      "grad_norm": 2.015625,
      "learning_rate": 2.36647180986078e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2398030.0,
      "step": 5491
    },
    {
      "epoch": 0.4224290439196985,
      "grad_norm": 3.90625,
      "learning_rate": 2.3663564341204524e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2398352.0,
      "step": 5492
    },
    {
      "epoch": 0.42250596107991695,
      "grad_norm": 3.15625,
      "learning_rate": 2.3662410583801246e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2398894.0,
      "step": 5493
    },
    {
      "epoch": 0.42258287824013535,
      "grad_norm": 3.015625,
      "learning_rate": 2.3661256826397973e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2399438.0,
      "step": 5494
    },
    {
      "epoch": 0.4226597954003538,
      "grad_norm": 4.8125,
      "learning_rate": 2.366010306899469e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2399762.0,
      "step": 5495
    },
    {
      "epoch": 0.42273671256057227,
      "grad_norm": 4.59375,
      "learning_rate": 2.3658949311591414e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 2400072.0,
      "step": 5496
    },
    {
      "epoch": 0.42281362972079073,
      "grad_norm": 3.421875,
      "learning_rate": 2.365779555418814e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2400636.0,
      "step": 5497
    },
    {
      "epoch": 0.42289054688100913,
      "grad_norm": 3.0,
      "learning_rate": 2.3656641796784863e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2400958.0,
      "step": 5498
    },
    {
      "epoch": 0.4229674640412276,
      "grad_norm": 12.0,
      "learning_rate": 2.3655488039381586e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2401258.0,
      "step": 5499
    },
    {
      "epoch": 0.42304438120144605,
      "grad_norm": 5.78125,
      "learning_rate": 2.3654334281978313e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2401620.0,
      "step": 5500
    },
    {
      "epoch": 0.4231212983616645,
      "grad_norm": 5.6875,
      "learning_rate": 2.365318052457503e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2401960.0,
      "step": 5501
    },
    {
      "epoch": 0.4231982155218829,
      "grad_norm": 2.453125,
      "learning_rate": 2.365202676717176e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2402500.0,
      "step": 5502
    },
    {
      "epoch": 0.4232751326821014,
      "grad_norm": 7.40625,
      "learning_rate": 2.365087300976848e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2402826.0,
      "step": 5503
    },
    {
      "epoch": 0.42335204984231983,
      "grad_norm": 5.5,
      "learning_rate": 2.3649719252365203e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2403124.0,
      "step": 5504
    },
    {
      "epoch": 0.4234289670025383,
      "grad_norm": 2.875,
      "learning_rate": 2.3648565494961926e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 2403656.0,
      "step": 5505
    },
    {
      "epoch": 0.4235058841627567,
      "grad_norm": 10.875,
      "learning_rate": 2.364741173755865e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2403912.0,
      "step": 5506
    },
    {
      "epoch": 0.42358280132297516,
      "grad_norm": 8.75,
      "learning_rate": 2.364625798015537e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2404196.0,
      "step": 5507
    },
    {
      "epoch": 0.4236597184831936,
      "grad_norm": 2.90625,
      "learning_rate": 2.3645104222752098e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 2404718.0,
      "step": 5508
    },
    {
      "epoch": 0.423736635643412,
      "grad_norm": 4.1875,
      "learning_rate": 2.364395046534882e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2405068.0,
      "step": 5509
    },
    {
      "epoch": 0.4238135528036305,
      "grad_norm": 9.75,
      "learning_rate": 2.3642796707945543e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2405382.0,
      "step": 5510
    },
    {
      "epoch": 0.42389046996384894,
      "grad_norm": 4.53125,
      "learning_rate": 2.364164295054227e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2405690.0,
      "step": 5511
    },
    {
      "epoch": 0.4239673871240674,
      "grad_norm": 3.453125,
      "learning_rate": 2.364048919313899e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 2406248.0,
      "step": 5512
    },
    {
      "epoch": 0.4240443042842858,
      "grad_norm": 2.015625,
      "learning_rate": 2.363933543573571e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2406788.0,
      "step": 5513
    },
    {
      "epoch": 0.42412122144450426,
      "grad_norm": 7.59375,
      "learning_rate": 2.3638181678332437e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 2407066.0,
      "step": 5514
    },
    {
      "epoch": 0.4241981386047227,
      "grad_norm": 5.59375,
      "learning_rate": 2.363702792092916e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2407348.0,
      "step": 5515
    },
    {
      "epoch": 0.4242750557649412,
      "grad_norm": 46.75,
      "learning_rate": 2.3635874163525883e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2407938.0,
      "step": 5516
    },
    {
      "epoch": 0.4243519729251596,
      "grad_norm": 2.828125,
      "learning_rate": 2.3634720406122605e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 2408454.0,
      "step": 5517
    },
    {
      "epoch": 0.42442889008537804,
      "grad_norm": 2.734375,
      "learning_rate": 2.3633566648719328e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 2409476.0,
      "step": 5518
    },
    {
      "epoch": 0.4245058072455965,
      "grad_norm": 4.28125,
      "learning_rate": 2.3632412891316055e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 2410040.0,
      "step": 5519
    },
    {
      "epoch": 0.42458272440581496,
      "grad_norm": 7.09375,
      "learning_rate": 2.3631259133912777e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 2410300.0,
      "step": 5520
    },
    {
      "epoch": 0.42465964156603336,
      "grad_norm": 2.171875,
      "learning_rate": 2.36301053765095e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 2411084.0,
      "step": 5521
    },
    {
      "epoch": 0.4247365587262518,
      "grad_norm": 3.234375,
      "learning_rate": 2.3628951619106222e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2411640.0,
      "step": 5522
    },
    {
      "epoch": 0.4248134758864703,
      "grad_norm": 5.6875,
      "learning_rate": 2.3627797861702945e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 2411970.0,
      "step": 5523
    },
    {
      "epoch": 0.42489039304668874,
      "grad_norm": 2.078125,
      "learning_rate": 2.3626644104299668e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2412538.0,
      "step": 5524
    },
    {
      "epoch": 0.42496731020690715,
      "grad_norm": 4.5,
      "learning_rate": 2.3625490346896394e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 2413114.0,
      "step": 5525
    },
    {
      "epoch": 0.4250442273671256,
      "grad_norm": 4.6875,
      "learning_rate": 2.3624336589493117e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2413418.0,
      "step": 5526
    },
    {
      "epoch": 0.42512114452734406,
      "grad_norm": 8.625,
      "learning_rate": 2.362318283208984e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 2413726.0,
      "step": 5527
    },
    {
      "epoch": 0.42519806168756247,
      "grad_norm": 5.0,
      "learning_rate": 2.362202907468656e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2414048.0,
      "step": 5528
    },
    {
      "epoch": 0.4252749788477809,
      "grad_norm": 5.40625,
      "learning_rate": 2.3620875317283285e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 2414362.0,
      "step": 5529
    },
    {
      "epoch": 0.4253518960079994,
      "grad_norm": 2.8125,
      "learning_rate": 2.361972155988001e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 2415150.0,
      "step": 5530
    },
    {
      "epoch": 0.42542881316821785,
      "grad_norm": 4.21875,
      "learning_rate": 2.3618567802476734e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2415740.0,
      "step": 5531
    },
    {
      "epoch": 0.42550573032843625,
      "grad_norm": 5.8125,
      "learning_rate": 2.3617414045073457e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2416026.0,
      "step": 5532
    },
    {
      "epoch": 0.4255826474886547,
      "grad_norm": 2.390625,
      "learning_rate": 2.361626028767018e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2416574.0,
      "step": 5533
    },
    {
      "epoch": 0.42565956464887317,
      "grad_norm": 3.25,
      "learning_rate": 2.36151065302669e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2417160.0,
      "step": 5534
    },
    {
      "epoch": 0.4257364818090916,
      "grad_norm": 4.46875,
      "learning_rate": 2.3613952772863624e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2417488.0,
      "step": 5535
    },
    {
      "epoch": 0.42581339896931003,
      "grad_norm": 5.46875,
      "learning_rate": 2.361279901546035e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2417776.0,
      "step": 5536
    },
    {
      "epoch": 0.4258903161295285,
      "grad_norm": 10.3125,
      "learning_rate": 2.3611645258057074e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2418054.0,
      "step": 5537
    },
    {
      "epoch": 0.42596723328974695,
      "grad_norm": 10.8125,
      "learning_rate": 2.3610491500653796e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2418366.0,
      "step": 5538
    },
    {
      "epoch": 0.4260441504499654,
      "grad_norm": 3.6875,
      "learning_rate": 2.360933774325052e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2418884.0,
      "step": 5539
    },
    {
      "epoch": 0.4261210676101838,
      "grad_norm": 6.3125,
      "learning_rate": 2.360818398584724e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 2419218.0,
      "step": 5540
    },
    {
      "epoch": 0.42619798477040227,
      "grad_norm": 2.484375,
      "learning_rate": 2.360703022844397e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2419782.0,
      "step": 5541
    },
    {
      "epoch": 0.42627490193062073,
      "grad_norm": 7.9375,
      "learning_rate": 2.360587647104069e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2420066.0,
      "step": 5542
    },
    {
      "epoch": 0.4263518190908392,
      "grad_norm": 10.5,
      "learning_rate": 2.3604722713637414e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 2420342.0,
      "step": 5543
    },
    {
      "epoch": 0.4264287362510576,
      "grad_norm": 3.125,
      "learning_rate": 2.3603568956234136e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2420888.0,
      "step": 5544
    },
    {
      "epoch": 0.42650565341127605,
      "grad_norm": 9.3125,
      "learning_rate": 2.360241519883086e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2421144.0,
      "step": 5545
    },
    {
      "epoch": 0.4265825705714945,
      "grad_norm": 6.65625,
      "learning_rate": 2.360126144142758e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2421488.0,
      "step": 5546
    },
    {
      "epoch": 0.42665948773171297,
      "grad_norm": 6.0,
      "learning_rate": 2.360010768402431e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 2421798.0,
      "step": 5547
    },
    {
      "epoch": 0.4267364048919314,
      "grad_norm": 2.390625,
      "learning_rate": 2.359895392662103e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2422360.0,
      "step": 5548
    },
    {
      "epoch": 0.42681332205214983,
      "grad_norm": 2.8125,
      "learning_rate": 2.3597800169217753e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2422944.0,
      "step": 5549
    },
    {
      "epoch": 0.4268902392123683,
      "grad_norm": 5.84375,
      "learning_rate": 2.3596646411814476e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2423258.0,
      "step": 5550
    },
    {
      "epoch": 0.4269671563725867,
      "grad_norm": 1.7421875,
      "learning_rate": 2.35954926544112e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 2423780.0,
      "step": 5551
    },
    {
      "epoch": 0.42704407353280516,
      "grad_norm": 3.9375,
      "learning_rate": 2.3594338897007925e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2424334.0,
      "step": 5552
    },
    {
      "epoch": 0.4271209906930236,
      "grad_norm": 6.21875,
      "learning_rate": 2.3593185139604648e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2424624.0,
      "step": 5553
    },
    {
      "epoch": 0.4271979078532421,
      "grad_norm": 4.65625,
      "learning_rate": 2.359203138220137e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 2425216.0,
      "step": 5554
    },
    {
      "epoch": 0.4272748250134605,
      "grad_norm": 2.84375,
      "learning_rate": 2.3590877624798093e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2425772.0,
      "step": 5555
    },
    {
      "epoch": 0.42735174217367894,
      "grad_norm": 5.0625,
      "learning_rate": 2.3589723867394816e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2426088.0,
      "step": 5556
    },
    {
      "epoch": 0.4274286593338974,
      "grad_norm": 5.0,
      "learning_rate": 2.358857010999154e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2426652.0,
      "step": 5557
    },
    {
      "epoch": 0.42750557649411586,
      "grad_norm": 6.3125,
      "learning_rate": 2.3587416352588265e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2426968.0,
      "step": 5558
    },
    {
      "epoch": 0.42758249365433426,
      "grad_norm": 6.46875,
      "learning_rate": 2.3586262595184988e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2427298.0,
      "step": 5559
    },
    {
      "epoch": 0.4276594108145527,
      "grad_norm": 2.296875,
      "learning_rate": 2.358510883778171e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2427854.0,
      "step": 5560
    },
    {
      "epoch": 0.4277363279747712,
      "grad_norm": 3.078125,
      "learning_rate": 2.3583955080378433e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 2428432.0,
      "step": 5561
    },
    {
      "epoch": 0.42781324513498964,
      "grad_norm": 8.3125,
      "learning_rate": 2.3582801322975155e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9967637658119202,
      "num_tokens": 2428744.0,
      "step": 5562
    },
    {
      "epoch": 0.42789016229520804,
      "grad_norm": 6.96875,
      "learning_rate": 2.3581647565571878e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2429084.0,
      "step": 5563
    },
    {
      "epoch": 0.4279670794554265,
      "grad_norm": 4.1875,
      "learning_rate": 2.3580493808168605e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2429408.0,
      "step": 5564
    },
    {
      "epoch": 0.42804399661564496,
      "grad_norm": 3.71875,
      "learning_rate": 2.3579340050765327e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2429754.0,
      "step": 5565
    },
    {
      "epoch": 0.4281209137758634,
      "grad_norm": 3.8125,
      "learning_rate": 2.357818629336205e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2430294.0,
      "step": 5566
    },
    {
      "epoch": 0.4281978309360818,
      "grad_norm": 1.875,
      "learning_rate": 2.3577032535958772e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 2431118.0,
      "step": 5567
    },
    {
      "epoch": 0.4282747480963003,
      "grad_norm": 5.15625,
      "learning_rate": 2.3575878778555495e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2431428.0,
      "step": 5568
    },
    {
      "epoch": 0.42835166525651874,
      "grad_norm": 3.890625,
      "learning_rate": 2.357472502115222e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2431728.0,
      "step": 5569
    },
    {
      "epoch": 0.42842858241673715,
      "grad_norm": 4.625,
      "learning_rate": 2.3573571263748944e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 2432040.0,
      "step": 5570
    },
    {
      "epoch": 0.4285054995769556,
      "grad_norm": 2.71875,
      "learning_rate": 2.3572417506345663e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2432572.0,
      "step": 5571
    },
    {
      "epoch": 0.42858241673717407,
      "grad_norm": 8.625,
      "learning_rate": 2.357126374894239e-06,
      "loss": 0.0371,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2432832.0,
      "step": 5572
    },
    {
      "epoch": 0.4286593338973925,
      "grad_norm": 8.375,
      "learning_rate": 2.3570109991539112e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 2433132.0,
      "step": 5573
    },
    {
      "epoch": 0.42873625105761093,
      "grad_norm": 4.1875,
      "learning_rate": 2.3568956234135835e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2433448.0,
      "step": 5574
    },
    {
      "epoch": 0.4288131682178294,
      "grad_norm": 4.0625,
      "learning_rate": 2.356780247673256e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 2433782.0,
      "step": 5575
    },
    {
      "epoch": 0.42889008537804785,
      "grad_norm": 4.78125,
      "learning_rate": 2.3566648719329284e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 2434130.0,
      "step": 5576
    },
    {
      "epoch": 0.4289670025382663,
      "grad_norm": 2.375,
      "learning_rate": 2.3565494961926007e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2434678.0,
      "step": 5577
    },
    {
      "epoch": 0.4290439196984847,
      "grad_norm": 6.375,
      "learning_rate": 2.356434120452273e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 2434974.0,
      "step": 5578
    },
    {
      "epoch": 0.42912083685870317,
      "grad_norm": 3.40625,
      "learning_rate": 2.356318744711945e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 2435550.0,
      "step": 5579
    },
    {
      "epoch": 0.42919775401892163,
      "grad_norm": 2.421875,
      "learning_rate": 2.356203368971618e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2436370.0,
      "step": 5580
    },
    {
      "epoch": 0.4292746711791401,
      "grad_norm": 7.4375,
      "learning_rate": 2.35608799323129e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 2436708.0,
      "step": 5581
    },
    {
      "epoch": 0.4293515883393585,
      "grad_norm": 3.90625,
      "learning_rate": 2.355972617490962e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2437030.0,
      "step": 5582
    },
    {
      "epoch": 0.42942850549957695,
      "grad_norm": 4.96875,
      "learning_rate": 2.3558572417506346e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2437342.0,
      "step": 5583
    },
    {
      "epoch": 0.4295054226597954,
      "grad_norm": 4.09375,
      "learning_rate": 2.355741866010307e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2437636.0,
      "step": 5584
    },
    {
      "epoch": 0.42958233982001387,
      "grad_norm": 2.03125,
      "learning_rate": 2.355626490269979e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2438444.0,
      "step": 5585
    },
    {
      "epoch": 0.4296592569802323,
      "grad_norm": 2.390625,
      "learning_rate": 2.355511114529652e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2439012.0,
      "step": 5586
    },
    {
      "epoch": 0.42973617414045073,
      "grad_norm": 2.265625,
      "learning_rate": 2.3553957387893237e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2439832.0,
      "step": 5587
    },
    {
      "epoch": 0.4298130913006692,
      "grad_norm": 3.5,
      "learning_rate": 2.3552803630489964e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2440152.0,
      "step": 5588
    },
    {
      "epoch": 0.42989000846088765,
      "grad_norm": 6.84375,
      "learning_rate": 2.3551649873086686e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 2440454.0,
      "step": 5589
    },
    {
      "epoch": 0.42996692562110606,
      "grad_norm": 6.90625,
      "learning_rate": 2.355049611568341e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2440738.0,
      "step": 5590
    },
    {
      "epoch": 0.4300438427813245,
      "grad_norm": 1.9765625,
      "learning_rate": 2.3549342358280136e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2441308.0,
      "step": 5591
    },
    {
      "epoch": 0.430120759941543,
      "grad_norm": 7.375,
      "learning_rate": 2.354818860087686e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 2441594.0,
      "step": 5592
    },
    {
      "epoch": 0.4301976771017614,
      "grad_norm": 2.625,
      "learning_rate": 2.3547034843473576e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 2442164.0,
      "step": 5593
    },
    {
      "epoch": 0.43027459426197984,
      "grad_norm": 2.75,
      "learning_rate": 2.3545881086070303e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2442750.0,
      "step": 5594
    },
    {
      "epoch": 0.4303515114221983,
      "grad_norm": 2.375,
      "learning_rate": 2.3544727328667026e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2443268.0,
      "step": 5595
    },
    {
      "epoch": 0.43042842858241676,
      "grad_norm": 5.15625,
      "learning_rate": 2.354357357126375e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 2443850.0,
      "step": 5596
    },
    {
      "epoch": 0.43050534574263516,
      "grad_norm": 4.9375,
      "learning_rate": 2.3542419813860475e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2444134.0,
      "step": 5597
    },
    {
      "epoch": 0.4305822629028536,
      "grad_norm": 9.25,
      "learning_rate": 2.3541266056457194e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 2444492.0,
      "step": 5598
    },
    {
      "epoch": 0.4306591800630721,
      "grad_norm": 5.46875,
      "learning_rate": 2.354011229905392e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 2444822.0,
      "step": 5599
    },
    {
      "epoch": 0.43073609722329054,
      "grad_norm": 2.515625,
      "learning_rate": 2.3538958541650643e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2445362.0,
      "step": 5600
    },
    {
      "epoch": 0.43081301438350894,
      "grad_norm": 2.765625,
      "learning_rate": 2.3537804784247366e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 2445884.0,
      "step": 5601
    },
    {
      "epoch": 0.4308899315437274,
      "grad_norm": 2.1875,
      "learning_rate": 2.3536651026844092e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2446458.0,
      "step": 5602
    },
    {
      "epoch": 0.43096684870394586,
      "grad_norm": 3.828125,
      "learning_rate": 2.3535497269440815e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2446742.0,
      "step": 5603
    },
    {
      "epoch": 0.4310437658641643,
      "grad_norm": 10.6875,
      "learning_rate": 2.3534343512037533e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2447092.0,
      "step": 5604
    },
    {
      "epoch": 0.4311206830243827,
      "grad_norm": 8.375,
      "learning_rate": 2.353318975463426e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2447432.0,
      "step": 5605
    },
    {
      "epoch": 0.4311976001846012,
      "grad_norm": 8.4375,
      "learning_rate": 2.3532035997230983e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2447772.0,
      "step": 5606
    },
    {
      "epoch": 0.43127451734481964,
      "grad_norm": 3.546875,
      "learning_rate": 2.3530882239827705e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2448316.0,
      "step": 5607
    },
    {
      "epoch": 0.4313514345050381,
      "grad_norm": 1.921875,
      "learning_rate": 2.352972848242443e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 2449126.0,
      "step": 5608
    },
    {
      "epoch": 0.4314283516652565,
      "grad_norm": 5.625,
      "learning_rate": 2.352857472502115e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2449472.0,
      "step": 5609
    },
    {
      "epoch": 0.43150526882547496,
      "grad_norm": 4.8125,
      "learning_rate": 2.3527420967617877e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2449784.0,
      "step": 5610
    },
    {
      "epoch": 0.4315821859856934,
      "grad_norm": 4.125,
      "learning_rate": 2.35262672102146e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2450380.0,
      "step": 5611
    },
    {
      "epoch": 0.4316591031459118,
      "grad_norm": 4.96875,
      "learning_rate": 2.3525113452811322e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2450692.0,
      "step": 5612
    },
    {
      "epoch": 0.4317360203061303,
      "grad_norm": 4.90625,
      "learning_rate": 2.352395969540805e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2450998.0,
      "step": 5613
    },
    {
      "epoch": 0.43181293746634875,
      "grad_norm": 6.125,
      "learning_rate": 2.352280593800477e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 2451318.0,
      "step": 5614
    },
    {
      "epoch": 0.4318898546265672,
      "grad_norm": 5.375,
      "learning_rate": 2.352165218060149e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 2451662.0,
      "step": 5615
    },
    {
      "epoch": 0.4319667717867856,
      "grad_norm": 3.828125,
      "learning_rate": 2.3520498423198217e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2451986.0,
      "step": 5616
    },
    {
      "epoch": 0.43204368894700407,
      "grad_norm": 5.375,
      "learning_rate": 2.351934466579494e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 2452284.0,
      "step": 5617
    },
    {
      "epoch": 0.4321206061072225,
      "grad_norm": 4.15625,
      "learning_rate": 2.3518190908391662e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2452610.0,
      "step": 5618
    },
    {
      "epoch": 0.432197523267441,
      "grad_norm": 7.71875,
      "learning_rate": 2.351703715098839e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 2452916.0,
      "step": 5619
    },
    {
      "epoch": 0.4322744404276594,
      "grad_norm": 4.625,
      "learning_rate": 2.3515883393585107e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2453222.0,
      "step": 5620
    },
    {
      "epoch": 0.43235135758787785,
      "grad_norm": 7.0,
      "learning_rate": 2.3514729636181834e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 2453484.0,
      "step": 5621
    },
    {
      "epoch": 0.4324282747480963,
      "grad_norm": 3.609375,
      "learning_rate": 2.3513575878778557e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2453804.0,
      "step": 5622
    },
    {
      "epoch": 0.43250519190831477,
      "grad_norm": 3.703125,
      "learning_rate": 2.351242212137528e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2454358.0,
      "step": 5623
    },
    {
      "epoch": 0.43258210906853317,
      "grad_norm": 2.375,
      "learning_rate": 2.3511268363972e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2454876.0,
      "step": 5624
    },
    {
      "epoch": 0.43265902622875163,
      "grad_norm": 4.34375,
      "learning_rate": 2.3510114606568724e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2455188.0,
      "step": 5625
    },
    {
      "epoch": 0.4327359433889701,
      "grad_norm": 3.640625,
      "learning_rate": 2.3508960849165447e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2455778.0,
      "step": 5626
    },
    {
      "epoch": 0.43281286054918855,
      "grad_norm": 2.53125,
      "learning_rate": 2.3507807091762174e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2456322.0,
      "step": 5627
    },
    {
      "epoch": 0.43288977770940695,
      "grad_norm": 5.34375,
      "learning_rate": 2.3506653334358896e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2456608.0,
      "step": 5628
    },
    {
      "epoch": 0.4329666948696254,
      "grad_norm": 4.25,
      "learning_rate": 2.350549957695562e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 2456956.0,
      "step": 5629
    },
    {
      "epoch": 0.43304361202984387,
      "grad_norm": 3.75,
      "learning_rate": 2.3504345819552346e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2457532.0,
      "step": 5630
    },
    {
      "epoch": 0.4331205291900623,
      "grad_norm": 9.25,
      "learning_rate": 2.3503192062149064e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 2457784.0,
      "step": 5631
    },
    {
      "epoch": 0.43319744635028073,
      "grad_norm": 4.375,
      "learning_rate": 2.3502038304745787e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2458102.0,
      "step": 5632
    },
    {
      "epoch": 0.4332743635104992,
      "grad_norm": 7.0,
      "learning_rate": 2.3500884547342514e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2458462.0,
      "step": 5633
    },
    {
      "epoch": 0.43335128067071765,
      "grad_norm": 3.546875,
      "learning_rate": 2.3499730789939236e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2459288.0,
      "step": 5634
    },
    {
      "epoch": 0.43342819783093606,
      "grad_norm": 6.71875,
      "learning_rate": 2.349857703253596e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2459594.0,
      "step": 5635
    },
    {
      "epoch": 0.4335051149911545,
      "grad_norm": 2.65625,
      "learning_rate": 2.349742327513268e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2460150.0,
      "step": 5636
    },
    {
      "epoch": 0.433582032151373,
      "grad_norm": 3.34375,
      "learning_rate": 2.3496269517729404e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2460462.0,
      "step": 5637
    },
    {
      "epoch": 0.43365894931159144,
      "grad_norm": 2.71875,
      "learning_rate": 2.349511576032613e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 2461020.0,
      "step": 5638
    },
    {
      "epoch": 0.43373586647180984,
      "grad_norm": 3.5,
      "learning_rate": 2.3493962002922853e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2461342.0,
      "step": 5639
    },
    {
      "epoch": 0.4338127836320283,
      "grad_norm": 3.515625,
      "learning_rate": 2.3492808245519576e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2461884.0,
      "step": 5640
    },
    {
      "epoch": 0.43388970079224676,
      "grad_norm": 3.1875,
      "learning_rate": 2.3491654488116303e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2462454.0,
      "step": 5641
    },
    {
      "epoch": 0.4339666179524652,
      "grad_norm": 2.109375,
      "learning_rate": 2.349050073071302e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2463236.0,
      "step": 5642
    },
    {
      "epoch": 0.4340435351126836,
      "grad_norm": 3.28125,
      "learning_rate": 2.3489346973309744e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2463762.0,
      "step": 5643
    },
    {
      "epoch": 0.4341204522729021,
      "grad_norm": 5.125,
      "learning_rate": 2.348819321590647e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2464046.0,
      "step": 5644
    },
    {
      "epoch": 0.43419736943312054,
      "grad_norm": 3.28125,
      "learning_rate": 2.3487039458503193e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2464626.0,
      "step": 5645
    },
    {
      "epoch": 0.434274286593339,
      "grad_norm": 4.125,
      "learning_rate": 2.3485885701099916e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2465166.0,
      "step": 5646
    },
    {
      "epoch": 0.4343512037535574,
      "grad_norm": 5.40625,
      "learning_rate": 2.348473194369664e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2465514.0,
      "step": 5647
    },
    {
      "epoch": 0.43442812091377586,
      "grad_norm": 8.6875,
      "learning_rate": 2.348357818629336e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 2465814.0,
      "step": 5648
    },
    {
      "epoch": 0.4345050380739943,
      "grad_norm": 2.140625,
      "learning_rate": 2.3482424428890088e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 2466598.0,
      "step": 5649
    },
    {
      "epoch": 0.4345819552342128,
      "grad_norm": 4.78125,
      "learning_rate": 2.348127067148681e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2466882.0,
      "step": 5650
    },
    {
      "epoch": 0.4346588723944312,
      "grad_norm": 3.375,
      "learning_rate": 2.3480116914083533e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2467472.0,
      "step": 5651
    },
    {
      "epoch": 0.43473578955464964,
      "grad_norm": 1.96875,
      "learning_rate": 2.347896315668026e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2468252.0,
      "step": 5652
    },
    {
      "epoch": 0.4348127067148681,
      "grad_norm": 3.09375,
      "learning_rate": 2.347780939927698e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 2468820.0,
      "step": 5653
    },
    {
      "epoch": 0.4348896238750865,
      "grad_norm": 5.25,
      "learning_rate": 2.34766556418737e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2469144.0,
      "step": 5654
    },
    {
      "epoch": 0.43496654103530497,
      "grad_norm": 4.65625,
      "learning_rate": 2.3475501884470427e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 2469744.0,
      "step": 5655
    },
    {
      "epoch": 0.4350434581955234,
      "grad_norm": 3.71875,
      "learning_rate": 2.347434812706715e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2470056.0,
      "step": 5656
    },
    {
      "epoch": 0.4351203753557419,
      "grad_norm": 3.375,
      "learning_rate": 2.3473194369663872e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2470596.0,
      "step": 5657
    },
    {
      "epoch": 0.4351972925159603,
      "grad_norm": 7.0625,
      "learning_rate": 2.3472040612260595e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2470948.0,
      "step": 5658
    },
    {
      "epoch": 0.43527420967617875,
      "grad_norm": 2.328125,
      "learning_rate": 2.3470886854857318e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2471466.0,
      "step": 5659
    },
    {
      "epoch": 0.4353511268363972,
      "grad_norm": 2.328125,
      "learning_rate": 2.3469733097454044e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2472046.0,
      "step": 5660
    },
    {
      "epoch": 0.43542804399661567,
      "grad_norm": 3.1875,
      "learning_rate": 2.3468579340050767e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2472572.0,
      "step": 5661
    },
    {
      "epoch": 0.43550496115683407,
      "grad_norm": 5.21875,
      "learning_rate": 2.346742558264749e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2472902.0,
      "step": 5662
    },
    {
      "epoch": 0.43558187831705253,
      "grad_norm": 5.0625,
      "learning_rate": 2.3466271825244212e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 2473242.0,
      "step": 5663
    },
    {
      "epoch": 0.435658795477271,
      "grad_norm": 7.3125,
      "learning_rate": 2.3465118067840935e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2473598.0,
      "step": 5664
    },
    {
      "epoch": 0.43573571263748945,
      "grad_norm": 7.0625,
      "learning_rate": 2.3463964310437657e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2473908.0,
      "step": 5665
    },
    {
      "epoch": 0.43581262979770785,
      "grad_norm": 2.234375,
      "learning_rate": 2.3462810553034384e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 2474432.0,
      "step": 5666
    },
    {
      "epoch": 0.4358895469579263,
      "grad_norm": 3.078125,
      "learning_rate": 2.3461656795631107e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2475008.0,
      "step": 5667
    },
    {
      "epoch": 0.43596646411814477,
      "grad_norm": 4.09375,
      "learning_rate": 2.346050303822783e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2475358.0,
      "step": 5668
    },
    {
      "epoch": 0.43604338127836323,
      "grad_norm": 7.21875,
      "learning_rate": 2.345934928082455e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2475642.0,
      "step": 5669
    },
    {
      "epoch": 0.43612029843858163,
      "grad_norm": 4.28125,
      "learning_rate": 2.3458195523421274e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2475986.0,
      "step": 5670
    },
    {
      "epoch": 0.4361972155988001,
      "grad_norm": 7.09375,
      "learning_rate": 2.3457041766018e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2476268.0,
      "step": 5671
    },
    {
      "epoch": 0.43627413275901855,
      "grad_norm": 3.578125,
      "learning_rate": 2.3455888008614724e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 2476614.0,
      "step": 5672
    },
    {
      "epoch": 0.43635104991923696,
      "grad_norm": 2.59375,
      "learning_rate": 2.3454734251211446e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2477184.0,
      "step": 5673
    },
    {
      "epoch": 0.4364279670794554,
      "grad_norm": 2.53125,
      "learning_rate": 2.345358049380817e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2477740.0,
      "step": 5674
    },
    {
      "epoch": 0.4365048842396739,
      "grad_norm": 1.9375,
      "learning_rate": 2.345242673640489e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2478314.0,
      "step": 5675
    },
    {
      "epoch": 0.43658180139989233,
      "grad_norm": 6.625,
      "learning_rate": 2.3451272979001614e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 2478670.0,
      "step": 5676
    },
    {
      "epoch": 0.43665871856011074,
      "grad_norm": 5.375,
      "learning_rate": 2.345011922159834e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2479000.0,
      "step": 5677
    },
    {
      "epoch": 0.4367356357203292,
      "grad_norm": 4.09375,
      "learning_rate": 2.3448965464195064e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2479596.0,
      "step": 5678
    },
    {
      "epoch": 0.43681255288054766,
      "grad_norm": 12.5,
      "learning_rate": 2.3447811706791786e-06,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 2479848.0,
      "step": 5679
    },
    {
      "epoch": 0.4368894700407661,
      "grad_norm": 3.5,
      "learning_rate": 2.344665794938851e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2480368.0,
      "step": 5680
    },
    {
      "epoch": 0.4369663872009845,
      "grad_norm": 2.171875,
      "learning_rate": 2.344550419198523e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9908925294876099,
      "num_tokens": 2480918.0,
      "step": 5681
    },
    {
      "epoch": 0.437043304361203,
      "grad_norm": 6.9375,
      "learning_rate": 2.3444350434581954e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2481200.0,
      "step": 5682
    },
    {
      "epoch": 0.43712022152142144,
      "grad_norm": 11.9375,
      "learning_rate": 2.344319667717868e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2481500.0,
      "step": 5683
    },
    {
      "epoch": 0.4371971386816399,
      "grad_norm": 4.03125,
      "learning_rate": 2.3442042919775403e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2481822.0,
      "step": 5684
    },
    {
      "epoch": 0.4372740558418583,
      "grad_norm": 5.25,
      "learning_rate": 2.3440889162372126e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2482152.0,
      "step": 5685
    },
    {
      "epoch": 0.43735097300207676,
      "grad_norm": 3.4375,
      "learning_rate": 2.343973540496885e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2482714.0,
      "step": 5686
    },
    {
      "epoch": 0.4374278901622952,
      "grad_norm": 5.78125,
      "learning_rate": 2.343858164756557e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2483042.0,
      "step": 5687
    },
    {
      "epoch": 0.4375048073225137,
      "grad_norm": 5.625,
      "learning_rate": 2.3437427890162298e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 2483642.0,
      "step": 5688
    },
    {
      "epoch": 0.4375817244827321,
      "grad_norm": 6.28125,
      "learning_rate": 2.343627413275902e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2483942.0,
      "step": 5689
    },
    {
      "epoch": 0.43765864164295054,
      "grad_norm": 2.09375,
      "learning_rate": 2.343512037535574e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2484488.0,
      "step": 5690
    },
    {
      "epoch": 0.437735558803169,
      "grad_norm": 4.28125,
      "learning_rate": 2.3433966617952466e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2484812.0,
      "step": 5691
    },
    {
      "epoch": 0.4378124759633874,
      "grad_norm": 3.71875,
      "learning_rate": 2.343281286054919e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2485138.0,
      "step": 5692
    },
    {
      "epoch": 0.43788939312360586,
      "grad_norm": 2.15625,
      "learning_rate": 2.343165910314591e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 2485724.0,
      "step": 5693
    },
    {
      "epoch": 0.4379663102838243,
      "grad_norm": 5.0,
      "learning_rate": 2.3430505345742638e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2486264.0,
      "step": 5694
    },
    {
      "epoch": 0.4380432274440428,
      "grad_norm": 7.84375,
      "learning_rate": 2.342935158833936e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2486580.0,
      "step": 5695
    },
    {
      "epoch": 0.4381201446042612,
      "grad_norm": 3.921875,
      "learning_rate": 2.3428197830936083e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 2486932.0,
      "step": 5696
    },
    {
      "epoch": 0.43819706176447965,
      "grad_norm": 4.34375,
      "learning_rate": 2.3427044073532805e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 2487476.0,
      "step": 5697
    },
    {
      "epoch": 0.4382739789246981,
      "grad_norm": 2.03125,
      "learning_rate": 2.342589031612953e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 2488262.0,
      "step": 5698
    },
    {
      "epoch": 0.43835089608491656,
      "grad_norm": 4.25,
      "learning_rate": 2.3424736558726255e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9943661689758301,
      "num_tokens": 2488618.0,
      "step": 5699
    },
    {
      "epoch": 0.43842781324513497,
      "grad_norm": 2.609375,
      "learning_rate": 2.3423582801322977e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2489174.0,
      "step": 5700
    },
    {
      "epoch": 0.4385047304053534,
      "grad_norm": 3.265625,
      "learning_rate": 2.3422429043919696e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9966158866882324,
      "num_tokens": 2489766.0,
      "step": 5701
    },
    {
      "epoch": 0.4385816475655719,
      "grad_norm": 2.859375,
      "learning_rate": 2.3421275286516422e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 2490318.0,
      "step": 5702
    },
    {
      "epoch": 0.43865856472579035,
      "grad_norm": 7.0625,
      "learning_rate": 2.3420121529113145e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2490628.0,
      "step": 5703
    },
    {
      "epoch": 0.43873548188600875,
      "grad_norm": 4.0,
      "learning_rate": 2.3418967771709868e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9897785186767578,
      "num_tokens": 2491216.0,
      "step": 5704
    },
    {
      "epoch": 0.4388123990462272,
      "grad_norm": 4.09375,
      "learning_rate": 2.3417814014306594e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2491538.0,
      "step": 5705
    },
    {
      "epoch": 0.43888931620644567,
      "grad_norm": 4.5625,
      "learning_rate": 2.3416660256903317e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2492084.0,
      "step": 5706
    },
    {
      "epoch": 0.4389662333666641,
      "grad_norm": 10.6875,
      "learning_rate": 2.341550649950004e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 2492392.0,
      "step": 5707
    },
    {
      "epoch": 0.43904315052688253,
      "grad_norm": 7.4375,
      "learning_rate": 2.3414352742096762e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2492708.0,
      "step": 5708
    },
    {
      "epoch": 0.439120067687101,
      "grad_norm": 10.1875,
      "learning_rate": 2.3413198984693485e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 2493010.0,
      "step": 5709
    },
    {
      "epoch": 0.43919698484731945,
      "grad_norm": 10.375,
      "learning_rate": 2.341204522729021e-06,
      "loss": 0.0401,
      "mean_token_accuracy": 0.9784172773361206,
      "num_tokens": 2493290.0,
      "step": 5710
    },
    {
      "epoch": 0.4392739020075379,
      "grad_norm": 3.828125,
      "learning_rate": 2.3410891469886934e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2493636.0,
      "step": 5711
    },
    {
      "epoch": 0.4393508191677563,
      "grad_norm": 5.53125,
      "learning_rate": 2.3409737712483653e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 2493976.0,
      "step": 5712
    },
    {
      "epoch": 0.43942773632797477,
      "grad_norm": 5.15625,
      "learning_rate": 2.340858395508038e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2494310.0,
      "step": 5713
    },
    {
      "epoch": 0.43950465348819323,
      "grad_norm": 4.34375,
      "learning_rate": 2.34074301976771e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2494896.0,
      "step": 5714
    },
    {
      "epoch": 0.43958157064841163,
      "grad_norm": 2.359375,
      "learning_rate": 2.3406276440273825e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2495438.0,
      "step": 5715
    },
    {
      "epoch": 0.4396584878086301,
      "grad_norm": 7.34375,
      "learning_rate": 2.340512268287055e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2495724.0,
      "step": 5716
    },
    {
      "epoch": 0.43973540496884855,
      "grad_norm": 2.828125,
      "learning_rate": 2.3403968925467274e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 2496240.0,
      "step": 5717
    },
    {
      "epoch": 0.439812322129067,
      "grad_norm": 5.9375,
      "learning_rate": 2.3402815168063996e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2496544.0,
      "step": 5718
    },
    {
      "epoch": 0.4398892392892854,
      "grad_norm": 2.140625,
      "learning_rate": 2.340166141066072e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2497086.0,
      "step": 5719
    },
    {
      "epoch": 0.4399661564495039,
      "grad_norm": 7.15625,
      "learning_rate": 2.340050765325744e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 2497432.0,
      "step": 5720
    },
    {
      "epoch": 0.44004307360972233,
      "grad_norm": 5.0,
      "learning_rate": 2.339935389585417e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2497986.0,
      "step": 5721
    },
    {
      "epoch": 0.4401199907699408,
      "grad_norm": 4.625,
      "learning_rate": 2.339820013845089e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2498298.0,
      "step": 5722
    },
    {
      "epoch": 0.4401969079301592,
      "grad_norm": 8.625,
      "learning_rate": 2.339704638104761e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 2498654.0,
      "step": 5723
    },
    {
      "epoch": 0.44027382509037766,
      "grad_norm": 4.59375,
      "learning_rate": 2.3395892623644336e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 2498932.0,
      "step": 5724
    },
    {
      "epoch": 0.4403507422505961,
      "grad_norm": 4.53125,
      "learning_rate": 2.339473886624106e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2499252.0,
      "step": 5725
    },
    {
      "epoch": 0.4404276594108146,
      "grad_norm": 3.203125,
      "learning_rate": 2.339358510883778e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2499836.0,
      "step": 5726
    },
    {
      "epoch": 0.440504576571033,
      "grad_norm": 4.125,
      "learning_rate": 2.339243135143451e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2500148.0,
      "step": 5727
    },
    {
      "epoch": 0.44058149373125144,
      "grad_norm": 6.59375,
      "learning_rate": 2.3391277594031227e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 2500436.0,
      "step": 5728
    },
    {
      "epoch": 0.4406584108914699,
      "grad_norm": 2.0625,
      "learning_rate": 2.3390123836627953e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2501244.0,
      "step": 5729
    },
    {
      "epoch": 0.44073532805168836,
      "grad_norm": 7.28125,
      "learning_rate": 2.3388970079224676e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 2501520.0,
      "step": 5730
    },
    {
      "epoch": 0.44081224521190676,
      "grad_norm": 3.296875,
      "learning_rate": 2.33878163218214e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2502040.0,
      "step": 5731
    },
    {
      "epoch": 0.4408891623721252,
      "grad_norm": 4.5625,
      "learning_rate": 2.338666256441812e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 2502640.0,
      "step": 5732
    },
    {
      "epoch": 0.4409660795323437,
      "grad_norm": 9.8125,
      "learning_rate": 2.338550880701485e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 2502976.0,
      "step": 5733
    },
    {
      "epoch": 0.4410429966925621,
      "grad_norm": 2.03125,
      "learning_rate": 2.3384355049611566e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 2503760.0,
      "step": 5734
    },
    {
      "epoch": 0.44111991385278054,
      "grad_norm": 3.625,
      "learning_rate": 2.3383201292208293e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2504104.0,
      "step": 5735
    },
    {
      "epoch": 0.441196831012999,
      "grad_norm": 4.28125,
      "learning_rate": 2.3382047534805016e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2504700.0,
      "step": 5736
    },
    {
      "epoch": 0.44127374817321746,
      "grad_norm": 3.53125,
      "learning_rate": 2.338089377740174e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2505012.0,
      "step": 5737
    },
    {
      "epoch": 0.44135066533343587,
      "grad_norm": 1.953125,
      "learning_rate": 2.3379740019998465e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2505558.0,
      "step": 5738
    },
    {
      "epoch": 0.4414275824936543,
      "grad_norm": 4.5625,
      "learning_rate": 2.3378586262595183e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 2505884.0,
      "step": 5739
    },
    {
      "epoch": 0.4415044996538728,
      "grad_norm": 4.3125,
      "learning_rate": 2.3377432505191906e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9943342804908752,
      "num_tokens": 2506238.0,
      "step": 5740
    },
    {
      "epoch": 0.44158141681409124,
      "grad_norm": 6.875,
      "learning_rate": 2.3376278747788633e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2506544.0,
      "step": 5741
    },
    {
      "epoch": 0.44165833397430965,
      "grad_norm": 2.109375,
      "learning_rate": 2.3375124990385355e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2507106.0,
      "step": 5742
    },
    {
      "epoch": 0.4417352511345281,
      "grad_norm": 6.0625,
      "learning_rate": 2.337397123298208e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2507402.0,
      "step": 5743
    },
    {
      "epoch": 0.44181216829474657,
      "grad_norm": 3.921875,
      "learning_rate": 2.3372817475578805e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2507722.0,
      "step": 5744
    },
    {
      "epoch": 0.441889085454965,
      "grad_norm": 7.1875,
      "learning_rate": 2.3371663718175523e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2508062.0,
      "step": 5745
    },
    {
      "epoch": 0.44196600261518343,
      "grad_norm": 5.0625,
      "learning_rate": 2.337050996077225e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 2508674.0,
      "step": 5746
    },
    {
      "epoch": 0.4420429197754019,
      "grad_norm": 2.890625,
      "learning_rate": 2.3369356203368973e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2509222.0,
      "step": 5747
    },
    {
      "epoch": 0.44211983693562035,
      "grad_norm": 7.1875,
      "learning_rate": 2.3368202445965695e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2509504.0,
      "step": 5748
    },
    {
      "epoch": 0.4421967540958388,
      "grad_norm": 4.59375,
      "learning_rate": 2.336704868856242e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2509822.0,
      "step": 5749
    },
    {
      "epoch": 0.4422736712560572,
      "grad_norm": 5.8125,
      "learning_rate": 2.336589493115914e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9970674514770508,
      "num_tokens": 2510164.0,
      "step": 5750
    },
    {
      "epoch": 0.44235058841627567,
      "grad_norm": 7.53125,
      "learning_rate": 2.3364741173755863e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2510446.0,
      "step": 5751
    },
    {
      "epoch": 0.44242750557649413,
      "grad_norm": 7.40625,
      "learning_rate": 2.336358741635259e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2510728.0,
      "step": 5752
    },
    {
      "epoch": 0.4425044227367126,
      "grad_norm": 5.0625,
      "learning_rate": 2.3362433658949312e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2511050.0,
      "step": 5753
    },
    {
      "epoch": 0.442581339896931,
      "grad_norm": 11.1875,
      "learning_rate": 2.3361279901546035e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9781022071838379,
      "num_tokens": 2511326.0,
      "step": 5754
    },
    {
      "epoch": 0.44265825705714945,
      "grad_norm": 7.5625,
      "learning_rate": 2.336012614414276e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2511604.0,
      "step": 5755
    },
    {
      "epoch": 0.4427351742173679,
      "grad_norm": 5.34375,
      "learning_rate": 2.335897238673948e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 2511942.0,
      "step": 5756
    },
    {
      "epoch": 0.4428120913775863,
      "grad_norm": 3.453125,
      "learning_rate": 2.3357818629336207e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 2512436.0,
      "step": 5757
    },
    {
      "epoch": 0.4428890085378048,
      "grad_norm": 6.84375,
      "learning_rate": 2.335666487193293e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2512752.0,
      "step": 5758
    },
    {
      "epoch": 0.44296592569802323,
      "grad_norm": 3.859375,
      "learning_rate": 2.335551111452965e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2513068.0,
      "step": 5759
    },
    {
      "epoch": 0.4430428428582417,
      "grad_norm": 4.21875,
      "learning_rate": 2.335435735712638e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 2513378.0,
      "step": 5760
    },
    {
      "epoch": 0.4431197600184601,
      "grad_norm": 4.40625,
      "learning_rate": 2.3353203599723097e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2513948.0,
      "step": 5761
    },
    {
      "epoch": 0.44319667717867856,
      "grad_norm": 3.640625,
      "learning_rate": 2.335204984231982e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2514260.0,
      "step": 5762
    },
    {
      "epoch": 0.443273594338897,
      "grad_norm": 5.0625,
      "learning_rate": 2.3350896084916547e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2514846.0,
      "step": 5763
    },
    {
      "epoch": 0.4433505114991155,
      "grad_norm": 3.859375,
      "learning_rate": 2.334974232751327e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2515430.0,
      "step": 5764
    },
    {
      "epoch": 0.4434274286593339,
      "grad_norm": 5.875,
      "learning_rate": 2.334858857010999e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2515684.0,
      "step": 5765
    },
    {
      "epoch": 0.44350434581955234,
      "grad_norm": 3.921875,
      "learning_rate": 2.3347434812706714e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2516264.0,
      "step": 5766
    },
    {
      "epoch": 0.4435812629797708,
      "grad_norm": 4.0625,
      "learning_rate": 2.3346281055303437e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2516798.0,
      "step": 5767
    },
    {
      "epoch": 0.44365818013998926,
      "grad_norm": 3.921875,
      "learning_rate": 2.3345127297900164e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2517354.0,
      "step": 5768
    },
    {
      "epoch": 0.44373509730020766,
      "grad_norm": 5.8125,
      "learning_rate": 2.3343973540496886e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2517700.0,
      "step": 5769
    },
    {
      "epoch": 0.4438120144604261,
      "grad_norm": 2.890625,
      "learning_rate": 2.334281978309361e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2518248.0,
      "step": 5770
    },
    {
      "epoch": 0.4438889316206446,
      "grad_norm": 3.765625,
      "learning_rate": 2.3341666025690336e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 2518592.0,
      "step": 5771
    },
    {
      "epoch": 0.44396584878086304,
      "grad_norm": 9.125,
      "learning_rate": 2.3340512268287054e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2518850.0,
      "step": 5772
    },
    {
      "epoch": 0.44404276594108144,
      "grad_norm": 2.546875,
      "learning_rate": 2.3339358510883777e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2519394.0,
      "step": 5773
    },
    {
      "epoch": 0.4441196831012999,
      "grad_norm": 5.4375,
      "learning_rate": 2.3338204753480503e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2519728.0,
      "step": 5774
    },
    {
      "epoch": 0.44419660026151836,
      "grad_norm": 3.46875,
      "learning_rate": 2.3337050996077226e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 2520074.0,
      "step": 5775
    },
    {
      "epoch": 0.44427351742173676,
      "grad_norm": 5.34375,
      "learning_rate": 2.333589723867395e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2520382.0,
      "step": 5776
    },
    {
      "epoch": 0.4443504345819552,
      "grad_norm": 10.3125,
      "learning_rate": 2.333474348127067e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9808428883552551,
      "num_tokens": 2520644.0,
      "step": 5777
    },
    {
      "epoch": 0.4444273517421737,
      "grad_norm": 5.53125,
      "learning_rate": 2.3333589723867394e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 2520930.0,
      "step": 5778
    },
    {
      "epoch": 0.44450426890239214,
      "grad_norm": 3.828125,
      "learning_rate": 2.333243596646412e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2521242.0,
      "step": 5779
    },
    {
      "epoch": 0.44458118606261054,
      "grad_norm": 2.8125,
      "learning_rate": 2.3331282209060843e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2521790.0,
      "step": 5780
    },
    {
      "epoch": 0.444658103222829,
      "grad_norm": 2.0,
      "learning_rate": 2.3330128451657566e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 2522360.0,
      "step": 5781
    },
    {
      "epoch": 0.44473502038304746,
      "grad_norm": 2.75,
      "learning_rate": 2.332897469425429e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 2523120.0,
      "step": 5782
    },
    {
      "epoch": 0.4448119375432659,
      "grad_norm": 5.75,
      "learning_rate": 2.332782093685101e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2523452.0,
      "step": 5783
    },
    {
      "epoch": 0.4448888547034843,
      "grad_norm": 5.34375,
      "learning_rate": 2.3326667179447733e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 2523782.0,
      "step": 5784
    },
    {
      "epoch": 0.4449657718637028,
      "grad_norm": 5.96875,
      "learning_rate": 2.332551342204446e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 2524092.0,
      "step": 5785
    },
    {
      "epoch": 0.44504268902392125,
      "grad_norm": 2.671875,
      "learning_rate": 2.3324359664641183e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2524654.0,
      "step": 5786
    },
    {
      "epoch": 0.4451196061841397,
      "grad_norm": 3.046875,
      "learning_rate": 2.3323205907237905e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2525200.0,
      "step": 5787
    },
    {
      "epoch": 0.4451965233443581,
      "grad_norm": 7.78125,
      "learning_rate": 2.332205214983463e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2525482.0,
      "step": 5788
    },
    {
      "epoch": 0.44527344050457657,
      "grad_norm": 5.84375,
      "learning_rate": 2.332089839243135e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 2525838.0,
      "step": 5789
    },
    {
      "epoch": 0.445350357664795,
      "grad_norm": 4.1875,
      "learning_rate": 2.3319744635028073e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2526156.0,
      "step": 5790
    },
    {
      "epoch": 0.4454272748250135,
      "grad_norm": 3.484375,
      "learning_rate": 2.33185908776248e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 2526502.0,
      "step": 5791
    },
    {
      "epoch": 0.4455041919852319,
      "grad_norm": 3.5,
      "learning_rate": 2.3317437120221523e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2527098.0,
      "step": 5792
    },
    {
      "epoch": 0.44558110914545035,
      "grad_norm": 2.421875,
      "learning_rate": 2.3316283362818245e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 2527632.0,
      "step": 5793
    },
    {
      "epoch": 0.4456580263056688,
      "grad_norm": 4.875,
      "learning_rate": 2.3315129605414968e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2527930.0,
      "step": 5794
    },
    {
      "epoch": 0.4457349434658872,
      "grad_norm": 3.203125,
      "learning_rate": 2.331397584801169e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 2528238.0,
      "step": 5795
    },
    {
      "epoch": 0.44581186062610567,
      "grad_norm": 2.453125,
      "learning_rate": 2.3312822090608417e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2528812.0,
      "step": 5796
    },
    {
      "epoch": 0.44588877778632413,
      "grad_norm": 6.71875,
      "learning_rate": 2.331166833320514e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2529072.0,
      "step": 5797
    },
    {
      "epoch": 0.4459656949465426,
      "grad_norm": 4.71875,
      "learning_rate": 2.3310514575801862e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 2529406.0,
      "step": 5798
    },
    {
      "epoch": 0.446042612106761,
      "grad_norm": 1.9921875,
      "learning_rate": 2.3309360818398585e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 2530210.0,
      "step": 5799
    },
    {
      "epoch": 0.44611952926697945,
      "grad_norm": 8.8125,
      "learning_rate": 2.3308207060995307e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2530464.0,
      "step": 5800
    },
    {
      "epoch": 0.4461964464271979,
      "grad_norm": 3.15625,
      "learning_rate": 2.330705330359203e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2530782.0,
      "step": 5801
    },
    {
      "epoch": 0.44627336358741637,
      "grad_norm": 5.84375,
      "learning_rate": 2.3305899546188757e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2531090.0,
      "step": 5802
    },
    {
      "epoch": 0.4463502807476348,
      "grad_norm": 2.234375,
      "learning_rate": 2.330474578878548e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 2531874.0,
      "step": 5803
    },
    {
      "epoch": 0.44642719790785323,
      "grad_norm": 2.1875,
      "learning_rate": 2.33035920313822e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2532422.0,
      "step": 5804
    },
    {
      "epoch": 0.4465041150680717,
      "grad_norm": 5.09375,
      "learning_rate": 2.3302438273978925e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2532718.0,
      "step": 5805
    },
    {
      "epoch": 0.44658103222829015,
      "grad_norm": 2.3125,
      "learning_rate": 2.3301284516575647e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 2533504.0,
      "step": 5806
    },
    {
      "epoch": 0.44665794938850856,
      "grad_norm": 6.21875,
      "learning_rate": 2.3300130759172374e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 2533852.0,
      "step": 5807
    },
    {
      "epoch": 0.446734866548727,
      "grad_norm": 3.140625,
      "learning_rate": 2.3298977001769097e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2534146.0,
      "step": 5808
    },
    {
      "epoch": 0.4468117837089455,
      "grad_norm": 2.90625,
      "learning_rate": 2.329782324436582e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2534972.0,
      "step": 5809
    },
    {
      "epoch": 0.44688870086916394,
      "grad_norm": 2.140625,
      "learning_rate": 2.329666948696254e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2535552.0,
      "step": 5810
    },
    {
      "epoch": 0.44696561802938234,
      "grad_norm": 4.4375,
      "learning_rate": 2.3295515729559264e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2535836.0,
      "step": 5811
    },
    {
      "epoch": 0.4470425351896008,
      "grad_norm": 2.078125,
      "learning_rate": 2.3294361972155987e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 2536880.0,
      "step": 5812
    },
    {
      "epoch": 0.44711945234981926,
      "grad_norm": 3.84375,
      "learning_rate": 2.3293208214752714e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2537202.0,
      "step": 5813
    },
    {
      "epoch": 0.4471963695100377,
      "grad_norm": 4.25,
      "learning_rate": 2.3292054457349436e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2537552.0,
      "step": 5814
    },
    {
      "epoch": 0.4472732866702561,
      "grad_norm": 4.78125,
      "learning_rate": 2.329090069994616e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2537836.0,
      "step": 5815
    },
    {
      "epoch": 0.4473502038304746,
      "grad_norm": 4.5,
      "learning_rate": 2.328974694254288e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2538120.0,
      "step": 5816
    },
    {
      "epoch": 0.44742712099069304,
      "grad_norm": 4.5,
      "learning_rate": 2.3288593185139604e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2538436.0,
      "step": 5817
    },
    {
      "epoch": 0.44750403815091144,
      "grad_norm": 5.375,
      "learning_rate": 2.328743942773633e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 2538724.0,
      "step": 5818
    },
    {
      "epoch": 0.4475809553111299,
      "grad_norm": 3.484375,
      "learning_rate": 2.3286285670333053e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2539244.0,
      "step": 5819
    },
    {
      "epoch": 0.44765787247134836,
      "grad_norm": 8.375,
      "learning_rate": 2.3285131912929776e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 2539518.0,
      "step": 5820
    },
    {
      "epoch": 0.4477347896315668,
      "grad_norm": 4.375,
      "learning_rate": 2.32839781555265e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2539852.0,
      "step": 5821
    },
    {
      "epoch": 0.4478117067917852,
      "grad_norm": 3.265625,
      "learning_rate": 2.328282439812322e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2540416.0,
      "step": 5822
    },
    {
      "epoch": 0.4478886239520037,
      "grad_norm": 2.421875,
      "learning_rate": 2.3281670640719944e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2540962.0,
      "step": 5823
    },
    {
      "epoch": 0.44796554111222214,
      "grad_norm": 2.640625,
      "learning_rate": 2.328051688331667e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2541546.0,
      "step": 5824
    },
    {
      "epoch": 0.4480424582724406,
      "grad_norm": 5.4375,
      "learning_rate": 2.3279363125913393e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2541844.0,
      "step": 5825
    },
    {
      "epoch": 0.448119375432659,
      "grad_norm": 3.734375,
      "learning_rate": 2.3278209368510116e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2542128.0,
      "step": 5826
    },
    {
      "epoch": 0.44819629259287747,
      "grad_norm": 3.296875,
      "learning_rate": 2.327705561110684e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2542954.0,
      "step": 5827
    },
    {
      "epoch": 0.4482732097530959,
      "grad_norm": 7.21875,
      "learning_rate": 2.327590185370356e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2543258.0,
      "step": 5828
    },
    {
      "epoch": 0.4483501269133144,
      "grad_norm": 6.59375,
      "learning_rate": 2.3274748096300288e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 2543546.0,
      "step": 5829
    },
    {
      "epoch": 0.4484270440735328,
      "grad_norm": 1.953125,
      "learning_rate": 2.327359433889701e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2544088.0,
      "step": 5830
    },
    {
      "epoch": 0.44850396123375125,
      "grad_norm": 6.90625,
      "learning_rate": 2.327244058149373e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 2544390.0,
      "step": 5831
    },
    {
      "epoch": 0.4485808783939697,
      "grad_norm": 3.734375,
      "learning_rate": 2.3271286824090455e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2544708.0,
      "step": 5832
    },
    {
      "epoch": 0.44865779555418817,
      "grad_norm": 6.46875,
      "learning_rate": 2.327013306668718e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 2545040.0,
      "step": 5833
    },
    {
      "epoch": 0.44873471271440657,
      "grad_norm": 3.609375,
      "learning_rate": 2.32689793092839e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2545352.0,
      "step": 5834
    },
    {
      "epoch": 0.44881162987462503,
      "grad_norm": 2.65625,
      "learning_rate": 2.3267825551880627e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.996874988079071,
      "num_tokens": 2546636.0,
      "step": 5835
    },
    {
      "epoch": 0.4488885470348435,
      "grad_norm": 2.296875,
      "learning_rate": 2.326667179447735e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 2547422.0,
      "step": 5836
    },
    {
      "epoch": 0.4489654641950619,
      "grad_norm": 3.78125,
      "learning_rate": 2.3265518037074073e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2548008.0,
      "step": 5837
    },
    {
      "epoch": 0.44904238135528035,
      "grad_norm": 3.25,
      "learning_rate": 2.3264364279670795e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.990439772605896,
      "num_tokens": 2548532.0,
      "step": 5838
    },
    {
      "epoch": 0.4491192985154988,
      "grad_norm": 3.765625,
      "learning_rate": 2.3263210522267518e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2548852.0,
      "step": 5839
    },
    {
      "epoch": 0.44919621567571727,
      "grad_norm": 2.9375,
      "learning_rate": 2.326205676486424e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 2549368.0,
      "step": 5840
    },
    {
      "epoch": 0.4492731328359357,
      "grad_norm": 5.59375,
      "learning_rate": 2.3260903007460967e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 2549970.0,
      "step": 5841
    },
    {
      "epoch": 0.44935004999615413,
      "grad_norm": 6.6875,
      "learning_rate": 2.3259749250057685e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 2550254.0,
      "step": 5842
    },
    {
      "epoch": 0.4494269671563726,
      "grad_norm": 7.6875,
      "learning_rate": 2.3258595492654412e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 2550530.0,
      "step": 5843
    },
    {
      "epoch": 0.44950388431659105,
      "grad_norm": 6.40625,
      "learning_rate": 2.3257441735251135e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 2550836.0,
      "step": 5844
    },
    {
      "epoch": 0.44958080147680946,
      "grad_norm": 3.953125,
      "learning_rate": 2.3256287977847857e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 2551148.0,
      "step": 5845
    },
    {
      "epoch": 0.4496577186370279,
      "grad_norm": 5.40625,
      "learning_rate": 2.3255134220444584e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9970845580101013,
      "num_tokens": 2551492.0,
      "step": 5846
    },
    {
      "epoch": 0.4497346357972464,
      "grad_norm": 2.796875,
      "learning_rate": 2.3253980463041307e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9910233616828918,
      "num_tokens": 2552050.0,
      "step": 5847
    },
    {
      "epoch": 0.44981155295746483,
      "grad_norm": 2.25,
      "learning_rate": 2.3252826705638025e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9971292018890381,
      "num_tokens": 2553098.0,
      "step": 5848
    },
    {
      "epoch": 0.44988847011768324,
      "grad_norm": 6.375,
      "learning_rate": 2.325167294823475e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2553380.0,
      "step": 5849
    },
    {
      "epoch": 0.4499653872779017,
      "grad_norm": 3.25,
      "learning_rate": 2.3250519190831475e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 2553958.0,
      "step": 5850
    },
    {
      "epoch": 0.45004230443812016,
      "grad_norm": 4.8125,
      "learning_rate": 2.3249365433428197e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2554284.0,
      "step": 5851
    },
    {
      "epoch": 0.4501192215983386,
      "grad_norm": 8.0,
      "learning_rate": 2.3248211676024924e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2554564.0,
      "step": 5852
    },
    {
      "epoch": 0.450196138758557,
      "grad_norm": 8.8125,
      "learning_rate": 2.3247057918621642e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2554866.0,
      "step": 5853
    },
    {
      "epoch": 0.4502730559187755,
      "grad_norm": 4.03125,
      "learning_rate": 2.324590416121837e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2555186.0,
      "step": 5854
    },
    {
      "epoch": 0.45034997307899394,
      "grad_norm": 6.84375,
      "learning_rate": 2.324475040381509e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2555500.0,
      "step": 5855
    },
    {
      "epoch": 0.45042689023921234,
      "grad_norm": 4.0625,
      "learning_rate": 2.3243596646411814e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 2556336.0,
      "step": 5856
    },
    {
      "epoch": 0.4505038073994308,
      "grad_norm": 3.828125,
      "learning_rate": 2.324244288900854e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2556658.0,
      "step": 5857
    },
    {
      "epoch": 0.45058072455964926,
      "grad_norm": 8.3125,
      "learning_rate": 2.3241289131605264e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2556986.0,
      "step": 5858
    },
    {
      "epoch": 0.4506576417198677,
      "grad_norm": 4.4375,
      "learning_rate": 2.324013537420198e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2557332.0,
      "step": 5859
    },
    {
      "epoch": 0.4507345588800861,
      "grad_norm": 2.09375,
      "learning_rate": 2.323898161679871e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 2558136.0,
      "step": 5860
    },
    {
      "epoch": 0.4508114760403046,
      "grad_norm": 10.1875,
      "learning_rate": 2.323782785939543e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2558456.0,
      "step": 5861
    },
    {
      "epoch": 0.45088839320052304,
      "grad_norm": 6.09375,
      "learning_rate": 2.3236674101992154e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2558752.0,
      "step": 5862
    },
    {
      "epoch": 0.4509653103607415,
      "grad_norm": 3.828125,
      "learning_rate": 2.323552034458888e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2559342.0,
      "step": 5863
    },
    {
      "epoch": 0.4510422275209599,
      "grad_norm": 2.515625,
      "learning_rate": 2.32343665871856e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2560162.0,
      "step": 5864
    },
    {
      "epoch": 0.45111914468117836,
      "grad_norm": 5.65625,
      "learning_rate": 2.3233212829782326e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 2560754.0,
      "step": 5865
    },
    {
      "epoch": 0.4511960618413968,
      "grad_norm": 2.875,
      "learning_rate": 2.323205907237905e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2561316.0,
      "step": 5866
    },
    {
      "epoch": 0.4512729790016153,
      "grad_norm": 3.53125,
      "learning_rate": 2.323090531497577e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 2562132.0,
      "step": 5867
    },
    {
      "epoch": 0.4513498961618337,
      "grad_norm": 4.75,
      "learning_rate": 2.32297515575725e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 2562468.0,
      "step": 5868
    },
    {
      "epoch": 0.45142681332205215,
      "grad_norm": 2.125,
      "learning_rate": 2.3228597800169216e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2563288.0,
      "step": 5869
    },
    {
      "epoch": 0.4515037304822706,
      "grad_norm": 6.8125,
      "learning_rate": 2.322744404276594e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 2563598.0,
      "step": 5870
    },
    {
      "epoch": 0.45158064764248906,
      "grad_norm": 3.078125,
      "learning_rate": 2.3226290285362666e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2564178.0,
      "step": 5871
    },
    {
      "epoch": 0.45165756480270747,
      "grad_norm": 8.125,
      "learning_rate": 2.322513652795939e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2564508.0,
      "step": 5872
    },
    {
      "epoch": 0.4517344819629259,
      "grad_norm": 8.25,
      "learning_rate": 2.322398277055611e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 2564850.0,
      "step": 5873
    },
    {
      "epoch": 0.4518113991231444,
      "grad_norm": 2.953125,
      "learning_rate": 2.3222829013152838e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2565396.0,
      "step": 5874
    },
    {
      "epoch": 0.45188831628336285,
      "grad_norm": 7.15625,
      "learning_rate": 2.3221675255749556e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2565706.0,
      "step": 5875
    },
    {
      "epoch": 0.45196523344358125,
      "grad_norm": 5.9375,
      "learning_rate": 2.3220521498346283e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2566046.0,
      "step": 5876
    },
    {
      "epoch": 0.4520421506037997,
      "grad_norm": 6.6875,
      "learning_rate": 2.3219367740943005e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 2566366.0,
      "step": 5877
    },
    {
      "epoch": 0.45211906776401817,
      "grad_norm": 10.75,
      "learning_rate": 2.321821398353973e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9790209531784058,
      "num_tokens": 2566654.0,
      "step": 5878
    },
    {
      "epoch": 0.45219598492423657,
      "grad_norm": 5.21875,
      "learning_rate": 2.3217060226136455e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 2567016.0,
      "step": 5879
    },
    {
      "epoch": 0.45227290208445503,
      "grad_norm": 2.734375,
      "learning_rate": 2.3215906468733173e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2567562.0,
      "step": 5880
    },
    {
      "epoch": 0.4523498192446735,
      "grad_norm": 2.953125,
      "learning_rate": 2.3214752711329896e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 2568142.0,
      "step": 5881
    },
    {
      "epoch": 0.45242673640489195,
      "grad_norm": 5.53125,
      "learning_rate": 2.3213598953926623e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2568454.0,
      "step": 5882
    },
    {
      "epoch": 0.45250365356511035,
      "grad_norm": 5.03125,
      "learning_rate": 2.3212445196523345e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 2568790.0,
      "step": 5883
    },
    {
      "epoch": 0.4525805707253288,
      "grad_norm": 5.28125,
      "learning_rate": 2.3211291439120068e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 2569382.0,
      "step": 5884
    },
    {
      "epoch": 0.45265748788554727,
      "grad_norm": 10.3125,
      "learning_rate": 2.3210137681716795e-06,
      "loss": 0.0407,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2569638.0,
      "step": 5885
    },
    {
      "epoch": 0.45273440504576573,
      "grad_norm": 4.8125,
      "learning_rate": 2.3208983924313513e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2570192.0,
      "step": 5886
    },
    {
      "epoch": 0.45281132220598413,
      "grad_norm": 3.125,
      "learning_rate": 2.320783016691024e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 2570716.0,
      "step": 5887
    },
    {
      "epoch": 0.4528882393662026,
      "grad_norm": 2.96875,
      "learning_rate": 2.3206676409506962e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2571274.0,
      "step": 5888
    },
    {
      "epoch": 0.45296515652642105,
      "grad_norm": 4.40625,
      "learning_rate": 2.3205522652103685e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2571596.0,
      "step": 5889
    },
    {
      "epoch": 0.4530420736866395,
      "grad_norm": 3.984375,
      "learning_rate": 2.3204368894700407e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2571918.0,
      "step": 5890
    },
    {
      "epoch": 0.4531189908468579,
      "grad_norm": 5.5,
      "learning_rate": 2.320321513729713e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 2572230.0,
      "step": 5891
    },
    {
      "epoch": 0.4531959080070764,
      "grad_norm": 9.0625,
      "learning_rate": 2.3202061379893853e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2572566.0,
      "step": 5892
    },
    {
      "epoch": 0.45327282516729483,
      "grad_norm": 4.25,
      "learning_rate": 2.320090762249058e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2572886.0,
      "step": 5893
    },
    {
      "epoch": 0.4533497423275133,
      "grad_norm": 2.78125,
      "learning_rate": 2.31997538650873e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2573442.0,
      "step": 5894
    },
    {
      "epoch": 0.4534266594877317,
      "grad_norm": 5.28125,
      "learning_rate": 2.3198600107684025e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2573760.0,
      "step": 5895
    },
    {
      "epoch": 0.45350357664795016,
      "grad_norm": 5.40625,
      "learning_rate": 2.3197446350280747e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2574082.0,
      "step": 5896
    },
    {
      "epoch": 0.4535804938081686,
      "grad_norm": 7.4375,
      "learning_rate": 2.319629259287747e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2574364.0,
      "step": 5897
    },
    {
      "epoch": 0.453657410968387,
      "grad_norm": 2.765625,
      "learning_rate": 2.3195138835474192e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2574926.0,
      "step": 5898
    },
    {
      "epoch": 0.4537343281286055,
      "grad_norm": 6.625,
      "learning_rate": 2.319398507807092e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 2575264.0,
      "step": 5899
    },
    {
      "epoch": 0.45381124528882394,
      "grad_norm": 5.6875,
      "learning_rate": 2.319283132066764e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2575550.0,
      "step": 5900
    },
    {
      "epoch": 0.4538881624490424,
      "grad_norm": 2.3125,
      "learning_rate": 2.3191677563264364e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 2576360.0,
      "step": 5901
    },
    {
      "epoch": 0.4539650796092608,
      "grad_norm": 2.6875,
      "learning_rate": 2.3190523805861087e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2576906.0,
      "step": 5902
    },
    {
      "epoch": 0.45404199676947926,
      "grad_norm": 6.46875,
      "learning_rate": 2.318937004845781e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 2577184.0,
      "step": 5903
    },
    {
      "epoch": 0.4541189139296977,
      "grad_norm": 3.53125,
      "learning_rate": 2.3188216291054536e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2577506.0,
      "step": 5904
    },
    {
      "epoch": 0.4541958310899162,
      "grad_norm": 2.453125,
      "learning_rate": 2.318706253365126e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2578024.0,
      "step": 5905
    },
    {
      "epoch": 0.4542727482501346,
      "grad_norm": 9.5,
      "learning_rate": 2.318590877624798e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2578328.0,
      "step": 5906
    },
    {
      "epoch": 0.45434966541035304,
      "grad_norm": 2.984375,
      "learning_rate": 2.3184755018844704e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2578844.0,
      "step": 5907
    },
    {
      "epoch": 0.4544265825705715,
      "grad_norm": 9.4375,
      "learning_rate": 2.3183601261441427e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2579144.0,
      "step": 5908
    },
    {
      "epoch": 0.45450349973078996,
      "grad_norm": 2.625,
      "learning_rate": 2.318244750403815e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2579662.0,
      "step": 5909
    },
    {
      "epoch": 0.45458041689100837,
      "grad_norm": 12.5,
      "learning_rate": 2.3181293746634876e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2580002.0,
      "step": 5910
    },
    {
      "epoch": 0.4546573340512268,
      "grad_norm": 6.09375,
      "learning_rate": 2.31801399892316e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 2580340.0,
      "step": 5911
    },
    {
      "epoch": 0.4547342512114453,
      "grad_norm": 3.890625,
      "learning_rate": 2.317898623182832e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 2580862.0,
      "step": 5912
    },
    {
      "epoch": 0.45481116837166374,
      "grad_norm": 3.078125,
      "learning_rate": 2.3177832474425044e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2581410.0,
      "step": 5913
    },
    {
      "epoch": 0.45488808553188215,
      "grad_norm": 2.90625,
      "learning_rate": 2.3176678717021766e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 2581952.0,
      "step": 5914
    },
    {
      "epoch": 0.4549650026921006,
      "grad_norm": 3.453125,
      "learning_rate": 2.3175524959618493e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 2582740.0,
      "step": 5915
    },
    {
      "epoch": 0.45504191985231907,
      "grad_norm": 7.09375,
      "learning_rate": 2.3174371202215216e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2583040.0,
      "step": 5916
    },
    {
      "epoch": 0.45511883701253747,
      "grad_norm": 6.53125,
      "learning_rate": 2.317321744481194e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2583300.0,
      "step": 5917
    },
    {
      "epoch": 0.45519575417275593,
      "grad_norm": 6.875,
      "learning_rate": 2.317206368740866e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2583596.0,
      "step": 5918
    },
    {
      "epoch": 0.4552726713329744,
      "grad_norm": 6.15625,
      "learning_rate": 2.3170909930005383e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2583900.0,
      "step": 5919
    },
    {
      "epoch": 0.45534958849319285,
      "grad_norm": 4.5,
      "learning_rate": 2.3169756172602106e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9819819927215576,
      "num_tokens": 2584234.0,
      "step": 5920
    },
    {
      "epoch": 0.45542650565341125,
      "grad_norm": 5.09375,
      "learning_rate": 2.3168602415198833e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 2584830.0,
      "step": 5921
    },
    {
      "epoch": 0.4555034228136297,
      "grad_norm": 5.46875,
      "learning_rate": 2.3167448657795555e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2585116.0,
      "step": 5922
    },
    {
      "epoch": 0.45558033997384817,
      "grad_norm": 6.25,
      "learning_rate": 2.316629490039228e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 2585416.0,
      "step": 5923
    },
    {
      "epoch": 0.45565725713406663,
      "grad_norm": 4.5625,
      "learning_rate": 2.3165141142989e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2585980.0,
      "step": 5924
    },
    {
      "epoch": 0.45573417429428503,
      "grad_norm": 10.25,
      "learning_rate": 2.3163987385585723e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 2586358.0,
      "step": 5925
    },
    {
      "epoch": 0.4558110914545035,
      "grad_norm": 5.875,
      "learning_rate": 2.316283362818245e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2586690.0,
      "step": 5926
    },
    {
      "epoch": 0.45588800861472195,
      "grad_norm": 7.46875,
      "learning_rate": 2.3161679870779173e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 2587000.0,
      "step": 5927
    },
    {
      "epoch": 0.4559649257749404,
      "grad_norm": 2.203125,
      "learning_rate": 2.3160526113375895e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 2587808.0,
      "step": 5928
    },
    {
      "epoch": 0.4560418429351588,
      "grad_norm": 1.7578125,
      "learning_rate": 2.3159372355972618e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2588366.0,
      "step": 5929
    },
    {
      "epoch": 0.4561187600953773,
      "grad_norm": 9.1875,
      "learning_rate": 2.315821859856934e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.992337167263031,
      "num_tokens": 2588628.0,
      "step": 5930
    },
    {
      "epoch": 0.45619567725559573,
      "grad_norm": 3.40625,
      "learning_rate": 2.3157064841166063e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 2589230.0,
      "step": 5931
    },
    {
      "epoch": 0.4562725944158142,
      "grad_norm": 6.875,
      "learning_rate": 2.315591108376279e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2589536.0,
      "step": 5932
    },
    {
      "epoch": 0.4563495115760326,
      "grad_norm": 4.53125,
      "learning_rate": 2.3154757326359512e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9966777563095093,
      "num_tokens": 2589838.0,
      "step": 5933
    },
    {
      "epoch": 0.45642642873625106,
      "grad_norm": 8.5,
      "learning_rate": 2.3153603568956235e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9823943376541138,
      "num_tokens": 2590124.0,
      "step": 5934
    },
    {
      "epoch": 0.4565033458964695,
      "grad_norm": 4.28125,
      "learning_rate": 2.3152449811552957e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2590474.0,
      "step": 5935
    },
    {
      "epoch": 0.456580263056688,
      "grad_norm": 11.3125,
      "learning_rate": 2.315129605414968e-06,
      "loss": 0.0426,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 2590750.0,
      "step": 5936
    },
    {
      "epoch": 0.4566571802169064,
      "grad_norm": 2.25,
      "learning_rate": 2.3150142296746407e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9909256100654602,
      "num_tokens": 2591302.0,
      "step": 5937
    },
    {
      "epoch": 0.45673409737712484,
      "grad_norm": 8.125,
      "learning_rate": 2.314898853934313e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2591606.0,
      "step": 5938
    },
    {
      "epoch": 0.4568110145373433,
      "grad_norm": 1.96875,
      "learning_rate": 2.314783478193985e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2592402.0,
      "step": 5939
    },
    {
      "epoch": 0.4568879316975617,
      "grad_norm": 5.5,
      "learning_rate": 2.3146681024536575e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 2592688.0,
      "step": 5940
    },
    {
      "epoch": 0.45696484885778016,
      "grad_norm": 2.75,
      "learning_rate": 2.3145527267133297e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2593262.0,
      "step": 5941
    },
    {
      "epoch": 0.4570417660179986,
      "grad_norm": 3.4375,
      "learning_rate": 2.314437350973002e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9959595799446106,
      "num_tokens": 2593758.0,
      "step": 5942
    },
    {
      "epoch": 0.4571186831782171,
      "grad_norm": 3.71875,
      "learning_rate": 2.3143219752326747e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2594298.0,
      "step": 5943
    },
    {
      "epoch": 0.4571956003384355,
      "grad_norm": 2.515625,
      "learning_rate": 2.314206599492347e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2594888.0,
      "step": 5944
    },
    {
      "epoch": 0.45727251749865394,
      "grad_norm": 6.21875,
      "learning_rate": 2.314091223752019e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9895833134651184,
      "num_tokens": 2595178.0,
      "step": 5945
    },
    {
      "epoch": 0.4573494346588724,
      "grad_norm": 2.203125,
      "learning_rate": 2.3139758480116914e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2595736.0,
      "step": 5946
    },
    {
      "epoch": 0.45742635181909086,
      "grad_norm": 3.8125,
      "learning_rate": 2.3138604722713637e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 2596308.0,
      "step": 5947
    },
    {
      "epoch": 0.45750326897930926,
      "grad_norm": 3.234375,
      "learning_rate": 2.3137450965310364e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994575023651123,
      "num_tokens": 2596862.0,
      "step": 5948
    },
    {
      "epoch": 0.4575801861395277,
      "grad_norm": 3.328125,
      "learning_rate": 2.3136297207907086e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 2597430.0,
      "step": 5949
    },
    {
      "epoch": 0.4576571032997462,
      "grad_norm": 2.015625,
      "learning_rate": 2.313514345050381e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 2597954.0,
      "step": 5950
    },
    {
      "epoch": 0.45773402045996464,
      "grad_norm": 3.921875,
      "learning_rate": 2.313398969310053e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 2598252.0,
      "step": 5951
    },
    {
      "epoch": 0.45781093762018304,
      "grad_norm": 3.796875,
      "learning_rate": 2.3132835935697254e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2598576.0,
      "step": 5952
    },
    {
      "epoch": 0.4578878547804015,
      "grad_norm": 7.1875,
      "learning_rate": 2.3131682178293977e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2598858.0,
      "step": 5953
    },
    {
      "epoch": 0.45796477194061996,
      "grad_norm": 2.8125,
      "learning_rate": 2.3130528420890703e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2599414.0,
      "step": 5954
    },
    {
      "epoch": 0.4580416891008384,
      "grad_norm": 6.71875,
      "learning_rate": 2.3129374663487426e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 2599752.0,
      "step": 5955
    },
    {
      "epoch": 0.4581186062610568,
      "grad_norm": 2.65625,
      "learning_rate": 2.312822090608415e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2600296.0,
      "step": 5956
    },
    {
      "epoch": 0.4581955234212753,
      "grad_norm": 4.03125,
      "learning_rate": 2.312706714868087e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 2600622.0,
      "step": 5957
    },
    {
      "epoch": 0.45827244058149375,
      "grad_norm": 6.75,
      "learning_rate": 2.3125913391277594e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2600932.0,
      "step": 5958
    },
    {
      "epoch": 0.45834935774171215,
      "grad_norm": 2.140625,
      "learning_rate": 2.3124759633874316e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 2601492.0,
      "step": 5959
    },
    {
      "epoch": 0.4584262749019306,
      "grad_norm": 3.578125,
      "learning_rate": 2.3123605876471043e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2601814.0,
      "step": 5960
    },
    {
      "epoch": 0.45850319206214907,
      "grad_norm": 6.1875,
      "learning_rate": 2.3122452119067766e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2602132.0,
      "step": 5961
    },
    {
      "epoch": 0.4585801092223675,
      "grad_norm": 3.328125,
      "learning_rate": 2.312129836166449e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2602718.0,
      "step": 5962
    },
    {
      "epoch": 0.45865702638258593,
      "grad_norm": 3.53125,
      "learning_rate": 2.312014460426121e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 2603314.0,
      "step": 5963
    },
    {
      "epoch": 0.4587339435428044,
      "grad_norm": 3.65625,
      "learning_rate": 2.3118990846857934e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2603856.0,
      "step": 5964
    },
    {
      "epoch": 0.45881086070302285,
      "grad_norm": 4.15625,
      "learning_rate": 2.311783708945466e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2604180.0,
      "step": 5965
    },
    {
      "epoch": 0.4588877778632413,
      "grad_norm": 2.5625,
      "learning_rate": 2.3116683332051383e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2604754.0,
      "step": 5966
    },
    {
      "epoch": 0.4589646950234597,
      "grad_norm": 1.9453125,
      "learning_rate": 2.31155295746481e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2605316.0,
      "step": 5967
    },
    {
      "epoch": 0.45904161218367817,
      "grad_norm": 3.21875,
      "learning_rate": 2.311437581724483e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 2605890.0,
      "step": 5968
    },
    {
      "epoch": 0.45911852934389663,
      "grad_norm": 6.46875,
      "learning_rate": 2.311322205984155e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9840764403343201,
      "num_tokens": 2606206.0,
      "step": 5969
    },
    {
      "epoch": 0.4591954465041151,
      "grad_norm": 4.3125,
      "learning_rate": 2.3112068302438273e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2606726.0,
      "step": 5970
    },
    {
      "epoch": 0.4592723636643335,
      "grad_norm": 4.5,
      "learning_rate": 2.3110914545035e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2607048.0,
      "step": 5971
    },
    {
      "epoch": 0.45934928082455195,
      "grad_norm": 5.75,
      "learning_rate": 2.310976078763172e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2607612.0,
      "step": 5972
    },
    {
      "epoch": 0.4594261979847704,
      "grad_norm": 3.921875,
      "learning_rate": 2.3108607030228445e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2608156.0,
      "step": 5973
    },
    {
      "epoch": 0.45950311514498887,
      "grad_norm": 8.25,
      "learning_rate": 2.3107453272825168e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 2608410.0,
      "step": 5974
    },
    {
      "epoch": 0.4595800323052073,
      "grad_norm": 3.078125,
      "learning_rate": 2.310629951542189e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2608956.0,
      "step": 5975
    },
    {
      "epoch": 0.45965694946542573,
      "grad_norm": 2.1875,
      "learning_rate": 2.3105145758018617e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 2609760.0,
      "step": 5976
    },
    {
      "epoch": 0.4597338666256442,
      "grad_norm": 4.1875,
      "learning_rate": 2.310399200061534e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 2610070.0,
      "step": 5977
    },
    {
      "epoch": 0.45981078378586265,
      "grad_norm": 8.6875,
      "learning_rate": 2.310283824321206e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2610392.0,
      "step": 5978
    },
    {
      "epoch": 0.45988770094608106,
      "grad_norm": 10.875,
      "learning_rate": 2.3101684485808785e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2610752.0,
      "step": 5979
    },
    {
      "epoch": 0.4599646181062995,
      "grad_norm": 2.640625,
      "learning_rate": 2.3100530728405508e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2611326.0,
      "step": 5980
    },
    {
      "epoch": 0.460041535266518,
      "grad_norm": 5.78125,
      "learning_rate": 2.309937697100223e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2611888.0,
      "step": 5981
    },
    {
      "epoch": 0.4601184524267364,
      "grad_norm": 2.65625,
      "learning_rate": 2.3098223213598957e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2612434.0,
      "step": 5982
    },
    {
      "epoch": 0.46019536958695484,
      "grad_norm": 6.90625,
      "learning_rate": 2.3097069456195675e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2612740.0,
      "step": 5983
    },
    {
      "epoch": 0.4602722867471733,
      "grad_norm": 4.53125,
      "learning_rate": 2.30959156987924e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 2613052.0,
      "step": 5984
    },
    {
      "epoch": 0.46034920390739176,
      "grad_norm": 4.625,
      "learning_rate": 2.3094761941389125e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2613340.0,
      "step": 5985
    },
    {
      "epoch": 0.46042612106761016,
      "grad_norm": 2.6875,
      "learning_rate": 2.3093608183985847e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2613886.0,
      "step": 5986
    },
    {
      "epoch": 0.4605030382278286,
      "grad_norm": 13.4375,
      "learning_rate": 2.3092454426582574e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 2614186.0,
      "step": 5987
    },
    {
      "epoch": 0.4605799553880471,
      "grad_norm": 8.875,
      "learning_rate": 2.3091300669179297e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 2614514.0,
      "step": 5988
    },
    {
      "epoch": 0.46065687254826554,
      "grad_norm": 9.125,
      "learning_rate": 2.3090146911776015e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2614848.0,
      "step": 5989
    },
    {
      "epoch": 0.46073378970848394,
      "grad_norm": 6.875,
      "learning_rate": 2.308899315437274e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2615156.0,
      "step": 5990
    },
    {
      "epoch": 0.4608107068687024,
      "grad_norm": 5.46875,
      "learning_rate": 2.3087839396969464e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9982788562774658,
      "num_tokens": 2615738.0,
      "step": 5991
    },
    {
      "epoch": 0.46088762402892086,
      "grad_norm": 6.25,
      "learning_rate": 2.3086685639566187e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2616056.0,
      "step": 5992
    },
    {
      "epoch": 0.4609645411891393,
      "grad_norm": 3.140625,
      "learning_rate": 2.3085531882162914e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 2616578.0,
      "step": 5993
    },
    {
      "epoch": 0.4610414583493577,
      "grad_norm": 7.8125,
      "learning_rate": 2.3084378124759632e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2616874.0,
      "step": 5994
    },
    {
      "epoch": 0.4611183755095762,
      "grad_norm": 9.6875,
      "learning_rate": 2.308322436735636e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2617156.0,
      "step": 5995
    },
    {
      "epoch": 0.46119529266979464,
      "grad_norm": 6.4375,
      "learning_rate": 2.308207060995308e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9885057210922241,
      "num_tokens": 2617418.0,
      "step": 5996
    },
    {
      "epoch": 0.4612722098300131,
      "grad_norm": 4.0,
      "learning_rate": 2.3080916852549804e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 2617702.0,
      "step": 5997
    },
    {
      "epoch": 0.4613491269902315,
      "grad_norm": 5.78125,
      "learning_rate": 2.307976309514653e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.996835470199585,
      "num_tokens": 2618020.0,
      "step": 5998
    },
    {
      "epoch": 0.46142604415044997,
      "grad_norm": 7.125,
      "learning_rate": 2.3078609337743254e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 2618346.0,
      "step": 5999
    },
    {
      "epoch": 0.4615029613106684,
      "grad_norm": 13.6875,
      "learning_rate": 2.307745558033997e-06,
      "loss": 0.0293,
      "step": 6000
    },
    {
      "epoch": 0.4615029613106684,
      "eval_loss": 0.02120424434542656,
      "eval_mean_token_accuracy": 0.9902708601951599,
      "eval_num_tokens": 2618684.0,
      "eval_runtime": 21.0958,
      "eval_samples_per_second": 9.481,
      "eval_steps_per_second": 1.185,
      "step": 6000
    },
    {
      "epoch": 0.46157987847088683,
      "grad_norm": 4.0625,
      "learning_rate": 2.30763018229367e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9885631501674652,
      "num_tokens": 2619232.0,
      "step": 6001
    },
    {
      "epoch": 0.4616567956311053,
      "grad_norm": 2.671875,
      "learning_rate": 2.307514806553342e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 2620276.0,
      "step": 6002
    },
    {
      "epoch": 0.46173371279132375,
      "grad_norm": 5.75,
      "learning_rate": 2.3073994308130144e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2620588.0,
      "step": 6003
    },
    {
      "epoch": 0.4618106299515422,
      "grad_norm": 6.15625,
      "learning_rate": 2.307284055072687e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9792284965515137,
      "num_tokens": 2620926.0,
      "step": 6004
    },
    {
      "epoch": 0.4618875471117606,
      "grad_norm": 7.9375,
      "learning_rate": 2.307168679332359e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2621282.0,
      "step": 6005
    },
    {
      "epoch": 0.46196446427197907,
      "grad_norm": 7.15625,
      "learning_rate": 2.3070533035920316e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 2621592.0,
      "step": 6006
    },
    {
      "epoch": 0.46204138143219753,
      "grad_norm": 5.65625,
      "learning_rate": 2.306937927851704e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9926199316978455,
      "num_tokens": 2621864.0,
      "step": 6007
    },
    {
      "epoch": 0.462118298592416,
      "grad_norm": 6.875,
      "learning_rate": 2.306822552111376e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2622220.0,
      "step": 6008
    },
    {
      "epoch": 0.4621952157526344,
      "grad_norm": 3.78125,
      "learning_rate": 2.3067071763710484e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 2622518.0,
      "step": 6009
    },
    {
      "epoch": 0.46227213291285285,
      "grad_norm": 4.75,
      "learning_rate": 2.3065918006307206e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 2622828.0,
      "step": 6010
    },
    {
      "epoch": 0.4623490500730713,
      "grad_norm": 10.875,
      "learning_rate": 2.306476424890393e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9868995547294617,
      "num_tokens": 2623058.0,
      "step": 6011
    },
    {
      "epoch": 0.46242596723328977,
      "grad_norm": 6.53125,
      "learning_rate": 2.3063610491500656e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2623394.0,
      "step": 6012
    },
    {
      "epoch": 0.4625028843935082,
      "grad_norm": 3.203125,
      "learning_rate": 2.306245673409738e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9959595799446106,
      "num_tokens": 2623890.0,
      "step": 6013
    },
    {
      "epoch": 0.46257980155372663,
      "grad_norm": 4.5,
      "learning_rate": 2.30613029766941e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2624212.0,
      "step": 6014
    },
    {
      "epoch": 0.4626567187139451,
      "grad_norm": 7.3125,
      "learning_rate": 2.3060149219290828e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2624468.0,
      "step": 6015
    },
    {
      "epoch": 0.46273363587416355,
      "grad_norm": 3.640625,
      "learning_rate": 2.3058995461887546e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2625048.0,
      "step": 6016
    },
    {
      "epoch": 0.46281055303438196,
      "grad_norm": 2.71875,
      "learning_rate": 2.305784170448427e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2625610.0,
      "step": 6017
    },
    {
      "epoch": 0.4628874701946004,
      "grad_norm": 2.40625,
      "learning_rate": 2.3056687947080995e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2626194.0,
      "step": 6018
    },
    {
      "epoch": 0.4629643873548189,
      "grad_norm": 9.1875,
      "learning_rate": 2.3055534189677718e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 2626524.0,
      "step": 6019
    },
    {
      "epoch": 0.4630413045150373,
      "grad_norm": 6.25,
      "learning_rate": 2.305438043227444e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 2626860.0,
      "step": 6020
    },
    {
      "epoch": 0.46311822167525574,
      "grad_norm": 8.5,
      "learning_rate": 2.3053226674871163e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2627168.0,
      "step": 6021
    },
    {
      "epoch": 0.4631951388354742,
      "grad_norm": 2.875,
      "learning_rate": 2.3052072917467886e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 2627956.0,
      "step": 6022
    },
    {
      "epoch": 0.46327205599569266,
      "grad_norm": 4.59375,
      "learning_rate": 2.3050919160064612e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2628306.0,
      "step": 6023
    },
    {
      "epoch": 0.46334897315591106,
      "grad_norm": 4.5625,
      "learning_rate": 2.3049765402661335e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2628656.0,
      "step": 6024
    },
    {
      "epoch": 0.4634258903161295,
      "grad_norm": 3.53125,
      "learning_rate": 2.3048611645258058e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2629176.0,
      "step": 6025
    },
    {
      "epoch": 0.463502807476348,
      "grad_norm": 2.84375,
      "learning_rate": 2.3047457887854784e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 2630226.0,
      "step": 6026
    },
    {
      "epoch": 0.46357972463656644,
      "grad_norm": 5.90625,
      "learning_rate": 2.3046304130451503e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2630558.0,
      "step": 6027
    },
    {
      "epoch": 0.46365664179678484,
      "grad_norm": 3.3125,
      "learning_rate": 2.3045150373048225e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 2631078.0,
      "step": 6028
    },
    {
      "epoch": 0.4637335589570033,
      "grad_norm": 5.09375,
      "learning_rate": 2.304399661564495e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2631394.0,
      "step": 6029
    },
    {
      "epoch": 0.46381047611722176,
      "grad_norm": 4.0,
      "learning_rate": 2.3042842858241675e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2631688.0,
      "step": 6030
    },
    {
      "epoch": 0.4638873932774402,
      "grad_norm": 2.421875,
      "learning_rate": 2.3041689100838397e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 2632262.0,
      "step": 6031
    },
    {
      "epoch": 0.4639643104376586,
      "grad_norm": 5.25,
      "learning_rate": 2.304053534343512e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2632574.0,
      "step": 6032
    },
    {
      "epoch": 0.4640412275978771,
      "grad_norm": 4.4375,
      "learning_rate": 2.3039381586031842e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 2632918.0,
      "step": 6033
    },
    {
      "epoch": 0.46411814475809554,
      "grad_norm": 7.03125,
      "learning_rate": 2.303822782862857e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2633178.0,
      "step": 6034
    },
    {
      "epoch": 0.464195061918314,
      "grad_norm": 2.984375,
      "learning_rate": 2.303707407122529e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2633722.0,
      "step": 6035
    },
    {
      "epoch": 0.4642719790785324,
      "grad_norm": 5.03125,
      "learning_rate": 2.3035920313822014e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2634044.0,
      "step": 6036
    },
    {
      "epoch": 0.46434889623875086,
      "grad_norm": 7.0,
      "learning_rate": 2.3034766556418737e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 2634388.0,
      "step": 6037
    },
    {
      "epoch": 0.4644258133989693,
      "grad_norm": 5.90625,
      "learning_rate": 2.303361279901546e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2634694.0,
      "step": 6038
    },
    {
      "epoch": 0.4645027305591878,
      "grad_norm": 2.078125,
      "learning_rate": 2.3032459041612182e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2635502.0,
      "step": 6039
    },
    {
      "epoch": 0.4645796477194062,
      "grad_norm": 7.1875,
      "learning_rate": 2.303130528420891e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2635836.0,
      "step": 6040
    },
    {
      "epoch": 0.46465656487962465,
      "grad_norm": 5.03125,
      "learning_rate": 2.303015152680563e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2636186.0,
      "step": 6041
    },
    {
      "epoch": 0.4647334820398431,
      "grad_norm": 12.625,
      "learning_rate": 2.3028997769402354e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 2636526.0,
      "step": 6042
    },
    {
      "epoch": 0.4648103992000615,
      "grad_norm": 4.46875,
      "learning_rate": 2.3027844011999077e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2636848.0,
      "step": 6043
    },
    {
      "epoch": 0.46488731636027997,
      "grad_norm": 5.09375,
      "learning_rate": 2.30266902545958e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 2637204.0,
      "step": 6044
    },
    {
      "epoch": 0.4649642335204984,
      "grad_norm": 8.875,
      "learning_rate": 2.3025536497192526e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2637520.0,
      "step": 6045
    },
    {
      "epoch": 0.4650411506807169,
      "grad_norm": 2.96875,
      "learning_rate": 2.302438273978925e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2638100.0,
      "step": 6046
    },
    {
      "epoch": 0.4651180678409353,
      "grad_norm": 18.625,
      "learning_rate": 2.302322898238597e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 2638424.0,
      "step": 6047
    },
    {
      "epoch": 0.46519498500115375,
      "grad_norm": 8.375,
      "learning_rate": 2.3022075224982694e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 2638736.0,
      "step": 6048
    },
    {
      "epoch": 0.4652719021613722,
      "grad_norm": 12.0625,
      "learning_rate": 2.3020921467579416e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 2639042.0,
      "step": 6049
    },
    {
      "epoch": 0.46534881932159067,
      "grad_norm": 2.71875,
      "learning_rate": 2.301976771017614e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2639600.0,
      "step": 6050
    },
    {
      "epoch": 0.46542573648180907,
      "grad_norm": 2.15625,
      "learning_rate": 2.3018613952772866e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 2640410.0,
      "step": 6051
    },
    {
      "epoch": 0.46550265364202753,
      "grad_norm": 3.71875,
      "learning_rate": 2.301746019536959e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2640732.0,
      "step": 6052
    },
    {
      "epoch": 0.465579570802246,
      "grad_norm": 5.4375,
      "learning_rate": 2.301630643796631e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2641076.0,
      "step": 6053
    },
    {
      "epoch": 0.46565648796246445,
      "grad_norm": 5.28125,
      "learning_rate": 2.3015152680563034e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2641422.0,
      "step": 6054
    },
    {
      "epoch": 0.46573340512268285,
      "grad_norm": 2.625,
      "learning_rate": 2.3013998923159756e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2642202.0,
      "step": 6055
    },
    {
      "epoch": 0.4658103222829013,
      "grad_norm": 8.1875,
      "learning_rate": 2.3012845165756483e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2642456.0,
      "step": 6056
    },
    {
      "epoch": 0.46588723944311977,
      "grad_norm": 13.375,
      "learning_rate": 2.3011691408353206e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2642808.0,
      "step": 6057
    },
    {
      "epoch": 0.46596415660333823,
      "grad_norm": 4.625,
      "learning_rate": 2.301053765094993e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2643118.0,
      "step": 6058
    },
    {
      "epoch": 0.46604107376355663,
      "grad_norm": 5.21875,
      "learning_rate": 2.300938389354665e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 2643702.0,
      "step": 6059
    },
    {
      "epoch": 0.4661179909237751,
      "grad_norm": 4.84375,
      "learning_rate": 2.3008230136143373e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2644250.0,
      "step": 6060
    },
    {
      "epoch": 0.46619490808399355,
      "grad_norm": 2.84375,
      "learning_rate": 2.3007076378740096e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 2644792.0,
      "step": 6061
    },
    {
      "epoch": 0.46627182524421196,
      "grad_norm": 10.125,
      "learning_rate": 2.3005922621336823e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2645116.0,
      "step": 6062
    },
    {
      "epoch": 0.4663487424044304,
      "grad_norm": 14.5,
      "learning_rate": 2.3004768863933545e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2645444.0,
      "step": 6063
    },
    {
      "epoch": 0.4664256595646489,
      "grad_norm": 11.5625,
      "learning_rate": 2.3003615106530268e-06,
      "loss": 0.0425,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 2645676.0,
      "step": 6064
    },
    {
      "epoch": 0.46650257672486733,
      "grad_norm": 9.3125,
      "learning_rate": 2.300246134912699e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2645982.0,
      "step": 6065
    },
    {
      "epoch": 0.46657949388508574,
      "grad_norm": 3.265625,
      "learning_rate": 2.3001307591723713e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2646528.0,
      "step": 6066
    },
    {
      "epoch": 0.4666564110453042,
      "grad_norm": 8.375,
      "learning_rate": 2.3000153834320436e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9850187301635742,
      "num_tokens": 2646796.0,
      "step": 6067
    },
    {
      "epoch": 0.46673332820552266,
      "grad_norm": 2.9375,
      "learning_rate": 2.2999000076917162e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2647354.0,
      "step": 6068
    },
    {
      "epoch": 0.4668102453657411,
      "grad_norm": 8.8125,
      "learning_rate": 2.2997846319513885e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2647636.0,
      "step": 6069
    },
    {
      "epoch": 0.4668871625259595,
      "grad_norm": 4.46875,
      "learning_rate": 2.2996692562110608e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 2648208.0,
      "step": 6070
    },
    {
      "epoch": 0.466964079686178,
      "grad_norm": 7.03125,
      "learning_rate": 2.299553880470733e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2648540.0,
      "step": 6071
    },
    {
      "epoch": 0.46704099684639644,
      "grad_norm": 2.015625,
      "learning_rate": 2.2994385047304053e-06,
      "loss": 0.0068,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 2649584.0,
      "step": 6072
    },
    {
      "epoch": 0.4671179140066149,
      "grad_norm": 5.0,
      "learning_rate": 2.299323128990078e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2649906.0,
      "step": 6073
    },
    {
      "epoch": 0.4671948311668333,
      "grad_norm": 5.71875,
      "learning_rate": 2.2992077532497502e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2650250.0,
      "step": 6074
    },
    {
      "epoch": 0.46727174832705176,
      "grad_norm": 2.84375,
      "learning_rate": 2.299092377509422e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2650808.0,
      "step": 6075
    },
    {
      "epoch": 0.4673486654872702,
      "grad_norm": 4.65625,
      "learning_rate": 2.2989770017690947e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2651352.0,
      "step": 6076
    },
    {
      "epoch": 0.4674255826474887,
      "grad_norm": 6.28125,
      "learning_rate": 2.298861626028767e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 2651892.0,
      "step": 6077
    },
    {
      "epoch": 0.4675024998077071,
      "grad_norm": 2.515625,
      "learning_rate": 2.2987462502884392e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2652438.0,
      "step": 6078
    },
    {
      "epoch": 0.46757941696792554,
      "grad_norm": 3.796875,
      "learning_rate": 2.298630874548112e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 2653008.0,
      "step": 6079
    },
    {
      "epoch": 0.467656334128144,
      "grad_norm": 4.75,
      "learning_rate": 2.298515498807784e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2653598.0,
      "step": 6080
    },
    {
      "epoch": 0.4677332512883624,
      "grad_norm": 6.4375,
      "learning_rate": 2.2984001230674564e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 2653936.0,
      "step": 6081
    },
    {
      "epoch": 0.46781016844858087,
      "grad_norm": 2.09375,
      "learning_rate": 2.2982847473271287e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9987593293190002,
      "num_tokens": 2654744.0,
      "step": 6082
    },
    {
      "epoch": 0.4678870856087993,
      "grad_norm": 3.59375,
      "learning_rate": 2.298169371586801e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2655308.0,
      "step": 6083
    },
    {
      "epoch": 0.4679640027690178,
      "grad_norm": 2.171875,
      "learning_rate": 2.2980539958464736e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2655848.0,
      "step": 6084
    },
    {
      "epoch": 0.4680409199292362,
      "grad_norm": 2.34375,
      "learning_rate": 2.297938620106146e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2656630.0,
      "step": 6085
    },
    {
      "epoch": 0.46811783708945465,
      "grad_norm": 10.25,
      "learning_rate": 2.2978232443658177e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2656924.0,
      "step": 6086
    },
    {
      "epoch": 0.4681947542496731,
      "grad_norm": 5.875,
      "learning_rate": 2.2977078686254904e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2657268.0,
      "step": 6087
    },
    {
      "epoch": 0.46827167140989157,
      "grad_norm": 14.9375,
      "learning_rate": 2.2975924928851627e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2657546.0,
      "step": 6088
    },
    {
      "epoch": 0.46834858857010997,
      "grad_norm": 12.3125,
      "learning_rate": 2.297477117144835e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2657802.0,
      "step": 6089
    },
    {
      "epoch": 0.46842550573032843,
      "grad_norm": 15.5625,
      "learning_rate": 2.2973617414045076e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2658154.0,
      "step": 6090
    },
    {
      "epoch": 0.4685024228905469,
      "grad_norm": 2.109375,
      "learning_rate": 2.29724636566418e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 2658958.0,
      "step": 6091
    },
    {
      "epoch": 0.46857934005076535,
      "grad_norm": 8.5625,
      "learning_rate": 2.297130989923852e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2659252.0,
      "step": 6092
    },
    {
      "epoch": 0.46865625721098375,
      "grad_norm": 10.1875,
      "learning_rate": 2.2970156141835244e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2659580.0,
      "step": 6093
    },
    {
      "epoch": 0.4687331743712022,
      "grad_norm": 6.125,
      "learning_rate": 2.2969002384431966e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2660106.0,
      "step": 6094
    },
    {
      "epoch": 0.46881009153142067,
      "grad_norm": 6.40625,
      "learning_rate": 2.2967848627028693e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2660388.0,
      "step": 6095
    },
    {
      "epoch": 0.46888700869163913,
      "grad_norm": 13.5625,
      "learning_rate": 2.2966694869625416e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2660706.0,
      "step": 6096
    },
    {
      "epoch": 0.46896392585185753,
      "grad_norm": 6.59375,
      "learning_rate": 2.2965541112222134e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2660988.0,
      "step": 6097
    },
    {
      "epoch": 0.469040843012076,
      "grad_norm": 2.90625,
      "learning_rate": 2.296438735481886e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2661528.0,
      "step": 6098
    },
    {
      "epoch": 0.46911776017229445,
      "grad_norm": 8.5625,
      "learning_rate": 2.2963233597415584e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 2661818.0,
      "step": 6099
    },
    {
      "epoch": 0.4691946773325129,
      "grad_norm": 15.8125,
      "learning_rate": 2.2962079840012306e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 2662150.0,
      "step": 6100
    },
    {
      "epoch": 0.4692715944927313,
      "grad_norm": 4.3125,
      "learning_rate": 2.2960926082609033e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9974554777145386,
      "num_tokens": 2662938.0,
      "step": 6101
    },
    {
      "epoch": 0.4693485116529498,
      "grad_norm": 2.703125,
      "learning_rate": 2.2959772325205756e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2663484.0,
      "step": 6102
    },
    {
      "epoch": 0.46942542881316823,
      "grad_norm": 8.875,
      "learning_rate": 2.295861856780248e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2663770.0,
      "step": 6103
    },
    {
      "epoch": 0.46950234597338664,
      "grad_norm": 6.09375,
      "learning_rate": 2.29574648103992e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2664054.0,
      "step": 6104
    },
    {
      "epoch": 0.4695792631336051,
      "grad_norm": 8.5,
      "learning_rate": 2.2956311052995923e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 2664366.0,
      "step": 6105
    },
    {
      "epoch": 0.46965618029382356,
      "grad_norm": 2.40625,
      "learning_rate": 2.295515729559265e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 2665162.0,
      "step": 6106
    },
    {
      "epoch": 0.469733097454042,
      "grad_norm": 3.390625,
      "learning_rate": 2.2954003538189373e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2665718.0,
      "step": 6107
    },
    {
      "epoch": 0.4698100146142604,
      "grad_norm": 9.0625,
      "learning_rate": 2.295284978078609e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 2666056.0,
      "step": 6108
    },
    {
      "epoch": 0.4698869317744789,
      "grad_norm": 5.21875,
      "learning_rate": 2.2951696023382818e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2666378.0,
      "step": 6109
    },
    {
      "epoch": 0.46996384893469734,
      "grad_norm": 7.8125,
      "learning_rate": 2.295054226597954e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 2666690.0,
      "step": 6110
    },
    {
      "epoch": 0.4700407660949158,
      "grad_norm": 4.625,
      "learning_rate": 2.2949388508576263e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2667040.0,
      "step": 6111
    },
    {
      "epoch": 0.4701176832551342,
      "grad_norm": 2.03125,
      "learning_rate": 2.294823475117299e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 2667826.0,
      "step": 6112
    },
    {
      "epoch": 0.47019460041535266,
      "grad_norm": 11.125,
      "learning_rate": 2.294708099376971e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 2668164.0,
      "step": 6113
    },
    {
      "epoch": 0.4702715175755711,
      "grad_norm": 6.4375,
      "learning_rate": 2.2945927236366435e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 2668454.0,
      "step": 6114
    },
    {
      "epoch": 0.4703484347357896,
      "grad_norm": 4.71875,
      "learning_rate": 2.2944773478963158e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2668766.0,
      "step": 6115
    },
    {
      "epoch": 0.470425351896008,
      "grad_norm": 6.0,
      "learning_rate": 2.294361972155988e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2669086.0,
      "step": 6116
    },
    {
      "epoch": 0.47050226905622644,
      "grad_norm": 2.359375,
      "learning_rate": 2.2942465964156603e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2669632.0,
      "step": 6117
    },
    {
      "epoch": 0.4705791862164449,
      "grad_norm": 2.390625,
      "learning_rate": 2.294131220675333e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2670440.0,
      "step": 6118
    },
    {
      "epoch": 0.47065610337666336,
      "grad_norm": 7.1875,
      "learning_rate": 2.294015844935005e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 2671000.0,
      "step": 6119
    },
    {
      "epoch": 0.47073302053688176,
      "grad_norm": 6.90625,
      "learning_rate": 2.2939004691946775e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 2671346.0,
      "step": 6120
    },
    {
      "epoch": 0.4708099376971002,
      "grad_norm": 6.40625,
      "learning_rate": 2.2937850934543497e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2671668.0,
      "step": 6121
    },
    {
      "epoch": 0.4708868548573187,
      "grad_norm": 9.75,
      "learning_rate": 2.293669717714022e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2671976.0,
      "step": 6122
    },
    {
      "epoch": 0.4709637720175371,
      "grad_norm": 5.15625,
      "learning_rate": 2.2935543419736947e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2672320.0,
      "step": 6123
    },
    {
      "epoch": 0.47104068917775554,
      "grad_norm": 5.03125,
      "learning_rate": 2.2934389662333665e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2672664.0,
      "step": 6124
    },
    {
      "epoch": 0.471117606337974,
      "grad_norm": 15.6875,
      "learning_rate": 2.2933235904930388e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2672942.0,
      "step": 6125
    },
    {
      "epoch": 0.47119452349819246,
      "grad_norm": 9.25,
      "learning_rate": 2.2932082147527114e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 2673200.0,
      "step": 6126
    },
    {
      "epoch": 0.47127144065841087,
      "grad_norm": 8.375,
      "learning_rate": 2.2930928390123837e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 2673784.0,
      "step": 6127
    },
    {
      "epoch": 0.4713483578186293,
      "grad_norm": 7.28125,
      "learning_rate": 2.292977463272056e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2674364.0,
      "step": 6128
    },
    {
      "epoch": 0.4714252749788478,
      "grad_norm": 2.984375,
      "learning_rate": 2.2928620875317286e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2674908.0,
      "step": 6129
    },
    {
      "epoch": 0.47150219213906625,
      "grad_norm": 10.5625,
      "learning_rate": 2.2927467117914005e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2675228.0,
      "step": 6130
    },
    {
      "epoch": 0.47157910929928465,
      "grad_norm": 8.75,
      "learning_rate": 2.292631336051073e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2675818.0,
      "step": 6131
    },
    {
      "epoch": 0.4716560264595031,
      "grad_norm": 21.0,
      "learning_rate": 2.2925159603107454e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2676120.0,
      "step": 6132
    },
    {
      "epoch": 0.47173294361972157,
      "grad_norm": 22.5,
      "learning_rate": 2.2924005845704177e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 2676422.0,
      "step": 6133
    },
    {
      "epoch": 0.47180986077994,
      "grad_norm": 8.0625,
      "learning_rate": 2.2922852088300904e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2676742.0,
      "step": 6134
    },
    {
      "epoch": 0.47188677794015843,
      "grad_norm": 4.375,
      "learning_rate": 2.292169833089762e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 2677310.0,
      "step": 6135
    },
    {
      "epoch": 0.4719636951003769,
      "grad_norm": 10.9375,
      "learning_rate": 2.2920544573494345e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2677656.0,
      "step": 6136
    },
    {
      "epoch": 0.47204061226059535,
      "grad_norm": 12.6875,
      "learning_rate": 2.291939081609107e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 2677936.0,
      "step": 6137
    },
    {
      "epoch": 0.4721175294208138,
      "grad_norm": 2.890625,
      "learning_rate": 2.2918237058687794e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2678506.0,
      "step": 6138
    },
    {
      "epoch": 0.4721944465810322,
      "grad_norm": 6.84375,
      "learning_rate": 2.2917083301284516e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 2679050.0,
      "step": 6139
    },
    {
      "epoch": 0.47227136374125067,
      "grad_norm": 3.46875,
      "learning_rate": 2.291592954388124e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2679596.0,
      "step": 6140
    },
    {
      "epoch": 0.47234828090146913,
      "grad_norm": 12.5625,
      "learning_rate": 2.291477578647796e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2679930.0,
      "step": 6141
    },
    {
      "epoch": 0.4724251980616876,
      "grad_norm": 2.953125,
      "learning_rate": 2.291362202907469e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2680504.0,
      "step": 6142
    },
    {
      "epoch": 0.472502115221906,
      "grad_norm": 5.75,
      "learning_rate": 2.291246827167141e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2681052.0,
      "step": 6143
    },
    {
      "epoch": 0.47257903238212445,
      "grad_norm": 4.21875,
      "learning_rate": 2.2911314514268134e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2681370.0,
      "step": 6144
    },
    {
      "epoch": 0.4726559495423429,
      "grad_norm": 5.5,
      "learning_rate": 2.291016075686486e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 2681704.0,
      "step": 6145
    },
    {
      "epoch": 0.4727328667025613,
      "grad_norm": 8.25,
      "learning_rate": 2.290900699946158e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2682010.0,
      "step": 6146
    },
    {
      "epoch": 0.4728097838627798,
      "grad_norm": 14.9375,
      "learning_rate": 2.29078532420583e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2682328.0,
      "step": 6147
    },
    {
      "epoch": 0.47288670102299823,
      "grad_norm": 8.3125,
      "learning_rate": 2.290669948465503e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2682610.0,
      "step": 6148
    },
    {
      "epoch": 0.4729636181832167,
      "grad_norm": 17.25,
      "learning_rate": 2.290554572725175e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2682906.0,
      "step": 6149
    },
    {
      "epoch": 0.4730405353434351,
      "grad_norm": 10.125,
      "learning_rate": 2.2904391969848473e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2683208.0,
      "step": 6150
    },
    {
      "epoch": 0.47311745250365356,
      "grad_norm": 3.515625,
      "learning_rate": 2.2903238212445196e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2683748.0,
      "step": 6151
    },
    {
      "epoch": 0.473194369663872,
      "grad_norm": 5.78125,
      "learning_rate": 2.290208445504192e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2684296.0,
      "step": 6152
    },
    {
      "epoch": 0.4732712868240905,
      "grad_norm": 10.9375,
      "learning_rate": 2.2900930697638645e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 2684636.0,
      "step": 6153
    },
    {
      "epoch": 0.4733482039843089,
      "grad_norm": 7.84375,
      "learning_rate": 2.289977694023537e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 2684914.0,
      "step": 6154
    },
    {
      "epoch": 0.47342512114452734,
      "grad_norm": 16.375,
      "learning_rate": 2.289862318283209e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2685212.0,
      "step": 6155
    },
    {
      "epoch": 0.4735020383047458,
      "grad_norm": 3.53125,
      "learning_rate": 2.2897469425428817e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2685758.0,
      "step": 6156
    },
    {
      "epoch": 0.47357895546496426,
      "grad_norm": 22.125,
      "learning_rate": 2.2896315668025536e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2686108.0,
      "step": 6157
    },
    {
      "epoch": 0.47365587262518266,
      "grad_norm": 19.875,
      "learning_rate": 2.289516191062226e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 2686416.0,
      "step": 6158
    },
    {
      "epoch": 0.4737327897854011,
      "grad_norm": 8.625,
      "learning_rate": 2.2894008153218985e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9970414042472839,
      "num_tokens": 2686756.0,
      "step": 6159
    },
    {
      "epoch": 0.4738097069456196,
      "grad_norm": 21.125,
      "learning_rate": 2.2892854395815708e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2687034.0,
      "step": 6160
    },
    {
      "epoch": 0.47388662410583804,
      "grad_norm": 7.375,
      "learning_rate": 2.289170063841243e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 2687394.0,
      "step": 6161
    },
    {
      "epoch": 0.47396354126605644,
      "grad_norm": 7.375,
      "learning_rate": 2.2890546881009153e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 2687760.0,
      "step": 6162
    },
    {
      "epoch": 0.4740404584262749,
      "grad_norm": 6.15625,
      "learning_rate": 2.2889393123605875e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2688110.0,
      "step": 6163
    },
    {
      "epoch": 0.47411737558649336,
      "grad_norm": 21.125,
      "learning_rate": 2.2888239366202602e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2688410.0,
      "step": 6164
    },
    {
      "epoch": 0.47419429274671177,
      "grad_norm": 7.6875,
      "learning_rate": 2.2887085608799325e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2688956.0,
      "step": 6165
    },
    {
      "epoch": 0.4742712099069302,
      "grad_norm": 2.46875,
      "learning_rate": 2.2885931851396047e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2689530.0,
      "step": 6166
    },
    {
      "epoch": 0.4743481270671487,
      "grad_norm": 6.90625,
      "learning_rate": 2.288477809399277e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2689852.0,
      "step": 6167
    },
    {
      "epoch": 0.47442504422736714,
      "grad_norm": 9.3125,
      "learning_rate": 2.2883624336589493e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 2690426.0,
      "step": 6168
    },
    {
      "epoch": 0.47450196138758555,
      "grad_norm": 8.625,
      "learning_rate": 2.2882470579186215e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 2690994.0,
      "step": 6169
    },
    {
      "epoch": 0.474578878547804,
      "grad_norm": 12.5625,
      "learning_rate": 2.288131682178294e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2691326.0,
      "step": 6170
    },
    {
      "epoch": 0.47465579570802247,
      "grad_norm": 17.875,
      "learning_rate": 2.2880163064379664e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2691608.0,
      "step": 6171
    },
    {
      "epoch": 0.4747327128682409,
      "grad_norm": 25.625,
      "learning_rate": 2.2879009306976387e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 2691884.0,
      "step": 6172
    },
    {
      "epoch": 0.47480963002845933,
      "grad_norm": 25.0,
      "learning_rate": 2.287785554957311e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 2692212.0,
      "step": 6173
    },
    {
      "epoch": 0.4748865471886778,
      "grad_norm": 14.0,
      "learning_rate": 2.2876701792169832e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 2692498.0,
      "step": 6174
    },
    {
      "epoch": 0.47496346434889625,
      "grad_norm": 7.21875,
      "learning_rate": 2.2875548034766555e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2692844.0,
      "step": 6175
    },
    {
      "epoch": 0.4750403815091147,
      "grad_norm": 13.375,
      "learning_rate": 2.287439427736328e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2693136.0,
      "step": 6176
    },
    {
      "epoch": 0.4751172986693331,
      "grad_norm": 7.0625,
      "learning_rate": 2.2873240519960004e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2693418.0,
      "step": 6177
    },
    {
      "epoch": 0.47519421582955157,
      "grad_norm": 7.03125,
      "learning_rate": 2.2872086762556727e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2693702.0,
      "step": 6178
    },
    {
      "epoch": 0.47527113298977003,
      "grad_norm": 8.4375,
      "learning_rate": 2.287093300515345e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2694254.0,
      "step": 6179
    },
    {
      "epoch": 0.4753480501499885,
      "grad_norm": 21.875,
      "learning_rate": 2.286977924775017e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2694828.0,
      "step": 6180
    },
    {
      "epoch": 0.4754249673102069,
      "grad_norm": 4.53125,
      "learning_rate": 2.28686254903469e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2695132.0,
      "step": 6181
    },
    {
      "epoch": 0.47550188447042535,
      "grad_norm": 5.1875,
      "learning_rate": 2.286747173294362e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 2695702.0,
      "step": 6182
    },
    {
      "epoch": 0.4755788016306438,
      "grad_norm": 6.875,
      "learning_rate": 2.2866317975540344e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2696030.0,
      "step": 6183
    },
    {
      "epoch": 0.4756557187908622,
      "grad_norm": 4.09375,
      "learning_rate": 2.2865164218137067e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2696364.0,
      "step": 6184
    },
    {
      "epoch": 0.4757326359510807,
      "grad_norm": 4.6875,
      "learning_rate": 2.286401046073379e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 2696702.0,
      "step": 6185
    },
    {
      "epoch": 0.47580955311129913,
      "grad_norm": 11.75,
      "learning_rate": 2.286285670333051e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2697264.0,
      "step": 6186
    },
    {
      "epoch": 0.4758864702715176,
      "grad_norm": 28.875,
      "learning_rate": 2.286170294592724e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2697624.0,
      "step": 6187
    },
    {
      "epoch": 0.475963387431736,
      "grad_norm": 7.4375,
      "learning_rate": 2.286054918852396e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 2698204.0,
      "step": 6188
    },
    {
      "epoch": 0.47604030459195446,
      "grad_norm": 9.375,
      "learning_rate": 2.2859395431120684e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2698482.0,
      "step": 6189
    },
    {
      "epoch": 0.4761172217521729,
      "grad_norm": 10.125,
      "learning_rate": 2.2858241673717406e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2698764.0,
      "step": 6190
    },
    {
      "epoch": 0.4761941389123914,
      "grad_norm": 11.8125,
      "learning_rate": 2.285708791631413e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2699082.0,
      "step": 6191
    },
    {
      "epoch": 0.4762710560726098,
      "grad_norm": 15.0625,
      "learning_rate": 2.2855934158910856e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2699438.0,
      "step": 6192
    },
    {
      "epoch": 0.47634797323282824,
      "grad_norm": 3.078125,
      "learning_rate": 2.285478040150758e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 2699988.0,
      "step": 6193
    },
    {
      "epoch": 0.4764248903930467,
      "grad_norm": 6.0,
      "learning_rate": 2.28536266441043e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2700524.0,
      "step": 6194
    },
    {
      "epoch": 0.47650180755326516,
      "grad_norm": 6.84375,
      "learning_rate": 2.2852472886701023e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2700810.0,
      "step": 6195
    },
    {
      "epoch": 0.47657872471348356,
      "grad_norm": 2.546875,
      "learning_rate": 2.2851319129297746e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2701362.0,
      "step": 6196
    },
    {
      "epoch": 0.476655641873702,
      "grad_norm": 12.0,
      "learning_rate": 2.285016537189447e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2701722.0,
      "step": 6197
    },
    {
      "epoch": 0.4767325590339205,
      "grad_norm": 2.578125,
      "learning_rate": 2.2849011614491195e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 2702308.0,
      "step": 6198
    },
    {
      "epoch": 0.47680947619413894,
      "grad_norm": 12.6875,
      "learning_rate": 2.284785785708792e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2702602.0,
      "step": 6199
    },
    {
      "epoch": 0.47688639335435734,
      "grad_norm": 18.0,
      "learning_rate": 2.284670409968464e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 2702910.0,
      "step": 6200
    },
    {
      "epoch": 0.4769633105145758,
      "grad_norm": 13.0,
      "learning_rate": 2.2845550342281363e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2703226.0,
      "step": 6201
    },
    {
      "epoch": 0.47704022767479426,
      "grad_norm": 9.3125,
      "learning_rate": 2.2844396584878086e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9893428087234497,
      "num_tokens": 2703790.0,
      "step": 6202
    },
    {
      "epoch": 0.4771171448350127,
      "grad_norm": 7.75,
      "learning_rate": 2.2843242827474812e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2704108.0,
      "step": 6203
    },
    {
      "epoch": 0.4771940619952311,
      "grad_norm": 9.9375,
      "learning_rate": 2.2842089070071535e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2704402.0,
      "step": 6204
    },
    {
      "epoch": 0.4772709791554496,
      "grad_norm": 7.03125,
      "learning_rate": 2.2840935312668258e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9940476417541504,
      "num_tokens": 2704740.0,
      "step": 6205
    },
    {
      "epoch": 0.47734789631566804,
      "grad_norm": 10.375,
      "learning_rate": 2.283978155526498e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 2705078.0,
      "step": 6206
    },
    {
      "epoch": 0.47742481347588644,
      "grad_norm": 13.625,
      "learning_rate": 2.2838627797861703e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 2705386.0,
      "step": 6207
    },
    {
      "epoch": 0.4775017306361049,
      "grad_norm": 5.0,
      "learning_rate": 2.2837474040458425e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 2705734.0,
      "step": 6208
    },
    {
      "epoch": 0.47757864779632336,
      "grad_norm": 6.0,
      "learning_rate": 2.2836320283055152e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2706050.0,
      "step": 6209
    },
    {
      "epoch": 0.4776555649565418,
      "grad_norm": 7.09375,
      "learning_rate": 2.2835166525651875e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2706332.0,
      "step": 6210
    },
    {
      "epoch": 0.4777324821167602,
      "grad_norm": 10.5,
      "learning_rate": 2.2834012768248597e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2707158.0,
      "step": 6211
    },
    {
      "epoch": 0.4778093992769787,
      "grad_norm": 4.53125,
      "learning_rate": 2.283285901084532e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2707678.0,
      "step": 6212
    },
    {
      "epoch": 0.47788631643719715,
      "grad_norm": 6.40625,
      "learning_rate": 2.2831705253442043e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2707932.0,
      "step": 6213
    },
    {
      "epoch": 0.4779632335974156,
      "grad_norm": 14.3125,
      "learning_rate": 2.283055149603877e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.98525071144104,
      "num_tokens": 2708272.0,
      "step": 6214
    },
    {
      "epoch": 0.478040150757634,
      "grad_norm": 5.28125,
      "learning_rate": 2.282939773863549e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2708554.0,
      "step": 6215
    },
    {
      "epoch": 0.47811706791785247,
      "grad_norm": 4.28125,
      "learning_rate": 2.282824398123221e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2708860.0,
      "step": 6216
    },
    {
      "epoch": 0.4781939850780709,
      "grad_norm": 17.25,
      "learning_rate": 2.2827090223828937e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2709206.0,
      "step": 6217
    },
    {
      "epoch": 0.4782709022382894,
      "grad_norm": 9.0625,
      "learning_rate": 2.282593646642566e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 2709482.0,
      "step": 6218
    },
    {
      "epoch": 0.4783478193985078,
      "grad_norm": 4.375,
      "learning_rate": 2.2824782709022382e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 2710262.0,
      "step": 6219
    },
    {
      "epoch": 0.47842473655872625,
      "grad_norm": 35.5,
      "learning_rate": 2.282362895161911e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 2710600.0,
      "step": 6220
    },
    {
      "epoch": 0.4785016537189447,
      "grad_norm": 6.84375,
      "learning_rate": 2.282247519421583e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9966996908187866,
      "num_tokens": 2710904.0,
      "step": 6221
    },
    {
      "epoch": 0.47857857087916317,
      "grad_norm": 9.1875,
      "learning_rate": 2.2821321436812554e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9858155846595764,
      "num_tokens": 2711188.0,
      "step": 6222
    },
    {
      "epoch": 0.47865548803938157,
      "grad_norm": 10.5625,
      "learning_rate": 2.2820167679409277e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 2711768.0,
      "step": 6223
    },
    {
      "epoch": 0.47873240519960003,
      "grad_norm": 3.65625,
      "learning_rate": 2.2819013922006e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 2712340.0,
      "step": 6224
    },
    {
      "epoch": 0.4788093223598185,
      "grad_norm": 17.875,
      "learning_rate": 2.281786016460272e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 2712666.0,
      "step": 6225
    },
    {
      "epoch": 0.4788862395200369,
      "grad_norm": 2.34375,
      "learning_rate": 2.281670640719945e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 2713470.0,
      "step": 6226
    },
    {
      "epoch": 0.47896315668025535,
      "grad_norm": 6.5,
      "learning_rate": 2.2815552649796167e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2713764.0,
      "step": 6227
    },
    {
      "epoch": 0.4790400738404738,
      "grad_norm": 8.875,
      "learning_rate": 2.2814398892392894e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2714048.0,
      "step": 6228
    },
    {
      "epoch": 0.47911699100069227,
      "grad_norm": 4.84375,
      "learning_rate": 2.2813245134989617e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2714364.0,
      "step": 6229
    },
    {
      "epoch": 0.4791939081609107,
      "grad_norm": 4.5,
      "learning_rate": 2.281209137758634e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2714890.0,
      "step": 6230
    },
    {
      "epoch": 0.47927082532112913,
      "grad_norm": 3.09375,
      "learning_rate": 2.2810937620183066e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 2715414.0,
      "step": 6231
    },
    {
      "epoch": 0.4793477424813476,
      "grad_norm": 9.75,
      "learning_rate": 2.280978386277979e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2715724.0,
      "step": 6232
    },
    {
      "epoch": 0.47942465964156605,
      "grad_norm": 7.15625,
      "learning_rate": 2.2808630105376507e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 2716014.0,
      "step": 6233
    },
    {
      "epoch": 0.47950157680178446,
      "grad_norm": 4.59375,
      "learning_rate": 2.2807476347973234e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2716578.0,
      "step": 6234
    },
    {
      "epoch": 0.4795784939620029,
      "grad_norm": 5.09375,
      "learning_rate": 2.2806322590569956e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 2716928.0,
      "step": 6235
    },
    {
      "epoch": 0.4796554111222214,
      "grad_norm": 5.5,
      "learning_rate": 2.280516883316668e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2717248.0,
      "step": 6236
    },
    {
      "epoch": 0.47973232828243983,
      "grad_norm": 2.921875,
      "learning_rate": 2.2804015075763406e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 2717808.0,
      "step": 6237
    },
    {
      "epoch": 0.47980924544265824,
      "grad_norm": 6.34375,
      "learning_rate": 2.2802861318360124e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2718372.0,
      "step": 6238
    },
    {
      "epoch": 0.4798861626028767,
      "grad_norm": 19.125,
      "learning_rate": 2.280170756095685e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2718722.0,
      "step": 6239
    },
    {
      "epoch": 0.47996307976309516,
      "grad_norm": 7.6875,
      "learning_rate": 2.2800553803553573e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2719038.0,
      "step": 6240
    },
    {
      "epoch": 0.4800399969233136,
      "grad_norm": 5.28125,
      "learning_rate": 2.2799400046150296e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9970845580101013,
      "num_tokens": 2719382.0,
      "step": 6241
    },
    {
      "epoch": 0.480116914083532,
      "grad_norm": 4.90625,
      "learning_rate": 2.2798246288747023e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 2719702.0,
      "step": 6242
    },
    {
      "epoch": 0.4801938312437505,
      "grad_norm": 2.0625,
      "learning_rate": 2.279709253134374e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 2720262.0,
      "step": 6243
    },
    {
      "epoch": 0.48027074840396894,
      "grad_norm": 7.90625,
      "learning_rate": 2.2795938773940464e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2720802.0,
      "step": 6244
    },
    {
      "epoch": 0.48034766556418734,
      "grad_norm": 2.90625,
      "learning_rate": 2.279478501653719e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 2721386.0,
      "step": 6245
    },
    {
      "epoch": 0.4804245827244058,
      "grad_norm": 15.8125,
      "learning_rate": 2.2793631259133913e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 2721718.0,
      "step": 6246
    },
    {
      "epoch": 0.48050149988462426,
      "grad_norm": 3.375,
      "learning_rate": 2.2792477501730636e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9968847632408142,
      "num_tokens": 2722040.0,
      "step": 6247
    },
    {
      "epoch": 0.4805784170448427,
      "grad_norm": 2.328125,
      "learning_rate": 2.2791323744327363e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 2723062.0,
      "step": 6248
    },
    {
      "epoch": 0.4806553342050611,
      "grad_norm": 9.25,
      "learning_rate": 2.279016998692408e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2723362.0,
      "step": 6249
    },
    {
      "epoch": 0.4807322513652796,
      "grad_norm": 7.0625,
      "learning_rate": 2.2789016229520808e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2723674.0,
      "step": 6250
    },
    {
      "epoch": 0.48080916852549804,
      "grad_norm": 8.875,
      "learning_rate": 2.278786247211753e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2724254.0,
      "step": 6251
    },
    {
      "epoch": 0.4808860856857165,
      "grad_norm": 4.84375,
      "learning_rate": 2.2786708714714253e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 2725070.0,
      "step": 6252
    },
    {
      "epoch": 0.4809630028459349,
      "grad_norm": 7.59375,
      "learning_rate": 2.278555495731098e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 2725384.0,
      "step": 6253
    },
    {
      "epoch": 0.48103992000615337,
      "grad_norm": 11.0,
      "learning_rate": 2.27844011999077e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 2725986.0,
      "step": 6254
    },
    {
      "epoch": 0.4811168371663718,
      "grad_norm": 4.875,
      "learning_rate": 2.278324744250442e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2726548.0,
      "step": 6255
    },
    {
      "epoch": 0.4811937543265903,
      "grad_norm": 22.5,
      "learning_rate": 2.2782093685101147e-06,
      "loss": 0.0371,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2726802.0,
      "step": 6256
    },
    {
      "epoch": 0.4812706714868087,
      "grad_norm": 5.53125,
      "learning_rate": 2.278093992769787e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2727320.0,
      "step": 6257
    },
    {
      "epoch": 0.48134758864702715,
      "grad_norm": 14.8125,
      "learning_rate": 2.2779786170294593e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2727650.0,
      "step": 6258
    },
    {
      "epoch": 0.4814245058072456,
      "grad_norm": 11.5,
      "learning_rate": 2.277863241289132e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2727960.0,
      "step": 6259
    },
    {
      "epoch": 0.48150142296746407,
      "grad_norm": 4.40625,
      "learning_rate": 2.2777478655488038e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2728282.0,
      "step": 6260
    },
    {
      "epoch": 0.48157834012768247,
      "grad_norm": 17.625,
      "learning_rate": 2.2776324898084765e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 2728562.0,
      "step": 6261
    },
    {
      "epoch": 0.48165525728790093,
      "grad_norm": 3.96875,
      "learning_rate": 2.2775171140681487e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2729104.0,
      "step": 6262
    },
    {
      "epoch": 0.4817321744481194,
      "grad_norm": 6.9375,
      "learning_rate": 2.277401738327821e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 2729444.0,
      "step": 6263
    },
    {
      "epoch": 0.48180909160833785,
      "grad_norm": 7.03125,
      "learning_rate": 2.2772863625874937e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2729780.0,
      "step": 6264
    },
    {
      "epoch": 0.48188600876855625,
      "grad_norm": 14.1875,
      "learning_rate": 2.2771709868471655e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 2730098.0,
      "step": 6265
    },
    {
      "epoch": 0.4819629259287747,
      "grad_norm": 2.390625,
      "learning_rate": 2.2770556111068377e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2730638.0,
      "step": 6266
    },
    {
      "epoch": 0.48203984308899317,
      "grad_norm": 3.9375,
      "learning_rate": 2.2769402353665104e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2730946.0,
      "step": 6267
    },
    {
      "epoch": 0.4821167602492116,
      "grad_norm": 4.40625,
      "learning_rate": 2.2768248596261827e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 2731466.0,
      "step": 6268
    },
    {
      "epoch": 0.48219367740943003,
      "grad_norm": 10.125,
      "learning_rate": 2.276709483885855e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2732022.0,
      "step": 6269
    },
    {
      "epoch": 0.4822705945696485,
      "grad_norm": 11.0625,
      "learning_rate": 2.2765941081455276e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 2732330.0,
      "step": 6270
    },
    {
      "epoch": 0.48234751172986695,
      "grad_norm": 8.9375,
      "learning_rate": 2.2764787324051995e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2732648.0,
      "step": 6271
    },
    {
      "epoch": 0.48242442889008536,
      "grad_norm": 13.125,
      "learning_rate": 2.276363356664872e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2733184.0,
      "step": 6272
    },
    {
      "epoch": 0.4825013460503038,
      "grad_norm": 12.0,
      "learning_rate": 2.2762479809245444e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2733770.0,
      "step": 6273
    },
    {
      "epoch": 0.4825782632105223,
      "grad_norm": 4.34375,
      "learning_rate": 2.2761326051842167e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2734314.0,
      "step": 6274
    },
    {
      "epoch": 0.48265518037074073,
      "grad_norm": 14.625,
      "learning_rate": 2.276017229443889e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2734592.0,
      "step": 6275
    },
    {
      "epoch": 0.48273209753095914,
      "grad_norm": 3.25,
      "learning_rate": 2.275901853703561e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 2735134.0,
      "step": 6276
    },
    {
      "epoch": 0.4828090146911776,
      "grad_norm": 4.25,
      "learning_rate": 2.2757864779632334e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2735442.0,
      "step": 6277
    },
    {
      "epoch": 0.48288593185139606,
      "grad_norm": 4.5625,
      "learning_rate": 2.275671102222906e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2735758.0,
      "step": 6278
    },
    {
      "epoch": 0.4829628490116145,
      "grad_norm": 10.5625,
      "learning_rate": 2.2755557264825784e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2736344.0,
      "step": 6279
    },
    {
      "epoch": 0.4830397661718329,
      "grad_norm": 10.625,
      "learning_rate": 2.2754403507422506e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2736636.0,
      "step": 6280
    },
    {
      "epoch": 0.4831166833320514,
      "grad_norm": 28.625,
      "learning_rate": 2.275324975001923e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9862637519836426,
      "num_tokens": 2737002.0,
      "step": 6281
    },
    {
      "epoch": 0.48319360049226984,
      "grad_norm": 15.8125,
      "learning_rate": 2.275209599261595e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 2737304.0,
      "step": 6282
    },
    {
      "epoch": 0.4832705176524883,
      "grad_norm": 21.0,
      "learning_rate": 2.2750942235212674e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 2737638.0,
      "step": 6283
    },
    {
      "epoch": 0.4833474348127067,
      "grad_norm": 10.3125,
      "learning_rate": 2.27497884778094e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2737982.0,
      "step": 6284
    },
    {
      "epoch": 0.48342435197292516,
      "grad_norm": 5.125,
      "learning_rate": 2.2748634720406123e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 2738538.0,
      "step": 6285
    },
    {
      "epoch": 0.4835012691331436,
      "grad_norm": 4.375,
      "learning_rate": 2.2747480963002846e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2738832.0,
      "step": 6286
    },
    {
      "epoch": 0.483578186293362,
      "grad_norm": 8.75,
      "learning_rate": 2.274632720559957e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2739182.0,
      "step": 6287
    },
    {
      "epoch": 0.4836551034535805,
      "grad_norm": 3.921875,
      "learning_rate": 2.274517344819629e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2739758.0,
      "step": 6288
    },
    {
      "epoch": 0.48373202061379894,
      "grad_norm": 5.65625,
      "learning_rate": 2.274401969079302e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2740108.0,
      "step": 6289
    },
    {
      "epoch": 0.4838089377740174,
      "grad_norm": 12.4375,
      "learning_rate": 2.274286593338974e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2740402.0,
      "step": 6290
    },
    {
      "epoch": 0.4838858549342358,
      "grad_norm": 12.375,
      "learning_rate": 2.2741712175986463e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2740920.0,
      "step": 6291
    },
    {
      "epoch": 0.48396277209445426,
      "grad_norm": 29.375,
      "learning_rate": 2.2740558418583186e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 2741244.0,
      "step": 6292
    },
    {
      "epoch": 0.4840396892546727,
      "grad_norm": 13.875,
      "learning_rate": 2.273940466117991e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 2741764.0,
      "step": 6293
    },
    {
      "epoch": 0.4841166064148912,
      "grad_norm": 20.5,
      "learning_rate": 2.273825090377663e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 2742056.0,
      "step": 6294
    },
    {
      "epoch": 0.4841935235751096,
      "grad_norm": 10.4375,
      "learning_rate": 2.2737097146373358e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 2742656.0,
      "step": 6295
    },
    {
      "epoch": 0.48427044073532804,
      "grad_norm": 9.1875,
      "learning_rate": 2.273594338897008e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 2742986.0,
      "step": 6296
    },
    {
      "epoch": 0.4843473578955465,
      "grad_norm": 3.5,
      "learning_rate": 2.2734789631566803e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2743526.0,
      "step": 6297
    },
    {
      "epoch": 0.48442427505576496,
      "grad_norm": 3.59375,
      "learning_rate": 2.2733635874163525e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 2744306.0,
      "step": 6298
    },
    {
      "epoch": 0.48450119221598337,
      "grad_norm": 15.8125,
      "learning_rate": 2.273248211676025e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2744632.0,
      "step": 6299
    },
    {
      "epoch": 0.4845781093762018,
      "grad_norm": 25.625,
      "learning_rate": 2.2731328359356975e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2744988.0,
      "step": 6300
    },
    {
      "epoch": 0.4846550265364203,
      "grad_norm": 8.3125,
      "learning_rate": 2.2730174601953697e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 2745544.0,
      "step": 6301
    },
    {
      "epoch": 0.48473194369663875,
      "grad_norm": 11.625,
      "learning_rate": 2.272902084455042e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2746100.0,
      "step": 6302
    },
    {
      "epoch": 0.48480886085685715,
      "grad_norm": 3.875,
      "learning_rate": 2.2727867087147143e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 2746650.0,
      "step": 6303
    },
    {
      "epoch": 0.4848857780170756,
      "grad_norm": 9.125,
      "learning_rate": 2.2726713329743865e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2746944.0,
      "step": 6304
    },
    {
      "epoch": 0.48496269517729407,
      "grad_norm": 4.9375,
      "learning_rate": 2.2725559572340588e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 2747264.0,
      "step": 6305
    },
    {
      "epoch": 0.48503961233751247,
      "grad_norm": 8.3125,
      "learning_rate": 2.2724405814937315e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 1.0,
      "num_tokens": 2747604.0,
      "step": 6306
    },
    {
      "epoch": 0.48511652949773093,
      "grad_norm": 8.0,
      "learning_rate": 2.2723252057534037e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2747928.0,
      "step": 6307
    },
    {
      "epoch": 0.4851934466579494,
      "grad_norm": 15.0,
      "learning_rate": 2.272209830013076e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 2748240.0,
      "step": 6308
    },
    {
      "epoch": 0.48527036381816785,
      "grad_norm": 2.375,
      "learning_rate": 2.2720944542727482e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 2749026.0,
      "step": 6309
    },
    {
      "epoch": 0.48534728097838625,
      "grad_norm": 10.5,
      "learning_rate": 2.2719790785324205e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9889298677444458,
      "num_tokens": 2749298.0,
      "step": 6310
    },
    {
      "epoch": 0.4854241981386047,
      "grad_norm": 8.25,
      "learning_rate": 2.271863702792093e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2749592.0,
      "step": 6311
    },
    {
      "epoch": 0.48550111529882317,
      "grad_norm": 2.296875,
      "learning_rate": 2.2717483270517654e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 2750176.0,
      "step": 6312
    },
    {
      "epoch": 0.48557803245904163,
      "grad_norm": 23.375,
      "learning_rate": 2.2716329513114377e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 2750478.0,
      "step": 6313
    },
    {
      "epoch": 0.48565494961926003,
      "grad_norm": 21.625,
      "learning_rate": 2.27151757557111e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 2751078.0,
      "step": 6314
    },
    {
      "epoch": 0.4857318667794785,
      "grad_norm": 5.0625,
      "learning_rate": 2.271402199830782e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 2751598.0,
      "step": 6315
    },
    {
      "epoch": 0.48580878393969695,
      "grad_norm": 13.5,
      "learning_rate": 2.2712868240904545e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 2751942.0,
      "step": 6316
    },
    {
      "epoch": 0.4858857010999154,
      "grad_norm": 3.265625,
      "learning_rate": 2.271171448350127e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2752460.0,
      "step": 6317
    },
    {
      "epoch": 0.4859626182601338,
      "grad_norm": 22.875,
      "learning_rate": 2.2710560726097994e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 2752756.0,
      "step": 6318
    },
    {
      "epoch": 0.4860395354203523,
      "grad_norm": 19.375,
      "learning_rate": 2.2709406968694717e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 2753108.0,
      "step": 6319
    },
    {
      "epoch": 0.48611645258057073,
      "grad_norm": 6.40625,
      "learning_rate": 2.270825321129144e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 2753664.0,
      "step": 6320
    },
    {
      "epoch": 0.4861933697407892,
      "grad_norm": 5.28125,
      "learning_rate": 2.270709945388816e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2754190.0,
      "step": 6321
    },
    {
      "epoch": 0.4862702869010076,
      "grad_norm": 9.5,
      "learning_rate": 2.270594569648489e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2754474.0,
      "step": 6322
    },
    {
      "epoch": 0.48634720406122606,
      "grad_norm": 9.125,
      "learning_rate": 2.270479193908161e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2754824.0,
      "step": 6323
    },
    {
      "epoch": 0.4864241212214445,
      "grad_norm": 5.0625,
      "learning_rate": 2.2703638181678334e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 2755420.0,
      "step": 6324
    },
    {
      "epoch": 0.486501038381663,
      "grad_norm": 33.5,
      "learning_rate": 2.2702484424275056e-06,
      "loss": 0.0422,
      "mean_token_accuracy": 0.9825327396392822,
      "num_tokens": 2755650.0,
      "step": 6325
    },
    {
      "epoch": 0.4865779555418814,
      "grad_norm": 13.1875,
      "learning_rate": 2.270133066687178e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2755968.0,
      "step": 6326
    },
    {
      "epoch": 0.48665487270209984,
      "grad_norm": 2.21875,
      "learning_rate": 2.27001769094685e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2756788.0,
      "step": 6327
    },
    {
      "epoch": 0.4867317898623183,
      "grad_norm": 22.0,
      "learning_rate": 2.269902315206523e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9773463010787964,
      "num_tokens": 2757100.0,
      "step": 6328
    },
    {
      "epoch": 0.4868087070225367,
      "grad_norm": 5.5,
      "learning_rate": 2.269786939466195e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9967532753944397,
      "num_tokens": 2757410.0,
      "step": 6329
    },
    {
      "epoch": 0.48688562418275516,
      "grad_norm": 4.1875,
      "learning_rate": 2.2696715637258673e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 2757710.0,
      "step": 6330
    },
    {
      "epoch": 0.4869625413429736,
      "grad_norm": 5.46875,
      "learning_rate": 2.2695561879855396e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2758274.0,
      "step": 6331
    },
    {
      "epoch": 0.4870394585031921,
      "grad_norm": 23.75,
      "learning_rate": 2.269440812245212e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 2758586.0,
      "step": 6332
    },
    {
      "epoch": 0.4871163756634105,
      "grad_norm": 15.75,
      "learning_rate": 2.2693254365048845e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 2758874.0,
      "step": 6333
    },
    {
      "epoch": 0.48719329282362894,
      "grad_norm": 8.875,
      "learning_rate": 2.269210060764557e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2759196.0,
      "step": 6334
    },
    {
      "epoch": 0.4872702099838474,
      "grad_norm": 4.125,
      "learning_rate": 2.269094685024229e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2759520.0,
      "step": 6335
    },
    {
      "epoch": 0.48734712714406586,
      "grad_norm": 13.0,
      "learning_rate": 2.2689793092839013e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 2759846.0,
      "step": 6336
    },
    {
      "epoch": 0.48742404430428427,
      "grad_norm": 3.984375,
      "learning_rate": 2.2688639335435736e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2760398.0,
      "step": 6337
    },
    {
      "epoch": 0.4875009614645027,
      "grad_norm": 9.1875,
      "learning_rate": 2.268748557803246e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9888682961463928,
      "num_tokens": 2760938.0,
      "step": 6338
    },
    {
      "epoch": 0.4875778786247212,
      "grad_norm": 10.375,
      "learning_rate": 2.2686331820629185e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 2761288.0,
      "step": 6339
    },
    {
      "epoch": 0.48765479578493964,
      "grad_norm": 5.9375,
      "learning_rate": 2.2685178063225908e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 2761804.0,
      "step": 6340
    },
    {
      "epoch": 0.48773171294515805,
      "grad_norm": 6.125,
      "learning_rate": 2.268402430582263e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 2762118.0,
      "step": 6341
    },
    {
      "epoch": 0.4878086301053765,
      "grad_norm": 8.8125,
      "learning_rate": 2.2682870548419353e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 2762942.0,
      "step": 6342
    },
    {
      "epoch": 0.48788554726559497,
      "grad_norm": 8.75,
      "learning_rate": 2.2681716791016075e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 2763260.0,
      "step": 6343
    },
    {
      "epoch": 0.4879624644258134,
      "grad_norm": 10.6875,
      "learning_rate": 2.26805630336128e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 2763752.0,
      "step": 6344
    },
    {
      "epoch": 0.48803938158603183,
      "grad_norm": 14.625,
      "learning_rate": 2.2679409276209525e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2764332.0,
      "step": 6345
    },
    {
      "epoch": 0.4881162987462503,
      "grad_norm": 10.625,
      "learning_rate": 2.2678255518806243e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2764894.0,
      "step": 6346
    },
    {
      "epoch": 0.48819321590646875,
      "grad_norm": 10.375,
      "learning_rate": 2.267710176140297e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 2765188.0,
      "step": 6347
    },
    {
      "epoch": 0.48827013306668715,
      "grad_norm": 15.125,
      "learning_rate": 2.2675948003999693e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2765732.0,
      "step": 6348
    },
    {
      "epoch": 0.4883470502269056,
      "grad_norm": 30.625,
      "learning_rate": 2.2674794246596415e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2766018.0,
      "step": 6349
    },
    {
      "epoch": 0.48842396738712407,
      "grad_norm": 22.0,
      "learning_rate": 2.267364048919314e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2766378.0,
      "step": 6350
    },
    {
      "epoch": 0.48850088454734253,
      "grad_norm": 9.5,
      "learning_rate": 2.2672486731789865e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2766958.0,
      "step": 6351
    },
    {
      "epoch": 0.48857780170756093,
      "grad_norm": 10.3125,
      "learning_rate": 2.2671332974386583e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 2767540.0,
      "step": 6352
    },
    {
      "epoch": 0.4886547188677794,
      "grad_norm": 13.1875,
      "learning_rate": 2.267017921698331e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 2767872.0,
      "step": 6353
    },
    {
      "epoch": 0.48873163602799785,
      "grad_norm": 27.25,
      "learning_rate": 2.2669025459580032e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 2768156.0,
      "step": 6354
    },
    {
      "epoch": 0.4888085531882163,
      "grad_norm": 22.5,
      "learning_rate": 2.2667871702176755e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 2768500.0,
      "step": 6355
    },
    {
      "epoch": 0.4888854703484347,
      "grad_norm": 18.375,
      "learning_rate": 2.266671794477348e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 2769068.0,
      "step": 6356
    },
    {
      "epoch": 0.4889623875086532,
      "grad_norm": 9.75,
      "learning_rate": 2.26655641873702e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 2769360.0,
      "step": 6357
    },
    {
      "epoch": 0.48903930466887163,
      "grad_norm": 14.625,
      "learning_rate": 2.2664410429966927e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 2769932.0,
      "step": 6358
    },
    {
      "epoch": 0.4891162218290901,
      "grad_norm": 5.0,
      "learning_rate": 2.266325667256365e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 2770746.0,
      "step": 6359
    },
    {
      "epoch": 0.4891931389893085,
      "grad_norm": 17.875,
      "learning_rate": 2.266210291516037e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 2771240.0,
      "step": 6360
    },
    {
      "epoch": 0.48927005614952696,
      "grad_norm": 22.875,
      "learning_rate": 2.26609491577571e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 2771546.0,
      "step": 6361
    },
    {
      "epoch": 0.4893469733097454,
      "grad_norm": 20.625,
      "learning_rate": 2.265979540035382e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2771864.0,
      "step": 6362
    },
    {
      "epoch": 0.4894238904699639,
      "grad_norm": 4.25,
      "learning_rate": 2.265864164295054e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2772406.0,
      "step": 6363
    },
    {
      "epoch": 0.4895008076301823,
      "grad_norm": 7.53125,
      "learning_rate": 2.2657487885547267e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2772690.0,
      "step": 6364
    },
    {
      "epoch": 0.48957772479040074,
      "grad_norm": 23.375,
      "learning_rate": 2.265633412814399e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9850187301635742,
      "num_tokens": 2772958.0,
      "step": 6365
    },
    {
      "epoch": 0.4896546419506192,
      "grad_norm": 7.15625,
      "learning_rate": 2.265518037074071e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2773512.0,
      "step": 6366
    },
    {
      "epoch": 0.48973155911083766,
      "grad_norm": 12.875,
      "learning_rate": 2.265402661333744e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2774332.0,
      "step": 6367
    },
    {
      "epoch": 0.48980847627105606,
      "grad_norm": 4.78125,
      "learning_rate": 2.2652872855934157e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2774906.0,
      "step": 6368
    },
    {
      "epoch": 0.4898853934312745,
      "grad_norm": 8.1875,
      "learning_rate": 2.2651719098530884e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 2775482.0,
      "step": 6369
    },
    {
      "epoch": 0.489962310591493,
      "grad_norm": 9.875,
      "learning_rate": 2.2650565341127606e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2775818.0,
      "step": 6370
    },
    {
      "epoch": 0.4900392277517114,
      "grad_norm": 14.9375,
      "learning_rate": 2.264941158372433e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2776134.0,
      "step": 6371
    },
    {
      "epoch": 0.49011614491192984,
      "grad_norm": 15.9375,
      "learning_rate": 2.2648257826321056e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 2776468.0,
      "step": 6372
    },
    {
      "epoch": 0.4901930620721483,
      "grad_norm": 21.25,
      "learning_rate": 2.264710406891778e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 2776746.0,
      "step": 6373
    },
    {
      "epoch": 0.49026997923236676,
      "grad_norm": 5.375,
      "learning_rate": 2.2645950311514497e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2777310.0,
      "step": 6374
    },
    {
      "epoch": 0.49034689639258516,
      "grad_norm": 7.78125,
      "learning_rate": 2.2644796554111223e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2777868.0,
      "step": 6375
    },
    {
      "epoch": 0.4904238135528036,
      "grad_norm": 10.375,
      "learning_rate": 2.2643642796707946e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2778414.0,
      "step": 6376
    },
    {
      "epoch": 0.4905007307130221,
      "grad_norm": 20.125,
      "learning_rate": 2.264248903930467e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2778726.0,
      "step": 6377
    },
    {
      "epoch": 0.49057764787324054,
      "grad_norm": 2.65625,
      "learning_rate": 2.2641335281901395e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2779274.0,
      "step": 6378
    },
    {
      "epoch": 0.49065456503345894,
      "grad_norm": 11.75,
      "learning_rate": 2.2640181524498114e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2779572.0,
      "step": 6379
    },
    {
      "epoch": 0.4907314821936774,
      "grad_norm": 25.75,
      "learning_rate": 2.263902776709484e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2779850.0,
      "step": 6380
    },
    {
      "epoch": 0.49080839935389586,
      "grad_norm": 9.1875,
      "learning_rate": 2.2637874009691563e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 2780176.0,
      "step": 6381
    },
    {
      "epoch": 0.4908853165141143,
      "grad_norm": 15.0625,
      "learning_rate": 2.2636720252288286e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9931740760803223,
      "num_tokens": 2780470.0,
      "step": 6382
    },
    {
      "epoch": 0.4909622336743327,
      "grad_norm": 3.359375,
      "learning_rate": 2.2635566494885013e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 2781044.0,
      "step": 6383
    },
    {
      "epoch": 0.4910391508345512,
      "grad_norm": 3.859375,
      "learning_rate": 2.263441273748173e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2781592.0,
      "step": 6384
    },
    {
      "epoch": 0.49111606799476965,
      "grad_norm": 6.84375,
      "learning_rate": 2.2633258980078454e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9961832165718079,
      "num_tokens": 2782380.0,
      "step": 6385
    },
    {
      "epoch": 0.4911929851549881,
      "grad_norm": 13.625,
      "learning_rate": 2.263210522267518e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2782938.0,
      "step": 6386
    },
    {
      "epoch": 0.4912699023152065,
      "grad_norm": 12.125,
      "learning_rate": 2.2630951465271903e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 2783276.0,
      "step": 6387
    },
    {
      "epoch": 0.49134681947542497,
      "grad_norm": 8.4375,
      "learning_rate": 2.2629797707868626e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2783558.0,
      "step": 6388
    },
    {
      "epoch": 0.4914237366356434,
      "grad_norm": 12.75,
      "learning_rate": 2.2628643950465352e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2784122.0,
      "step": 6389
    },
    {
      "epoch": 0.49150065379586183,
      "grad_norm": 5.5625,
      "learning_rate": 2.262749019306207e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2784444.0,
      "step": 6390
    },
    {
      "epoch": 0.4915775709560803,
      "grad_norm": 6.25,
      "learning_rate": 2.2626336435658797e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2784758.0,
      "step": 6391
    },
    {
      "epoch": 0.49165448811629875,
      "grad_norm": 10.5625,
      "learning_rate": 2.262518267825552e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2785316.0,
      "step": 6392
    },
    {
      "epoch": 0.4917314052765172,
      "grad_norm": 11.875,
      "learning_rate": 2.2624028920852243e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 2785650.0,
      "step": 6393
    },
    {
      "epoch": 0.4918083224367356,
      "grad_norm": 5.15625,
      "learning_rate": 2.2622875163448965e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2786196.0,
      "step": 6394
    },
    {
      "epoch": 0.49188523959695407,
      "grad_norm": 6.375,
      "learning_rate": 2.2621721406045688e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 2786518.0,
      "step": 6395
    },
    {
      "epoch": 0.49196215675717253,
      "grad_norm": 4.6875,
      "learning_rate": 2.262056764864241e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 2787058.0,
      "step": 6396
    },
    {
      "epoch": 0.492039073917391,
      "grad_norm": 6.5,
      "learning_rate": 2.2619413891239137e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2787612.0,
      "step": 6397
    },
    {
      "epoch": 0.4921159910776094,
      "grad_norm": 9.3125,
      "learning_rate": 2.261826013383586e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 2788436.0,
      "step": 6398
    },
    {
      "epoch": 0.49219290823782785,
      "grad_norm": 4.28125,
      "learning_rate": 2.2617106376432582e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2788978.0,
      "step": 6399
    },
    {
      "epoch": 0.4922698253980463,
      "grad_norm": 8.8125,
      "learning_rate": 2.261595261902931e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2789296.0,
      "step": 6400
    },
    {
      "epoch": 0.49234674255826477,
      "grad_norm": 14.625,
      "learning_rate": 2.2614798861626028e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9926199316978455,
      "num_tokens": 2789568.0,
      "step": 6401
    },
    {
      "epoch": 0.4924236597184832,
      "grad_norm": 26.875,
      "learning_rate": 2.261364510422275e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2789874.0,
      "step": 6402
    },
    {
      "epoch": 0.49250057687870163,
      "grad_norm": 19.5,
      "learning_rate": 2.2612491346819477e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 2790184.0,
      "step": 6403
    },
    {
      "epoch": 0.4925774940389201,
      "grad_norm": 25.125,
      "learning_rate": 2.26113375894162e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2790466.0,
      "step": 6404
    },
    {
      "epoch": 0.49265441119913855,
      "grad_norm": 23.25,
      "learning_rate": 2.261018383201292e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 2790780.0,
      "step": 6405
    },
    {
      "epoch": 0.49273132835935696,
      "grad_norm": 19.25,
      "learning_rate": 2.2609030074609645e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 2791090.0,
      "step": 6406
    },
    {
      "epoch": 0.4928082455195754,
      "grad_norm": 19.25,
      "learning_rate": 2.2607876317206367e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2791646.0,
      "step": 6407
    },
    {
      "epoch": 0.4928851626797939,
      "grad_norm": 3.390625,
      "learning_rate": 2.2606722559803094e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2792220.0,
      "step": 6408
    },
    {
      "epoch": 0.4929620798400123,
      "grad_norm": 10.5,
      "learning_rate": 2.2605568802399817e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2792506.0,
      "step": 6409
    },
    {
      "epoch": 0.49303899700023074,
      "grad_norm": 6.40625,
      "learning_rate": 2.260441504499654e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2793046.0,
      "step": 6410
    },
    {
      "epoch": 0.4931159141604492,
      "grad_norm": 3.25,
      "learning_rate": 2.2603261287593266e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2793592.0,
      "step": 6411
    },
    {
      "epoch": 0.49319283132066766,
      "grad_norm": 17.375,
      "learning_rate": 2.2602107530189984e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2794418.0,
      "step": 6412
    },
    {
      "epoch": 0.49326974848088606,
      "grad_norm": 17.25,
      "learning_rate": 2.2600953772786707e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2795004.0,
      "step": 6413
    },
    {
      "epoch": 0.4933466656411045,
      "grad_norm": 4.15625,
      "learning_rate": 2.2599800015383434e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2795552.0,
      "step": 6414
    },
    {
      "epoch": 0.493423582801323,
      "grad_norm": 6.46875,
      "learning_rate": 2.2598646257980156e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2795866.0,
      "step": 6415
    },
    {
      "epoch": 0.49350049996154144,
      "grad_norm": 27.25,
      "learning_rate": 2.259749250057688e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 2796150.0,
      "step": 6416
    },
    {
      "epoch": 0.49357741712175984,
      "grad_norm": 12.25,
      "learning_rate": 2.25963387431736e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2796506.0,
      "step": 6417
    },
    {
      "epoch": 0.4936543342819783,
      "grad_norm": 10.625,
      "learning_rate": 2.2595184985770324e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2796824.0,
      "step": 6418
    },
    {
      "epoch": 0.49373125144219676,
      "grad_norm": 4.3125,
      "learning_rate": 2.259403122836705e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 2797384.0,
      "step": 6419
    },
    {
      "epoch": 0.4938081686024152,
      "grad_norm": 6.96875,
      "learning_rate": 2.2592877470963774e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2797938.0,
      "step": 6420
    },
    {
      "epoch": 0.4938850857626336,
      "grad_norm": 18.625,
      "learning_rate": 2.2591723713560496e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2798502.0,
      "step": 6421
    },
    {
      "epoch": 0.4939620029228521,
      "grad_norm": 12.3125,
      "learning_rate": 2.259056995615722e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9904030561447144,
      "num_tokens": 2799024.0,
      "step": 6422
    },
    {
      "epoch": 0.49403892008307054,
      "grad_norm": 7.0625,
      "learning_rate": 2.258941619875394e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 2799310.0,
      "step": 6423
    },
    {
      "epoch": 0.494115837243289,
      "grad_norm": 2.640625,
      "learning_rate": 2.2588262441350664e-06,
      "loss": 0.0061,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 2800354.0,
      "step": 6424
    },
    {
      "epoch": 0.4941927544035074,
      "grad_norm": 22.25,
      "learning_rate": 2.258710868394739e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 2800626.0,
      "step": 6425
    },
    {
      "epoch": 0.49426967156372587,
      "grad_norm": 11.0625,
      "learning_rate": 2.2585954926544113e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 2801212.0,
      "step": 6426
    },
    {
      "epoch": 0.4943465887239443,
      "grad_norm": 13.0,
      "learning_rate": 2.2584801169140836e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 2801774.0,
      "step": 6427
    },
    {
      "epoch": 0.4944235058841628,
      "grad_norm": 5.03125,
      "learning_rate": 2.258364741173756e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 2802354.0,
      "step": 6428
    },
    {
      "epoch": 0.4945004230443812,
      "grad_norm": 2.453125,
      "learning_rate": 2.258249365433428e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.998123824596405,
      "num_tokens": 2802888.0,
      "step": 6429
    },
    {
      "epoch": 0.49457734020459965,
      "grad_norm": 16.125,
      "learning_rate": 2.2581339896931008e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2803204.0,
      "step": 6430
    },
    {
      "epoch": 0.4946542573648181,
      "grad_norm": 7.375,
      "learning_rate": 2.258018613952773e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2803760.0,
      "step": 6431
    },
    {
      "epoch": 0.4947311745250365,
      "grad_norm": 7.1875,
      "learning_rate": 2.2579032382124453e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2804094.0,
      "step": 6432
    },
    {
      "epoch": 0.49480809168525497,
      "grad_norm": 3.25,
      "learning_rate": 2.2577878624721176e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 2804650.0,
      "step": 6433
    },
    {
      "epoch": 0.49488500884547343,
      "grad_norm": 6.21875,
      "learning_rate": 2.25767248673179e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 2804984.0,
      "step": 6434
    },
    {
      "epoch": 0.4949619260056919,
      "grad_norm": 5.84375,
      "learning_rate": 2.257557110991462e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2805764.0,
      "step": 6435
    },
    {
      "epoch": 0.4950388431659103,
      "grad_norm": 48.25,
      "learning_rate": 2.2574417352511348e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2806068.0,
      "step": 6436
    },
    {
      "epoch": 0.49511576032612875,
      "grad_norm": 21.75,
      "learning_rate": 2.257326359510807e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 2806328.0,
      "step": 6437
    },
    {
      "epoch": 0.4951926774863472,
      "grad_norm": 4.5,
      "learning_rate": 2.2572109837704793e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2806874.0,
      "step": 6438
    },
    {
      "epoch": 0.49526959464656567,
      "grad_norm": 11.25,
      "learning_rate": 2.2570956080301515e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9928057789802551,
      "num_tokens": 2807154.0,
      "step": 6439
    },
    {
      "epoch": 0.4953465118067841,
      "grad_norm": 6.90625,
      "learning_rate": 2.2569802322898238e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 2807696.0,
      "step": 6440
    },
    {
      "epoch": 0.49542342896700253,
      "grad_norm": 27.5,
      "learning_rate": 2.2568648565494965e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2808046.0,
      "step": 6441
    },
    {
      "epoch": 0.495500346127221,
      "grad_norm": 4.03125,
      "learning_rate": 2.2567494808091687e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2808616.0,
      "step": 6442
    },
    {
      "epoch": 0.49557726328743945,
      "grad_norm": 28.625,
      "learning_rate": 2.256634105068841e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2808944.0,
      "step": 6443
    },
    {
      "epoch": 0.49565418044765785,
      "grad_norm": 19.625,
      "learning_rate": 2.2565187293285132e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 2809478.0,
      "step": 6444
    },
    {
      "epoch": 0.4957310976078763,
      "grad_norm": 5.90625,
      "learning_rate": 2.2564033535881855e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2810026.0,
      "step": 6445
    },
    {
      "epoch": 0.4958080147680948,
      "grad_norm": 31.0,
      "learning_rate": 2.2562879778478578e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2810324.0,
      "step": 6446
    },
    {
      "epoch": 0.49588493192831323,
      "grad_norm": 5.46875,
      "learning_rate": 2.2561726021075304e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 2810642.0,
      "step": 6447
    },
    {
      "epoch": 0.49596184908853164,
      "grad_norm": 40.25,
      "learning_rate": 2.2560572263672027e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2810948.0,
      "step": 6448
    },
    {
      "epoch": 0.4960387662487501,
      "grad_norm": 15.6875,
      "learning_rate": 2.255941850626875e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2811270.0,
      "step": 6449
    },
    {
      "epoch": 0.49611568340896856,
      "grad_norm": 23.375,
      "learning_rate": 2.255826474886547e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2811586.0,
      "step": 6450
    },
    {
      "epoch": 0.49619260056918696,
      "grad_norm": 19.5,
      "learning_rate": 2.2557110991462195e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 2812164.0,
      "step": 6451
    },
    {
      "epoch": 0.4962695177294054,
      "grad_norm": 21.25,
      "learning_rate": 2.2555957234058917e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 2812520.0,
      "step": 6452
    },
    {
      "epoch": 0.4963464348896239,
      "grad_norm": 14.8125,
      "learning_rate": 2.2554803476655644e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2813094.0,
      "step": 6453
    },
    {
      "epoch": 0.49642335204984234,
      "grad_norm": 28.625,
      "learning_rate": 2.2553649719252367e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2813428.0,
      "step": 6454
    },
    {
      "epoch": 0.49650026921006074,
      "grad_norm": 8.125,
      "learning_rate": 2.255249596184909e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2814018.0,
      "step": 6455
    },
    {
      "epoch": 0.4965771863702792,
      "grad_norm": 8.375,
      "learning_rate": 2.255134220444581e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2814580.0,
      "step": 6456
    },
    {
      "epoch": 0.49665410353049766,
      "grad_norm": 3.234375,
      "learning_rate": 2.2550188447042534e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2815126.0,
      "step": 6457
    },
    {
      "epoch": 0.4967310206907161,
      "grad_norm": 14.5625,
      "learning_rate": 2.254903468963926e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2815688.0,
      "step": 6458
    },
    {
      "epoch": 0.4968079378509345,
      "grad_norm": 52.0,
      "learning_rate": 2.2547880932235984e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2815946.0,
      "step": 6459
    },
    {
      "epoch": 0.496884855011153,
      "grad_norm": 45.5,
      "learning_rate": 2.2546727174832702e-06,
      "loss": 0.0361,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 2816222.0,
      "step": 6460
    },
    {
      "epoch": 0.49696177217137144,
      "grad_norm": 4.8125,
      "learning_rate": 2.254557341742943e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2817002.0,
      "step": 6461
    },
    {
      "epoch": 0.4970386893315899,
      "grad_norm": 14.5,
      "learning_rate": 2.254441966002615e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2817314.0,
      "step": 6462
    },
    {
      "epoch": 0.4971156064918083,
      "grad_norm": 21.625,
      "learning_rate": 2.2543265902622874e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2817642.0,
      "step": 6463
    },
    {
      "epoch": 0.49719252365202676,
      "grad_norm": 11.4375,
      "learning_rate": 2.25421121452196e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 2817898.0,
      "step": 6464
    },
    {
      "epoch": 0.4972694408122452,
      "grad_norm": 28.25,
      "learning_rate": 2.2540958387816324e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 2818230.0,
      "step": 6465
    },
    {
      "epoch": 0.4973463579724637,
      "grad_norm": 18.125,
      "learning_rate": 2.2539804630413046e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 2818534.0,
      "step": 6466
    },
    {
      "epoch": 0.4974232751326821,
      "grad_norm": 23.625,
      "learning_rate": 2.253865087300977e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2818858.0,
      "step": 6467
    },
    {
      "epoch": 0.49750019229290054,
      "grad_norm": 4.21875,
      "learning_rate": 2.253749711560649e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2819402.0,
      "step": 6468
    },
    {
      "epoch": 0.497577109453119,
      "grad_norm": 5.59375,
      "learning_rate": 2.253634335820322e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 2819926.0,
      "step": 6469
    },
    {
      "epoch": 0.4976540266133374,
      "grad_norm": 11.8125,
      "learning_rate": 2.253518960079994e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 2820262.0,
      "step": 6470
    },
    {
      "epoch": 0.49773094377355587,
      "grad_norm": 13.5625,
      "learning_rate": 2.253403584339666e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 2820564.0,
      "step": 6471
    },
    {
      "epoch": 0.4978078609337743,
      "grad_norm": 3.09375,
      "learning_rate": 2.2532882085993386e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2821110.0,
      "step": 6472
    },
    {
      "epoch": 0.4978847780939928,
      "grad_norm": 4.625,
      "learning_rate": 2.253172832859011e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 2821688.0,
      "step": 6473
    },
    {
      "epoch": 0.4979616952542112,
      "grad_norm": 15.9375,
      "learning_rate": 2.253057457118683e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2822514.0,
      "step": 6474
    },
    {
      "epoch": 0.49803861241442965,
      "grad_norm": 8.9375,
      "learning_rate": 2.2529420813783558e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2822772.0,
      "step": 6475
    },
    {
      "epoch": 0.4981155295746481,
      "grad_norm": 26.625,
      "learning_rate": 2.252826705638028e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 2823120.0,
      "step": 6476
    },
    {
      "epoch": 0.49819244673486657,
      "grad_norm": 6.9375,
      "learning_rate": 2.2527113298977003e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2823402.0,
      "step": 6477
    },
    {
      "epoch": 0.49826936389508497,
      "grad_norm": 10.8125,
      "learning_rate": 2.2525959541573726e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 2823682.0,
      "step": 6478
    },
    {
      "epoch": 0.49834628105530343,
      "grad_norm": 11.75,
      "learning_rate": 2.252480578417045e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2823960.0,
      "step": 6479
    },
    {
      "epoch": 0.4984231982155219,
      "grad_norm": 36.0,
      "learning_rate": 2.2523652026767175e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2824254.0,
      "step": 6480
    },
    {
      "epoch": 0.49850011537574035,
      "grad_norm": 5.5625,
      "learning_rate": 2.2522498269363898e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 2825068.0,
      "step": 6481
    },
    {
      "epoch": 0.49857703253595875,
      "grad_norm": 10.3125,
      "learning_rate": 2.2521344511960616e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2825372.0,
      "step": 6482
    },
    {
      "epoch": 0.4986539496961772,
      "grad_norm": 4.125,
      "learning_rate": 2.2520190754557343e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2825924.0,
      "step": 6483
    },
    {
      "epoch": 0.49873086685639567,
      "grad_norm": 31.625,
      "learning_rate": 2.2519036997154065e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2826246.0,
      "step": 6484
    },
    {
      "epoch": 0.49880778401661413,
      "grad_norm": 5.65625,
      "learning_rate": 2.2517883239750788e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2826530.0,
      "step": 6485
    },
    {
      "epoch": 0.49888470117683253,
      "grad_norm": 30.125,
      "learning_rate": 2.2516729482347515e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2826850.0,
      "step": 6486
    },
    {
      "epoch": 0.498961618337051,
      "grad_norm": 6.96875,
      "learning_rate": 2.2515575724944233e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2827414.0,
      "step": 6487
    },
    {
      "epoch": 0.49903853549726945,
      "grad_norm": 12.375,
      "learning_rate": 2.251442196754096e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2827930.0,
      "step": 6488
    },
    {
      "epoch": 0.4991154526574879,
      "grad_norm": 2.1875,
      "learning_rate": 2.2513268210137682e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2828494.0,
      "step": 6489
    },
    {
      "epoch": 0.4991923698177063,
      "grad_norm": 10.8125,
      "learning_rate": 2.2512114452734405e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 2828790.0,
      "step": 6490
    },
    {
      "epoch": 0.4992692869779248,
      "grad_norm": 4.53125,
      "learning_rate": 2.251096069533113e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2829380.0,
      "step": 6491
    },
    {
      "epoch": 0.49934620413814323,
      "grad_norm": 14.125,
      "learning_rate": 2.2509806937927854e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2829700.0,
      "step": 6492
    },
    {
      "epoch": 0.49942312129836164,
      "grad_norm": 3.859375,
      "learning_rate": 2.2508653180524573e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2830258.0,
      "step": 6493
    },
    {
      "epoch": 0.4995000384585801,
      "grad_norm": 15.8125,
      "learning_rate": 2.25074994231213e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 2830826.0,
      "step": 6494
    },
    {
      "epoch": 0.49957695561879856,
      "grad_norm": 23.75,
      "learning_rate": 2.2506345665718022e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 2831132.0,
      "step": 6495
    },
    {
      "epoch": 0.499653872779017,
      "grad_norm": 9.6875,
      "learning_rate": 2.2505191908314745e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2831460.0,
      "step": 6496
    },
    {
      "epoch": 0.4997307899392354,
      "grad_norm": 29.375,
      "learning_rate": 2.250403815091147e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 2831820.0,
      "step": 6497
    },
    {
      "epoch": 0.4998077070994539,
      "grad_norm": 9.6875,
      "learning_rate": 2.250288439350819e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 2832364.0,
      "step": 6498
    },
    {
      "epoch": 0.49988462425967234,
      "grad_norm": 23.0,
      "learning_rate": 2.2501730636104917e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9961089491844177,
      "num_tokens": 2832622.0,
      "step": 6499
    },
    {
      "epoch": 0.4999615414198908,
      "grad_norm": 46.5,
      "learning_rate": 2.250057687870164e-06,
      "loss": 0.0393,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2832922.0,
      "step": 6500
    },
    {
      "epoch": 0.5000384585801092,
      "grad_norm": 2.0625,
      "learning_rate": 2.249942312129836e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 2833446.0,
      "step": 6501
    },
    {
      "epoch": 0.5001153757403277,
      "grad_norm": 20.75,
      "learning_rate": 2.2498269363895084e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9940476417541504,
      "num_tokens": 2833784.0,
      "step": 6502
    },
    {
      "epoch": 0.5001922929005461,
      "grad_norm": 8.625,
      "learning_rate": 2.249711560649181e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2834074.0,
      "step": 6503
    },
    {
      "epoch": 0.5002692100607645,
      "grad_norm": 9.75,
      "learning_rate": 2.249596184908853e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 2834408.0,
      "step": 6504
    },
    {
      "epoch": 0.500346127220983,
      "grad_norm": 4.9375,
      "learning_rate": 2.2494808091685256e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2834954.0,
      "step": 6505
    },
    {
      "epoch": 0.5004230443812014,
      "grad_norm": 13.5625,
      "learning_rate": 2.249365433428198e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2835300.0,
      "step": 6506
    },
    {
      "epoch": 0.5004999615414198,
      "grad_norm": 11.625,
      "learning_rate": 2.24925005768787e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2835890.0,
      "step": 6507
    },
    {
      "epoch": 0.5005768787016384,
      "grad_norm": 3.234375,
      "learning_rate": 2.249134681947543e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2836464.0,
      "step": 6508
    },
    {
      "epoch": 0.5006537958618568,
      "grad_norm": 4.8125,
      "learning_rate": 2.2490193062072147e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2837272.0,
      "step": 6509
    },
    {
      "epoch": 0.5007307130220753,
      "grad_norm": 19.125,
      "learning_rate": 2.248903930466887e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2837596.0,
      "step": 6510
    },
    {
      "epoch": 0.5008076301822937,
      "grad_norm": 23.0,
      "learning_rate": 2.2487885547265596e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 2837876.0,
      "step": 6511
    },
    {
      "epoch": 0.5008845473425121,
      "grad_norm": 3.71875,
      "learning_rate": 2.248673178986232e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2838424.0,
      "step": 6512
    },
    {
      "epoch": 0.5009614645027306,
      "grad_norm": 4.78125,
      "learning_rate": 2.248557803245904e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 2839184.0,
      "step": 6513
    },
    {
      "epoch": 0.501038381662949,
      "grad_norm": 11.25,
      "learning_rate": 2.248442427505577e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2839502.0,
      "step": 6514
    },
    {
      "epoch": 0.5011152988231674,
      "grad_norm": 5.25,
      "learning_rate": 2.2483270517652486e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9936224222183228,
      "num_tokens": 2840288.0,
      "step": 6515
    },
    {
      "epoch": 0.5011922159833859,
      "grad_norm": 8.8125,
      "learning_rate": 2.2482116760249213e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9913344979286194,
      "num_tokens": 2840866.0,
      "step": 6516
    },
    {
      "epoch": 0.5012691331436043,
      "grad_norm": 20.875,
      "learning_rate": 2.2480963002845936e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 2841184.0,
      "step": 6517
    },
    {
      "epoch": 0.5013460503038227,
      "grad_norm": 15.5,
      "learning_rate": 2.247980924544266e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 2841480.0,
      "step": 6518
    },
    {
      "epoch": 0.5014229674640412,
      "grad_norm": 9.5,
      "learning_rate": 2.2478655488039385e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 2841976.0,
      "step": 6519
    },
    {
      "epoch": 0.5014998846242597,
      "grad_norm": 7.875,
      "learning_rate": 2.2477501730636104e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2842296.0,
      "step": 6520
    },
    {
      "epoch": 0.5015768017844782,
      "grad_norm": 10.875,
      "learning_rate": 2.2476347973232826e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2842610.0,
      "step": 6521
    },
    {
      "epoch": 0.5016537189446966,
      "grad_norm": 21.875,
      "learning_rate": 2.2475194215829553e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 2843190.0,
      "step": 6522
    },
    {
      "epoch": 0.501730636104915,
      "grad_norm": 5.3125,
      "learning_rate": 2.2474040458426276e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2843474.0,
      "step": 6523
    },
    {
      "epoch": 0.5018075532651335,
      "grad_norm": 12.3125,
      "learning_rate": 2.2472886701023e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2843780.0,
      "step": 6524
    },
    {
      "epoch": 0.5018844704253519,
      "grad_norm": 58.5,
      "learning_rate": 2.247173294361972e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2844144.0,
      "step": 6525
    },
    {
      "epoch": 0.5019613875855703,
      "grad_norm": 8.9375,
      "learning_rate": 2.2470579186216443e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 2844436.0,
      "step": 6526
    },
    {
      "epoch": 0.5020383047457888,
      "grad_norm": 5.65625,
      "learning_rate": 2.246942542881317e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 2845220.0,
      "step": 6527
    },
    {
      "epoch": 0.5021152219060072,
      "grad_norm": 38.0,
      "learning_rate": 2.2468271671409893e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2845550.0,
      "step": 6528
    },
    {
      "epoch": 0.5021921390662257,
      "grad_norm": 58.75,
      "learning_rate": 2.2467117914006615e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2845910.0,
      "step": 6529
    },
    {
      "epoch": 0.5022690562264441,
      "grad_norm": 29.75,
      "learning_rate": 2.2465964156603342e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 2846204.0,
      "step": 6530
    },
    {
      "epoch": 0.5023459733866625,
      "grad_norm": 9.8125,
      "learning_rate": 2.246481039920006e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2846528.0,
      "step": 6531
    },
    {
      "epoch": 0.502422890546881,
      "grad_norm": 26.0,
      "learning_rate": 2.2463656641796783e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.9928057789802551,
      "num_tokens": 2846808.0,
      "step": 6532
    },
    {
      "epoch": 0.5024998077070995,
      "grad_norm": 3.375,
      "learning_rate": 2.246250288439351e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2847398.0,
      "step": 6533
    },
    {
      "epoch": 0.5025767248673179,
      "grad_norm": 22.875,
      "learning_rate": 2.2461349126990232e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2847682.0,
      "step": 6534
    },
    {
      "epoch": 0.5026536420275364,
      "grad_norm": 18.125,
      "learning_rate": 2.2460195369586955e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2847966.0,
      "step": 6535
    },
    {
      "epoch": 0.5027305591877548,
      "grad_norm": 14.1875,
      "learning_rate": 2.2459041612183678e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2848310.0,
      "step": 6536
    },
    {
      "epoch": 0.5028074763479732,
      "grad_norm": 31.875,
      "learning_rate": 2.24578878547804e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 2848638.0,
      "step": 6537
    },
    {
      "epoch": 0.5028843935081917,
      "grad_norm": 6.09375,
      "learning_rate": 2.2456734097377127e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 2849418.0,
      "step": 6538
    },
    {
      "epoch": 0.5029613106684101,
      "grad_norm": 27.625,
      "learning_rate": 2.245558033997385e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 2849650.0,
      "step": 6539
    },
    {
      "epoch": 0.5030382278286286,
      "grad_norm": 27.0,
      "learning_rate": 2.2454426582570572e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9892086386680603,
      "num_tokens": 2849930.0,
      "step": 6540
    },
    {
      "epoch": 0.503115144988847,
      "grad_norm": 16.625,
      "learning_rate": 2.24532728251673e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 2850230.0,
      "step": 6541
    },
    {
      "epoch": 0.5031920621490654,
      "grad_norm": 6.25,
      "learning_rate": 2.2452119067764017e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2850538.0,
      "step": 6542
    },
    {
      "epoch": 0.5032689793092839,
      "grad_norm": 29.0,
      "learning_rate": 2.245096531036074e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2850884.0,
      "step": 6543
    },
    {
      "epoch": 0.5033458964695023,
      "grad_norm": 4.09375,
      "learning_rate": 2.2449811552957467e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2851166.0,
      "step": 6544
    },
    {
      "epoch": 0.5034228136297207,
      "grad_norm": 7.4375,
      "learning_rate": 2.244865779555419e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2851452.0,
      "step": 6545
    },
    {
      "epoch": 0.5034997307899393,
      "grad_norm": 5.0625,
      "learning_rate": 2.244750403815091e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2852004.0,
      "step": 6546
    },
    {
      "epoch": 0.5035766479501577,
      "grad_norm": 11.4375,
      "learning_rate": 2.2446350280747634e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 1.0,
      "num_tokens": 2852340.0,
      "step": 6547
    },
    {
      "epoch": 0.5036535651103762,
      "grad_norm": 24.375,
      "learning_rate": 2.2445196523344357e-06,
      "loss": 0.0401,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2852656.0,
      "step": 6548
    },
    {
      "epoch": 0.5037304822705946,
      "grad_norm": 18.75,
      "learning_rate": 2.2444042765941084e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2852984.0,
      "step": 6549
    },
    {
      "epoch": 0.503807399430813,
      "grad_norm": 14.4375,
      "learning_rate": 2.2442889008537806e-06,
      "loss": 0.0411,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2853304.0,
      "step": 6550
    },
    {
      "epoch": 0.5038843165910315,
      "grad_norm": 4.3125,
      "learning_rate": 2.244173525113453e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2853850.0,
      "step": 6551
    },
    {
      "epoch": 0.5039612337512499,
      "grad_norm": 41.25,
      "learning_rate": 2.244058149373125e-06,
      "loss": 0.0415,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2854210.0,
      "step": 6552
    },
    {
      "epoch": 0.5040381509114683,
      "grad_norm": 4.6875,
      "learning_rate": 2.2439427736327974e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2854498.0,
      "step": 6553
    },
    {
      "epoch": 0.5041150680716868,
      "grad_norm": 14.1875,
      "learning_rate": 2.2438273978924697e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 2855058.0,
      "step": 6554
    },
    {
      "epoch": 0.5041919852319052,
      "grad_norm": 14.0625,
      "learning_rate": 2.2437120221521424e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2855614.0,
      "step": 6555
    },
    {
      "epoch": 0.5042689023921236,
      "grad_norm": 9.625,
      "learning_rate": 2.2435966464118146e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 2855894.0,
      "step": 6556
    },
    {
      "epoch": 0.5043458195523421,
      "grad_norm": 4.59375,
      "learning_rate": 2.243481270671487e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 2856448.0,
      "step": 6557
    },
    {
      "epoch": 0.5044227367125605,
      "grad_norm": 14.6875,
      "learning_rate": 2.243365894931159e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 2856700.0,
      "step": 6558
    },
    {
      "epoch": 0.5044996538727791,
      "grad_norm": 3.296875,
      "learning_rate": 2.2432505191908314e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2857246.0,
      "step": 6559
    },
    {
      "epoch": 0.5045765710329975,
      "grad_norm": 7.125,
      "learning_rate": 2.2431351434505036e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2857530.0,
      "step": 6560
    },
    {
      "epoch": 0.5046534881932159,
      "grad_norm": 3.21875,
      "learning_rate": 2.2430197677101763e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2858086.0,
      "step": 6561
    },
    {
      "epoch": 0.5047304053534344,
      "grad_norm": 5.625,
      "learning_rate": 2.2429043919698486e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 2858666.0,
      "step": 6562
    },
    {
      "epoch": 0.5048073225136528,
      "grad_norm": 10.4375,
      "learning_rate": 2.242789016229521e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 2858976.0,
      "step": 6563
    },
    {
      "epoch": 0.5048842396738712,
      "grad_norm": 5.59375,
      "learning_rate": 2.242673640489193e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 2859246.0,
      "step": 6564
    },
    {
      "epoch": 0.5049611568340897,
      "grad_norm": 15.4375,
      "learning_rate": 2.2425582647488654e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2859548.0,
      "step": 6565
    },
    {
      "epoch": 0.5050380739943081,
      "grad_norm": 5.15625,
      "learning_rate": 2.242442889008538e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2859868.0,
      "step": 6566
    },
    {
      "epoch": 0.5051149911545266,
      "grad_norm": 2.21875,
      "learning_rate": 2.2423275132682103e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2860676.0,
      "step": 6567
    },
    {
      "epoch": 0.505191908314745,
      "grad_norm": 8.4375,
      "learning_rate": 2.2422121375278826e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2861004.0,
      "step": 6568
    },
    {
      "epoch": 0.5052688254749634,
      "grad_norm": 7.21875,
      "learning_rate": 2.242096761787555e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 2861320.0,
      "step": 6569
    },
    {
      "epoch": 0.505345742635182,
      "grad_norm": 3.109375,
      "learning_rate": 2.241981386047227e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 2862102.0,
      "step": 6570
    },
    {
      "epoch": 0.5054226597954004,
      "grad_norm": 3.09375,
      "learning_rate": 2.2418660103068993e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2862658.0,
      "step": 6571
    },
    {
      "epoch": 0.5054995769556188,
      "grad_norm": 5.90625,
      "learning_rate": 2.241750634566572e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2862986.0,
      "step": 6572
    },
    {
      "epoch": 0.5055764941158373,
      "grad_norm": 8.6875,
      "learning_rate": 2.2416352588262443e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 2863302.0,
      "step": 6573
    },
    {
      "epoch": 0.5056534112760557,
      "grad_norm": 8.75,
      "learning_rate": 2.2415198830859165e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2863620.0,
      "step": 6574
    },
    {
      "epoch": 0.5057303284362741,
      "grad_norm": 18.25,
      "learning_rate": 2.241404507345589e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2863950.0,
      "step": 6575
    },
    {
      "epoch": 0.5058072455964926,
      "grad_norm": 17.125,
      "learning_rate": 2.241289131605261e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2864278.0,
      "step": 6576
    },
    {
      "epoch": 0.505884162756711,
      "grad_norm": 10.6875,
      "learning_rate": 2.2411737558649337e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2864562.0,
      "step": 6577
    },
    {
      "epoch": 0.5059610799169295,
      "grad_norm": 3.75,
      "learning_rate": 2.241058380124606e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2865342.0,
      "step": 6578
    },
    {
      "epoch": 0.5060379970771479,
      "grad_norm": 7.84375,
      "learning_rate": 2.2409430043842782e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 2865932.0,
      "step": 6579
    },
    {
      "epoch": 0.5061149142373663,
      "grad_norm": 6.25,
      "learning_rate": 2.2408276286439505e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 2866250.0,
      "step": 6580
    },
    {
      "epoch": 0.5061918313975848,
      "grad_norm": 3.0625,
      "learning_rate": 2.2407122529036228e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 2866772.0,
      "step": 6581
    },
    {
      "epoch": 0.5062687485578032,
      "grad_norm": 18.25,
      "learning_rate": 2.240596877163295e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 2867072.0,
      "step": 6582
    },
    {
      "epoch": 0.5063456657180216,
      "grad_norm": 17.375,
      "learning_rate": 2.2404815014229677e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 2867374.0,
      "step": 6583
    },
    {
      "epoch": 0.5064225828782402,
      "grad_norm": 5.40625,
      "learning_rate": 2.24036612568264e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 2868132.0,
      "step": 6584
    },
    {
      "epoch": 0.5064995000384586,
      "grad_norm": 26.125,
      "learning_rate": 2.2402507499423122e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 2868494.0,
      "step": 6585
    },
    {
      "epoch": 0.5065764171986771,
      "grad_norm": 15.3125,
      "learning_rate": 2.2401353742019845e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 2868770.0,
      "step": 6586
    },
    {
      "epoch": 0.5066533343588955,
      "grad_norm": 3.515625,
      "learning_rate": 2.2400199984616567e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2869312.0,
      "step": 6587
    },
    {
      "epoch": 0.5067302515191139,
      "grad_norm": 12.5625,
      "learning_rate": 2.2399046227213294e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 2869656.0,
      "step": 6588
    },
    {
      "epoch": 0.5068071686793324,
      "grad_norm": 6.71875,
      "learning_rate": 2.2397892469810017e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9962476491928101,
      "num_tokens": 2870190.0,
      "step": 6589
    },
    {
      "epoch": 0.5068840858395508,
      "grad_norm": 10.25,
      "learning_rate": 2.2396738712406735e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 2870520.0,
      "step": 6590
    },
    {
      "epoch": 0.5069610029997692,
      "grad_norm": 3.40625,
      "learning_rate": 2.239558495500346e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2871094.0,
      "step": 6591
    },
    {
      "epoch": 0.5070379201599877,
      "grad_norm": 19.75,
      "learning_rate": 2.2394431197600184e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 2871400.0,
      "step": 6592
    },
    {
      "epoch": 0.5071148373202061,
      "grad_norm": 22.625,
      "learning_rate": 2.2393277440196907e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2871728.0,
      "step": 6593
    },
    {
      "epoch": 0.5071917544804245,
      "grad_norm": 9.75,
      "learning_rate": 2.2392123682793634e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2872290.0,
      "step": 6594
    },
    {
      "epoch": 0.507268671640643,
      "grad_norm": 17.625,
      "learning_rate": 2.2390969925390356e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 2872598.0,
      "step": 6595
    },
    {
      "epoch": 0.5073455888008614,
      "grad_norm": 13.25,
      "learning_rate": 2.238981616798708e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 2872926.0,
      "step": 6596
    },
    {
      "epoch": 0.50742250596108,
      "grad_norm": 31.0,
      "learning_rate": 2.23886624105838e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 2873298.0,
      "step": 6597
    },
    {
      "epoch": 0.5074994231212984,
      "grad_norm": 2.6875,
      "learning_rate": 2.2387508653180524e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2873872.0,
      "step": 6598
    },
    {
      "epoch": 0.5075763402815168,
      "grad_norm": 16.0,
      "learning_rate": 2.238635489577725e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2874178.0,
      "step": 6599
    },
    {
      "epoch": 0.5076532574417353,
      "grad_norm": 25.625,
      "learning_rate": 2.2385201138373974e-06,
      "loss": 0.0418,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 2874410.0,
      "step": 6600
    },
    {
      "epoch": 0.5077301746019537,
      "grad_norm": 6.53125,
      "learning_rate": 2.238404738097069e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 2874926.0,
      "step": 6601
    },
    {
      "epoch": 0.5078070917621721,
      "grad_norm": 8.8125,
      "learning_rate": 2.238289362356742e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 2875272.0,
      "step": 6602
    },
    {
      "epoch": 0.5078840089223906,
      "grad_norm": 11.3125,
      "learning_rate": 2.238173986616414e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 2875532.0,
      "step": 6603
    },
    {
      "epoch": 0.507960926082609,
      "grad_norm": 4.96875,
      "learning_rate": 2.2380586108760864e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 2876072.0,
      "step": 6604
    },
    {
      "epoch": 0.5080378432428275,
      "grad_norm": 3.65625,
      "learning_rate": 2.237943235135759e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 2876880.0,
      "step": 6605
    },
    {
      "epoch": 0.5081147604030459,
      "grad_norm": 2.9375,
      "learning_rate": 2.2378278593954313e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 2877454.0,
      "step": 6606
    },
    {
      "epoch": 0.5081916775632643,
      "grad_norm": 4.28125,
      "learning_rate": 2.2377124836551036e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 2878018.0,
      "step": 6607
    },
    {
      "epoch": 0.5082685947234828,
      "grad_norm": 33.25,
      "learning_rate": 2.237597107914776e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2878380.0,
      "step": 6608
    },
    {
      "epoch": 0.5083455118837013,
      "grad_norm": 14.0,
      "learning_rate": 2.237481732174448e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 2878912.0,
      "step": 6609
    },
    {
      "epoch": 0.5084224290439197,
      "grad_norm": 5.0625,
      "learning_rate": 2.2373663564341204e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 2879252.0,
      "step": 6610
    },
    {
      "epoch": 0.5084993462041382,
      "grad_norm": 11.125,
      "learning_rate": 2.237250980693793e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2879808.0,
      "step": 6611
    },
    {
      "epoch": 0.5085762633643566,
      "grad_norm": 5.15625,
      "learning_rate": 2.237135604953465e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2880130.0,
      "step": 6612
    },
    {
      "epoch": 0.5086531805245751,
      "grad_norm": 10.125,
      "learning_rate": 2.2370202292131376e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 2880438.0,
      "step": 6613
    },
    {
      "epoch": 0.5087300976847935,
      "grad_norm": 7.5,
      "learning_rate": 2.23690485347281e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2880996.0,
      "step": 6614
    },
    {
      "epoch": 0.5088070148450119,
      "grad_norm": 9.5,
      "learning_rate": 2.236789477732482e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2881308.0,
      "step": 6615
    },
    {
      "epoch": 0.5088839320052304,
      "grad_norm": 4.03125,
      "learning_rate": 2.2366741019921548e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2881860.0,
      "step": 6616
    },
    {
      "epoch": 0.5089608491654488,
      "grad_norm": 5.09375,
      "learning_rate": 2.236558726251827e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2882430.0,
      "step": 6617
    },
    {
      "epoch": 0.5090377663256672,
      "grad_norm": 4.625,
      "learning_rate": 2.236443350511499e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 2883256.0,
      "step": 6618
    },
    {
      "epoch": 0.5091146834858857,
      "grad_norm": 11.25,
      "learning_rate": 2.2363279747711715e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 2883518.0,
      "step": 6619
    },
    {
      "epoch": 0.5091916006461041,
      "grad_norm": 3.546875,
      "learning_rate": 2.236212599030844e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2884066.0,
      "step": 6620
    },
    {
      "epoch": 0.5092685178063225,
      "grad_norm": 11.875,
      "learning_rate": 2.236097223290516e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2884356.0,
      "step": 6621
    },
    {
      "epoch": 0.509345434966541,
      "grad_norm": 26.25,
      "learning_rate": 2.2359818475501887e-06,
      "loss": 0.0423,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 2884610.0,
      "step": 6622
    },
    {
      "epoch": 0.5094223521267595,
      "grad_norm": 13.125,
      "learning_rate": 2.2358664718098606e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 2884954.0,
      "step": 6623
    },
    {
      "epoch": 0.509499269286978,
      "grad_norm": 4.1875,
      "learning_rate": 2.2357510960695332e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2885472.0,
      "step": 6624
    },
    {
      "epoch": 0.5095761864471964,
      "grad_norm": 3.84375,
      "learning_rate": 2.2356357203292055e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 2886036.0,
      "step": 6625
    },
    {
      "epoch": 0.5096531036074148,
      "grad_norm": 4.03125,
      "learning_rate": 2.2355203445888778e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2886606.0,
      "step": 6626
    },
    {
      "epoch": 0.5097300207676333,
      "grad_norm": 13.125,
      "learning_rate": 2.2354049688485504e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2886862.0,
      "step": 6627
    },
    {
      "epoch": 0.5098069379278517,
      "grad_norm": 18.75,
      "learning_rate": 2.2352895931082223e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 2887420.0,
      "step": 6628
    },
    {
      "epoch": 0.5098838550880701,
      "grad_norm": 6.78125,
      "learning_rate": 2.2351742173678945e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 2887730.0,
      "step": 6629
    },
    {
      "epoch": 0.5099607722482886,
      "grad_norm": 6.46875,
      "learning_rate": 2.2350588416275672e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2887990.0,
      "step": 6630
    },
    {
      "epoch": 0.510037689408507,
      "grad_norm": 9.9375,
      "learning_rate": 2.2349434658872395e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2888306.0,
      "step": 6631
    },
    {
      "epoch": 0.5101146065687255,
      "grad_norm": 6.4375,
      "learning_rate": 2.2348280901469117e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2888852.0,
      "step": 6632
    },
    {
      "epoch": 0.5101915237289439,
      "grad_norm": 7.5,
      "learning_rate": 2.2347127144065844e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 2889130.0,
      "step": 6633
    },
    {
      "epoch": 0.5102684408891623,
      "grad_norm": 18.625,
      "learning_rate": 2.2345973386662563e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 2889470.0,
      "step": 6634
    },
    {
      "epoch": 0.5103453580493809,
      "grad_norm": 7.34375,
      "learning_rate": 2.234481962925929e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2889784.0,
      "step": 6635
    },
    {
      "epoch": 0.5104222752095993,
      "grad_norm": 28.625,
      "learning_rate": 2.234366587185601e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.9850187301635742,
      "num_tokens": 2890052.0,
      "step": 6636
    },
    {
      "epoch": 0.5104991923698177,
      "grad_norm": 24.875,
      "learning_rate": 2.2342512114452735e-06,
      "loss": 0.038,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 2890322.0,
      "step": 6637
    },
    {
      "epoch": 0.5105761095300362,
      "grad_norm": 20.875,
      "learning_rate": 2.234135835704946e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2890644.0,
      "step": 6638
    },
    {
      "epoch": 0.5106530266902546,
      "grad_norm": 15.125,
      "learning_rate": 2.234020459964618e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 2890960.0,
      "step": 6639
    },
    {
      "epoch": 0.510729943850473,
      "grad_norm": 18.375,
      "learning_rate": 2.2339050842242902e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2891266.0,
      "step": 6640
    },
    {
      "epoch": 0.5108068610106915,
      "grad_norm": 7.34375,
      "learning_rate": 2.233789708483963e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 2891838.0,
      "step": 6641
    },
    {
      "epoch": 0.5108837781709099,
      "grad_norm": 10.0625,
      "learning_rate": 2.233674332743635e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2892146.0,
      "step": 6642
    },
    {
      "epoch": 0.5109606953311284,
      "grad_norm": 13.25,
      "learning_rate": 2.2335589570033074e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 2892484.0,
      "step": 6643
    },
    {
      "epoch": 0.5110376124913468,
      "grad_norm": 7.09375,
      "learning_rate": 2.23344358126298e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 2892834.0,
      "step": 6644
    },
    {
      "epoch": 0.5111145296515652,
      "grad_norm": 11.875,
      "learning_rate": 2.233328205522652e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 2893092.0,
      "step": 6645
    },
    {
      "epoch": 0.5111914468117837,
      "grad_norm": 4.375,
      "learning_rate": 2.2332128297823246e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9980732202529907,
      "num_tokens": 2893612.0,
      "step": 6646
    },
    {
      "epoch": 0.5112683639720021,
      "grad_norm": 10.4375,
      "learning_rate": 2.233097454041997e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2893968.0,
      "step": 6647
    },
    {
      "epoch": 0.5113452811322206,
      "grad_norm": 3.390625,
      "learning_rate": 2.232982078301669e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2894516.0,
      "step": 6648
    },
    {
      "epoch": 0.5114221982924391,
      "grad_norm": 8.6875,
      "learning_rate": 2.232866702561342e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 2894828.0,
      "step": 6649
    },
    {
      "epoch": 0.5114991154526575,
      "grad_norm": 28.875,
      "learning_rate": 2.2327513268210137e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2895082.0,
      "step": 6650
    },
    {
      "epoch": 0.511576032612876,
      "grad_norm": 5.59375,
      "learning_rate": 2.232635951080686e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2895646.0,
      "step": 6651
    },
    {
      "epoch": 0.5116529497730944,
      "grad_norm": 16.625,
      "learning_rate": 2.2325205753403586e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.978723406791687,
      "num_tokens": 2895930.0,
      "step": 6652
    },
    {
      "epoch": 0.5117298669333128,
      "grad_norm": 5.40625,
      "learning_rate": 2.232405199600031e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 2896216.0,
      "step": 6653
    },
    {
      "epoch": 0.5118067840935313,
      "grad_norm": 8.25,
      "learning_rate": 2.232289823859703e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 2896536.0,
      "step": 6654
    },
    {
      "epoch": 0.5118837012537497,
      "grad_norm": 6.46875,
      "learning_rate": 2.232174448119376e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 2896876.0,
      "step": 6655
    },
    {
      "epoch": 0.5119606184139681,
      "grad_norm": 13.25,
      "learning_rate": 2.2320590723790476e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 2897432.0,
      "step": 6656
    },
    {
      "epoch": 0.5120375355741866,
      "grad_norm": 3.234375,
      "learning_rate": 2.2319436966387203e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2897978.0,
      "step": 6657
    },
    {
      "epoch": 0.512114452734405,
      "grad_norm": 6.625,
      "learning_rate": 2.2318283208983926e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 2898294.0,
      "step": 6658
    },
    {
      "epoch": 0.5121913698946234,
      "grad_norm": 28.5,
      "learning_rate": 2.231712945158065e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 2898644.0,
      "step": 6659
    },
    {
      "epoch": 0.512268287054842,
      "grad_norm": 2.96875,
      "learning_rate": 2.231597569417737e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 2899430.0,
      "step": 6660
    },
    {
      "epoch": 0.5123452042150604,
      "grad_norm": 11.4375,
      "learning_rate": 2.2314821936774093e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 2899686.0,
      "step": 6661
    },
    {
      "epoch": 0.5124221213752789,
      "grad_norm": 8.9375,
      "learning_rate": 2.2313668179370816e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 2899994.0,
      "step": 6662
    },
    {
      "epoch": 0.5124990385354973,
      "grad_norm": 12.9375,
      "learning_rate": 2.2312514421967543e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 2900330.0,
      "step": 6663
    },
    {
      "epoch": 0.5125759556957157,
      "grad_norm": 7.4375,
      "learning_rate": 2.2311360664564265e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 2900620.0,
      "step": 6664
    },
    {
      "epoch": 0.5126528728559342,
      "grad_norm": 12.1875,
      "learning_rate": 2.231020690716099e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 2900922.0,
      "step": 6665
    },
    {
      "epoch": 0.5127297900161526,
      "grad_norm": 4.59375,
      "learning_rate": 2.230905314975771e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 2901454.0,
      "step": 6666
    },
    {
      "epoch": 0.512806707176371,
      "grad_norm": 12.4375,
      "learning_rate": 2.2307899392354433e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9858757257461548,
      "num_tokens": 2901810.0,
      "step": 6667
    },
    {
      "epoch": 0.5128836243365895,
      "grad_norm": 2.59375,
      "learning_rate": 2.2306745634951156e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 2902368.0,
      "step": 6668
    },
    {
      "epoch": 0.5129605414968079,
      "grad_norm": 11.0625,
      "learning_rate": 2.2305591877547883e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 2902958.0,
      "step": 6669
    },
    {
      "epoch": 0.5130374586570264,
      "grad_norm": 32.75,
      "learning_rate": 2.2304438120144605e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2903262.0,
      "step": 6670
    },
    {
      "epoch": 0.5131143758172448,
      "grad_norm": 8.6875,
      "learning_rate": 2.2303284362741328e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2903806.0,
      "step": 6671
    },
    {
      "epoch": 0.5131912929774632,
      "grad_norm": 12.6875,
      "learning_rate": 2.230213060533805e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9971618056297302,
      "num_tokens": 2904866.0,
      "step": 6672
    },
    {
      "epoch": 0.5132682101376818,
      "grad_norm": 20.75,
      "learning_rate": 2.2300976847934773e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 2905120.0,
      "step": 6673
    },
    {
      "epoch": 0.5133451272979002,
      "grad_norm": 9.6875,
      "learning_rate": 2.22998230905315e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 2905462.0,
      "step": 6674
    },
    {
      "epoch": 0.5134220444581186,
      "grad_norm": 12.6875,
      "learning_rate": 2.2298669333128222e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 2906014.0,
      "step": 6675
    },
    {
      "epoch": 0.5134989616183371,
      "grad_norm": 7.59375,
      "learning_rate": 2.2297515575724945e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 2906326.0,
      "step": 6676
    },
    {
      "epoch": 0.5135758787785555,
      "grad_norm": 23.625,
      "learning_rate": 2.2296361818321667e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2906648.0,
      "step": 6677
    },
    {
      "epoch": 0.5136527959387739,
      "grad_norm": 11.6875,
      "learning_rate": 2.229520806091839e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 2907468.0,
      "step": 6678
    },
    {
      "epoch": 0.5137297130989924,
      "grad_norm": 40.75,
      "learning_rate": 2.2294054303515113e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2907802.0,
      "step": 6679
    },
    {
      "epoch": 0.5138066302592108,
      "grad_norm": 13.625,
      "learning_rate": 2.229290054611184e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9889298677444458,
      "num_tokens": 2908074.0,
      "step": 6680
    },
    {
      "epoch": 0.5138835474194293,
      "grad_norm": 9.4375,
      "learning_rate": 2.229174678870856e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 2908410.0,
      "step": 6681
    },
    {
      "epoch": 0.5139604645796477,
      "grad_norm": 12.1875,
      "learning_rate": 2.2290593031305285e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2908954.0,
      "step": 6682
    },
    {
      "epoch": 0.5140373817398661,
      "grad_norm": 5.03125,
      "learning_rate": 2.2289439273902007e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2909764.0,
      "step": 6683
    },
    {
      "epoch": 0.5141142989000846,
      "grad_norm": 3.21875,
      "learning_rate": 2.228828551649873e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2910574.0,
      "step": 6684
    },
    {
      "epoch": 0.514191216060303,
      "grad_norm": 23.875,
      "learning_rate": 2.2287131759095457e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 2910858.0,
      "step": 6685
    },
    {
      "epoch": 0.5142681332205215,
      "grad_norm": 2.34375,
      "learning_rate": 2.228597800169218e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 2911682.0,
      "step": 6686
    },
    {
      "epoch": 0.51434505038074,
      "grad_norm": 7.5,
      "learning_rate": 2.22848242442889e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2912202.0,
      "step": 6687
    },
    {
      "epoch": 0.5144219675409584,
      "grad_norm": 8.125,
      "learning_rate": 2.2283670486885624e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 2912772.0,
      "step": 6688
    },
    {
      "epoch": 0.5144988847011769,
      "grad_norm": 22.125,
      "learning_rate": 2.2282516729482347e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2913064.0,
      "step": 6689
    },
    {
      "epoch": 0.5145758018613953,
      "grad_norm": 4.25,
      "learning_rate": 2.228136297207907e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2913610.0,
      "step": 6690
    },
    {
      "epoch": 0.5146527190216137,
      "grad_norm": 24.125,
      "learning_rate": 2.2280209214675796e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2913894.0,
      "step": 6691
    },
    {
      "epoch": 0.5147296361818322,
      "grad_norm": 14.625,
      "learning_rate": 2.227905545727252e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2914256.0,
      "step": 6692
    },
    {
      "epoch": 0.5148065533420506,
      "grad_norm": 23.75,
      "learning_rate": 2.227790169986924e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 2914592.0,
      "step": 6693
    },
    {
      "epoch": 0.514883470502269,
      "grad_norm": 6.59375,
      "learning_rate": 2.2276747942465964e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 2914904.0,
      "step": 6694
    },
    {
      "epoch": 0.5149603876624875,
      "grad_norm": 3.75,
      "learning_rate": 2.2275594185062687e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 2915472.0,
      "step": 6695
    },
    {
      "epoch": 0.5150373048227059,
      "grad_norm": 11.6875,
      "learning_rate": 2.2274440427659413e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 2915728.0,
      "step": 6696
    },
    {
      "epoch": 0.5151142219829243,
      "grad_norm": 16.25,
      "learning_rate": 2.2273286670256136e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 2916290.0,
      "step": 6697
    },
    {
      "epoch": 0.5151911391431429,
      "grad_norm": 5.90625,
      "learning_rate": 2.227213291285286e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2916574.0,
      "step": 6698
    },
    {
      "epoch": 0.5152680563033613,
      "grad_norm": 15.1875,
      "learning_rate": 2.227097915544958e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2916876.0,
      "step": 6699
    },
    {
      "epoch": 0.5153449734635798,
      "grad_norm": 12.1875,
      "learning_rate": 2.2269825398046304e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 2917200.0,
      "step": 6700
    },
    {
      "epoch": 0.5154218906237982,
      "grad_norm": 21.25,
      "learning_rate": 2.2268671640643026e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 2917550.0,
      "step": 6701
    },
    {
      "epoch": 0.5154988077840166,
      "grad_norm": 10.0625,
      "learning_rate": 2.2267517883239753e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 2918358.0,
      "step": 6702
    },
    {
      "epoch": 0.5155757249442351,
      "grad_norm": 24.75,
      "learning_rate": 2.2266364125836476e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 2918710.0,
      "step": 6703
    },
    {
      "epoch": 0.5156526421044535,
      "grad_norm": 7.6875,
      "learning_rate": 2.22652103684332e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2918994.0,
      "step": 6704
    },
    {
      "epoch": 0.5157295592646719,
      "grad_norm": 46.0,
      "learning_rate": 2.226405661102992e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 2919326.0,
      "step": 6705
    },
    {
      "epoch": 0.5158064764248904,
      "grad_norm": 8.1875,
      "learning_rate": 2.2262902853626643e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9904030561447144,
      "num_tokens": 2919848.0,
      "step": 6706
    },
    {
      "epoch": 0.5158833935851088,
      "grad_norm": 22.0,
      "learning_rate": 2.226174909622337e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2920194.0,
      "step": 6707
    },
    {
      "epoch": 0.5159603107453273,
      "grad_norm": 2.609375,
      "learning_rate": 2.2260595338820093e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 2920734.0,
      "step": 6708
    },
    {
      "epoch": 0.5160372279055457,
      "grad_norm": 16.875,
      "learning_rate": 2.2259441581416815e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 2921090.0,
      "step": 6709
    },
    {
      "epoch": 0.5161141450657641,
      "grad_norm": 20.125,
      "learning_rate": 2.225828782401354e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 2921414.0,
      "step": 6710
    },
    {
      "epoch": 0.5161910622259827,
      "grad_norm": 30.375,
      "learning_rate": 2.225713406661026e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 2921748.0,
      "step": 6711
    },
    {
      "epoch": 0.5162679793862011,
      "grad_norm": 12.5,
      "learning_rate": 2.2255980309206983e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 2922038.0,
      "step": 6712
    },
    {
      "epoch": 0.5163448965464195,
      "grad_norm": 6.875,
      "learning_rate": 2.225482655180371e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2922338.0,
      "step": 6713
    },
    {
      "epoch": 0.516421813706638,
      "grad_norm": 14.75,
      "learning_rate": 2.2253672794400433e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2922908.0,
      "step": 6714
    },
    {
      "epoch": 0.5164987308668564,
      "grad_norm": 7.0625,
      "learning_rate": 2.2252519036997155e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 2923484.0,
      "step": 6715
    },
    {
      "epoch": 0.5165756480270748,
      "grad_norm": 17.625,
      "learning_rate": 2.2251365279593878e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 2923718.0,
      "step": 6716
    },
    {
      "epoch": 0.5166525651872933,
      "grad_norm": 8.875,
      "learning_rate": 2.22502115221906e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9826989769935608,
      "num_tokens": 2924008.0,
      "step": 6717
    },
    {
      "epoch": 0.5167294823475117,
      "grad_norm": 12.75,
      "learning_rate": 2.2249057764787327e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2924318.0,
      "step": 6718
    },
    {
      "epoch": 0.5168063995077302,
      "grad_norm": 4.71875,
      "learning_rate": 2.224790400738405e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9982425570487976,
      "num_tokens": 2924888.0,
      "step": 6719
    },
    {
      "epoch": 0.5168833166679486,
      "grad_norm": 3.890625,
      "learning_rate": 2.2246750249980772e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 2925458.0,
      "step": 6720
    },
    {
      "epoch": 0.516960233828167,
      "grad_norm": 22.625,
      "learning_rate": 2.2245596492577495e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2925762.0,
      "step": 6721
    },
    {
      "epoch": 0.5170371509883855,
      "grad_norm": 8.5625,
      "learning_rate": 2.2244442735174217e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 2926310.0,
      "step": 6722
    },
    {
      "epoch": 0.517114068148604,
      "grad_norm": 7.96875,
      "learning_rate": 2.224328897777094e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 2926598.0,
      "step": 6723
    },
    {
      "epoch": 0.5171909853088223,
      "grad_norm": 13.125,
      "learning_rate": 2.2242135220367667e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9788732528686523,
      "num_tokens": 2926884.0,
      "step": 6724
    },
    {
      "epoch": 0.5172679024690409,
      "grad_norm": 8.5625,
      "learning_rate": 2.224098146296439e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2927400.0,
      "step": 6725
    },
    {
      "epoch": 0.5173448196292593,
      "grad_norm": 5.4375,
      "learning_rate": 2.223982770556111e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9904030561447144,
      "num_tokens": 2927922.0,
      "step": 6726
    },
    {
      "epoch": 0.5174217367894778,
      "grad_norm": 14.125,
      "learning_rate": 2.2238673948157835e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2928742.0,
      "step": 6727
    },
    {
      "epoch": 0.5174986539496962,
      "grad_norm": 6.0625,
      "learning_rate": 2.2237520190754557e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 2929284.0,
      "step": 6728
    },
    {
      "epoch": 0.5175755711099146,
      "grad_norm": 8.25,
      "learning_rate": 2.223636643335128e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 2929544.0,
      "step": 6729
    },
    {
      "epoch": 0.5176524882701331,
      "grad_norm": 6.375,
      "learning_rate": 2.2235212675948007e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9960784316062927,
      "num_tokens": 2929800.0,
      "step": 6730
    },
    {
      "epoch": 0.5177294054303515,
      "grad_norm": 7.46875,
      "learning_rate": 2.2234058918544725e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 2930144.0,
      "step": 6731
    },
    {
      "epoch": 0.5178063225905699,
      "grad_norm": 19.25,
      "learning_rate": 2.223290516114145e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 2930466.0,
      "step": 6732
    },
    {
      "epoch": 0.5178832397507884,
      "grad_norm": 15.25,
      "learning_rate": 2.2231751403738174e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 2930744.0,
      "step": 6733
    },
    {
      "epoch": 0.5179601569110068,
      "grad_norm": 3.875,
      "learning_rate": 2.2230597646334897e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 2931288.0,
      "step": 6734
    },
    {
      "epoch": 0.5180370740712253,
      "grad_norm": 9.0,
      "learning_rate": 2.2229443888931624e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 2931836.0,
      "step": 6735
    },
    {
      "epoch": 0.5181139912314437,
      "grad_norm": 6.0,
      "learning_rate": 2.2228290131528346e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 2932420.0,
      "step": 6736
    },
    {
      "epoch": 0.5181909083916622,
      "grad_norm": 5.03125,
      "learning_rate": 2.2227136374125065e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2932770.0,
      "step": 6737
    },
    {
      "epoch": 0.5182678255518807,
      "grad_norm": 8.0625,
      "learning_rate": 2.222598261672179e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 2933114.0,
      "step": 6738
    },
    {
      "epoch": 0.5183447427120991,
      "grad_norm": 15.875,
      "learning_rate": 2.2224828859318514e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 2933454.0,
      "step": 6739
    },
    {
      "epoch": 0.5184216598723175,
      "grad_norm": 17.75,
      "learning_rate": 2.2223675101915237e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2933770.0,
      "step": 6740
    },
    {
      "epoch": 0.518498577032536,
      "grad_norm": 2.5625,
      "learning_rate": 2.2222521344511963e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9951969385147095,
      "num_tokens": 2934814.0,
      "step": 6741
    },
    {
      "epoch": 0.5185754941927544,
      "grad_norm": 2.859375,
      "learning_rate": 2.222136758710868e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 2935388.0,
      "step": 6742
    },
    {
      "epoch": 0.5186524113529728,
      "grad_norm": 14.5,
      "learning_rate": 2.222021382970541e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 2935694.0,
      "step": 6743
    },
    {
      "epoch": 0.5187293285131913,
      "grad_norm": 2.46875,
      "learning_rate": 2.221906007230213e-06,
      "loss": 0.0062,
      "mean_token_accuracy": 0.9990430474281311,
      "num_tokens": 2936742.0,
      "step": 6744
    },
    {
      "epoch": 0.5188062456734097,
      "grad_norm": 14.875,
      "learning_rate": 2.2217906314898854e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 2937046.0,
      "step": 6745
    },
    {
      "epoch": 0.5188831628336282,
      "grad_norm": 15.75,
      "learning_rate": 2.221675255749558e-06,
      "loss": 0.0371,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 2937302.0,
      "step": 6746
    },
    {
      "epoch": 0.5189600799938466,
      "grad_norm": 2.875,
      "learning_rate": 2.2215598800092303e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9987593293190002,
      "num_tokens": 2938110.0,
      "step": 6747
    },
    {
      "epoch": 0.519036997154065,
      "grad_norm": 25.125,
      "learning_rate": 2.221444504268902e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 2938434.0,
      "step": 6748
    },
    {
      "epoch": 0.5191139143142836,
      "grad_norm": 16.875,
      "learning_rate": 2.221329128528575e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 2938744.0,
      "step": 6749
    },
    {
      "epoch": 0.519190831474502,
      "grad_norm": 12.3125,
      "learning_rate": 2.221213752788247e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 2939094.0,
      "step": 6750
    },
    {
      "epoch": 0.5192677486347204,
      "grad_norm": 6.625,
      "learning_rate": 2.2210983770479193e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2939620.0,
      "step": 6751
    },
    {
      "epoch": 0.5193446657949389,
      "grad_norm": 5.28125,
      "learning_rate": 2.220983001307592e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2939954.0,
      "step": 6752
    },
    {
      "epoch": 0.5194215829551573,
      "grad_norm": 4.4375,
      "learning_rate": 2.220867625567264e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2940518.0,
      "step": 6753
    },
    {
      "epoch": 0.5194985001153758,
      "grad_norm": 7.78125,
      "learning_rate": 2.2207522498269365e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 2941070.0,
      "step": 6754
    },
    {
      "epoch": 0.5195754172755942,
      "grad_norm": 20.0,
      "learning_rate": 2.220636874086609e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 2941396.0,
      "step": 6755
    },
    {
      "epoch": 0.5196523344358126,
      "grad_norm": 5.8125,
      "learning_rate": 2.220521498346281e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 2941730.0,
      "step": 6756
    },
    {
      "epoch": 0.5197292515960311,
      "grad_norm": 18.5,
      "learning_rate": 2.2204061226059537e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2942054.0,
      "step": 6757
    },
    {
      "epoch": 0.5198061687562495,
      "grad_norm": 4.9375,
      "learning_rate": 2.220290746865626e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 2942578.0,
      "step": 6758
    },
    {
      "epoch": 0.5198830859164679,
      "grad_norm": 23.625,
      "learning_rate": 2.220175371125298e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 2942854.0,
      "step": 6759
    },
    {
      "epoch": 0.5199600030766864,
      "grad_norm": 4.78125,
      "learning_rate": 2.2200599953849705e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2943410.0,
      "step": 6760
    },
    {
      "epoch": 0.5200369202369048,
      "grad_norm": 6.25,
      "learning_rate": 2.2199446196446428e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 2943984.0,
      "step": 6761
    },
    {
      "epoch": 0.5201138373971232,
      "grad_norm": 3.03125,
      "learning_rate": 2.219829243904315e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2944530.0,
      "step": 6762
    },
    {
      "epoch": 0.5201907545573418,
      "grad_norm": 16.0,
      "learning_rate": 2.2197138681639877e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2944822.0,
      "step": 6763
    },
    {
      "epoch": 0.5202676717175602,
      "grad_norm": 2.078125,
      "learning_rate": 2.2195984924236595e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9974554777145386,
      "num_tokens": 2945610.0,
      "step": 6764
    },
    {
      "epoch": 0.5203445888777787,
      "grad_norm": 9.125,
      "learning_rate": 2.2194831166833322e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 2945934.0,
      "step": 6765
    },
    {
      "epoch": 0.5204215060379971,
      "grad_norm": 12.8125,
      "learning_rate": 2.2193677409430045e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2946216.0,
      "step": 6766
    },
    {
      "epoch": 0.5204984231982155,
      "grad_norm": 12.9375,
      "learning_rate": 2.2192523652026767e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 2946758.0,
      "step": 6767
    },
    {
      "epoch": 0.520575340358434,
      "grad_norm": 3.3125,
      "learning_rate": 2.2191369894623494e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2947304.0,
      "step": 6768
    },
    {
      "epoch": 0.5206522575186524,
      "grad_norm": 2.40625,
      "learning_rate": 2.2190216137220213e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 2947862.0,
      "step": 6769
    },
    {
      "epoch": 0.5207291746788708,
      "grad_norm": 10.0,
      "learning_rate": 2.2189062379816935e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2948176.0,
      "step": 6770
    },
    {
      "epoch": 0.5208060918390893,
      "grad_norm": 16.0,
      "learning_rate": 2.218790862241366e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 2948496.0,
      "step": 6771
    },
    {
      "epoch": 0.5208830089993077,
      "grad_norm": 7.46875,
      "learning_rate": 2.2186754865010385e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 2949032.0,
      "step": 6772
    },
    {
      "epoch": 0.5209599261595262,
      "grad_norm": 15.0,
      "learning_rate": 2.2185601107607107e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2949316.0,
      "step": 6773
    },
    {
      "epoch": 0.5210368433197446,
      "grad_norm": 9.0625,
      "learning_rate": 2.2184447350203834e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 2949626.0,
      "step": 6774
    },
    {
      "epoch": 0.521113760479963,
      "grad_norm": 2.65625,
      "learning_rate": 2.2183293592800552e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2950172.0,
      "step": 6775
    },
    {
      "epoch": 0.5211906776401816,
      "grad_norm": 18.25,
      "learning_rate": 2.218213983539728e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2950736.0,
      "step": 6776
    },
    {
      "epoch": 0.5212675948004,
      "grad_norm": 11.9375,
      "learning_rate": 2.2180986077994e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 2951556.0,
      "step": 6777
    },
    {
      "epoch": 0.5213445119606184,
      "grad_norm": 8.5625,
      "learning_rate": 2.2179832320590724e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 2951862.0,
      "step": 6778
    },
    {
      "epoch": 0.5214214291208369,
      "grad_norm": 7.65625,
      "learning_rate": 2.2178678563187447e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2952144.0,
      "step": 6779
    },
    {
      "epoch": 0.5214983462810553,
      "grad_norm": 7.59375,
      "learning_rate": 2.217752480578417e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 2952456.0,
      "step": 6780
    },
    {
      "epoch": 0.5215752634412737,
      "grad_norm": 14.625,
      "learning_rate": 2.217637104838089e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 2953252.0,
      "step": 6781
    },
    {
      "epoch": 0.5216521806014922,
      "grad_norm": 22.0,
      "learning_rate": 2.217521729097762e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2953554.0,
      "step": 6782
    },
    {
      "epoch": 0.5217290977617106,
      "grad_norm": 9.875,
      "learning_rate": 2.217406353357434e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2954080.0,
      "step": 6783
    },
    {
      "epoch": 0.5218060149219291,
      "grad_norm": 16.5,
      "learning_rate": 2.2172909776171064e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9845201373100281,
      "num_tokens": 2954404.0,
      "step": 6784
    },
    {
      "epoch": 0.5218829320821475,
      "grad_norm": 16.75,
      "learning_rate": 2.217175601876779e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2954706.0,
      "step": 6785
    },
    {
      "epoch": 0.5219598492423659,
      "grad_norm": 3.28125,
      "learning_rate": 2.217060226136451e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 2955230.0,
      "step": 6786
    },
    {
      "epoch": 0.5220367664025845,
      "grad_norm": 7.0,
      "learning_rate": 2.216944850396123e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 2955552.0,
      "step": 6787
    },
    {
      "epoch": 0.5221136835628029,
      "grad_norm": 9.0,
      "learning_rate": 2.216829474655796e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2956116.0,
      "step": 6788
    },
    {
      "epoch": 0.5221906007230213,
      "grad_norm": 3.828125,
      "learning_rate": 2.216714098915468e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 2956702.0,
      "step": 6789
    },
    {
      "epoch": 0.5222675178832398,
      "grad_norm": 4.15625,
      "learning_rate": 2.2165987231751404e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 2957248.0,
      "step": 6790
    },
    {
      "epoch": 0.5223444350434582,
      "grad_norm": 13.6875,
      "learning_rate": 2.2164833474348126e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 2957588.0,
      "step": 6791
    },
    {
      "epoch": 0.5224213522036767,
      "grad_norm": 2.578125,
      "learning_rate": 2.216367971694485e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 2958396.0,
      "step": 6792
    },
    {
      "epoch": 0.5224982693638951,
      "grad_norm": 3.890625,
      "learning_rate": 2.2162525959541576e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9936061501502991,
      "num_tokens": 2959180.0,
      "step": 6793
    },
    {
      "epoch": 0.5225751865241135,
      "grad_norm": 4.78125,
      "learning_rate": 2.21613722021383e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 2959532.0,
      "step": 6794
    },
    {
      "epoch": 0.522652103684332,
      "grad_norm": 12.4375,
      "learning_rate": 2.216021844473502e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2959856.0,
      "step": 6795
    },
    {
      "epoch": 0.5227290208445504,
      "grad_norm": 3.875,
      "learning_rate": 2.2159064687331743e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 2960374.0,
      "step": 6796
    },
    {
      "epoch": 0.5228059380047688,
      "grad_norm": 16.25,
      "learning_rate": 2.2157910929928466e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 2960684.0,
      "step": 6797
    },
    {
      "epoch": 0.5228828551649873,
      "grad_norm": 3.3125,
      "learning_rate": 2.215675717252519e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 2961208.0,
      "step": 6798
    },
    {
      "epoch": 0.5229597723252057,
      "grad_norm": 23.75,
      "learning_rate": 2.2155603415121915e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 2961512.0,
      "step": 6799
    },
    {
      "epoch": 0.5230366894854241,
      "grad_norm": 12.125,
      "learning_rate": 2.215444965771864e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 2961832.0,
      "step": 6800
    },
    {
      "epoch": 0.5231136066456427,
      "grad_norm": 9.125,
      "learning_rate": 2.215329590031536e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 2962086.0,
      "step": 6801
    },
    {
      "epoch": 0.5231905238058611,
      "grad_norm": 6.65625,
      "learning_rate": 2.2152142142912083e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 2962612.0,
      "step": 6802
    },
    {
      "epoch": 0.5232674409660796,
      "grad_norm": 5.375,
      "learning_rate": 2.2150988385508806e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 2963188.0,
      "step": 6803
    },
    {
      "epoch": 0.523344358126298,
      "grad_norm": 3.40625,
      "learning_rate": 2.2149834628105533e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 2963756.0,
      "step": 6804
    },
    {
      "epoch": 0.5234212752865164,
      "grad_norm": 14.375,
      "learning_rate": 2.2148680870702255e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2964082.0,
      "step": 6805
    },
    {
      "epoch": 0.5234981924467349,
      "grad_norm": 4.96875,
      "learning_rate": 2.2147527113298978e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 2964598.0,
      "step": 6806
    },
    {
      "epoch": 0.5235751096069533,
      "grad_norm": 7.6875,
      "learning_rate": 2.21463733558957e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 2965386.0,
      "step": 6807
    },
    {
      "epoch": 0.5236520267671717,
      "grad_norm": 15.9375,
      "learning_rate": 2.2145219598492423e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 2965686.0,
      "step": 6808
    },
    {
      "epoch": 0.5237289439273902,
      "grad_norm": 2.71875,
      "learning_rate": 2.2144065841089146e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9974358677864075,
      "num_tokens": 2966468.0,
      "step": 6809
    },
    {
      "epoch": 0.5238058610876086,
      "grad_norm": 2.234375,
      "learning_rate": 2.2142912083685872e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9975247383117676,
      "num_tokens": 2967278.0,
      "step": 6810
    },
    {
      "epoch": 0.5238827782478271,
      "grad_norm": 8.0625,
      "learning_rate": 2.2141758326282595e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2967822.0,
      "step": 6811
    },
    {
      "epoch": 0.5239596954080455,
      "grad_norm": 4.90625,
      "learning_rate": 2.2140604568879317e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 2968382.0,
      "step": 6812
    },
    {
      "epoch": 0.524036612568264,
      "grad_norm": 16.375,
      "learning_rate": 2.213945081147604e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 2968698.0,
      "step": 6813
    },
    {
      "epoch": 0.5241135297284825,
      "grad_norm": 6.21875,
      "learning_rate": 2.2138297054072763e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 2969502.0,
      "step": 6814
    },
    {
      "epoch": 0.5241904468887009,
      "grad_norm": 20.375,
      "learning_rate": 2.213714329666949e-06,
      "loss": 0.0419,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 2969784.0,
      "step": 6815
    },
    {
      "epoch": 0.5242673640489193,
      "grad_norm": 13.75,
      "learning_rate": 2.213598953926621e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9824561476707458,
      "num_tokens": 2970070.0,
      "step": 6816
    },
    {
      "epoch": 0.5243442812091378,
      "grad_norm": 11.375,
      "learning_rate": 2.2134835781862935e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 2970388.0,
      "step": 6817
    },
    {
      "epoch": 0.5244211983693562,
      "grad_norm": 7.375,
      "learning_rate": 2.2133682024459657e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2970906.0,
      "step": 6818
    },
    {
      "epoch": 0.5244981155295746,
      "grad_norm": 11.1875,
      "learning_rate": 2.213252826705638e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 2971222.0,
      "step": 6819
    },
    {
      "epoch": 0.5245750326897931,
      "grad_norm": 8.875,
      "learning_rate": 2.2131374509653102e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 2971556.0,
      "step": 6820
    },
    {
      "epoch": 0.5246519498500115,
      "grad_norm": 2.5,
      "learning_rate": 2.213022075224983e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 2972100.0,
      "step": 6821
    },
    {
      "epoch": 0.52472886701023,
      "grad_norm": 12.625,
      "learning_rate": 2.212906699484655e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 2972392.0,
      "step": 6822
    },
    {
      "epoch": 0.5248057841704484,
      "grad_norm": 6.0,
      "learning_rate": 2.2127913237443274e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 2973202.0,
      "step": 6823
    },
    {
      "epoch": 0.5248827013306668,
      "grad_norm": 6.25,
      "learning_rate": 2.2126759480039997e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 2973512.0,
      "step": 6824
    },
    {
      "epoch": 0.5249596184908853,
      "grad_norm": 21.375,
      "learning_rate": 2.212560572263672e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 2973872.0,
      "step": 6825
    },
    {
      "epoch": 0.5250365356511038,
      "grad_norm": 4.875,
      "learning_rate": 2.2124451965233446e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2974160.0,
      "step": 6826
    },
    {
      "epoch": 0.5251134528113222,
      "grad_norm": 4.0,
      "learning_rate": 2.212329820783017e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 2974706.0,
      "step": 6827
    },
    {
      "epoch": 0.5251903699715407,
      "grad_norm": 4.65625,
      "learning_rate": 2.212214445042689e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2975252.0,
      "step": 6828
    },
    {
      "epoch": 0.5252672871317591,
      "grad_norm": 3.625,
      "learning_rate": 2.2120990693023614e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 2975792.0,
      "step": 6829
    },
    {
      "epoch": 0.5253442042919776,
      "grad_norm": 12.9375,
      "learning_rate": 2.2119836935620337e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 2976378.0,
      "step": 6830
    },
    {
      "epoch": 0.525421121452196,
      "grad_norm": 10.5,
      "learning_rate": 2.211868317821706e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 2976942.0,
      "step": 6831
    },
    {
      "epoch": 0.5254980386124144,
      "grad_norm": 10.75,
      "learning_rate": 2.2117529420813786e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 2977224.0,
      "step": 6832
    },
    {
      "epoch": 0.5255749557726329,
      "grad_norm": 3.578125,
      "learning_rate": 2.211637566341051e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9973545074462891,
      "num_tokens": 2977982.0,
      "step": 6833
    },
    {
      "epoch": 0.5256518729328513,
      "grad_norm": 4.125,
      "learning_rate": 2.211522190600723e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 2978526.0,
      "step": 6834
    },
    {
      "epoch": 0.5257287900930697,
      "grad_norm": 6.90625,
      "learning_rate": 2.2114068148603954e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 2978852.0,
      "step": 6835
    },
    {
      "epoch": 0.5258057072532882,
      "grad_norm": 19.375,
      "learning_rate": 2.2112914391200676e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 2979144.0,
      "step": 6836
    },
    {
      "epoch": 0.5258826244135066,
      "grad_norm": 3.265625,
      "learning_rate": 2.21117606337974e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9942857027053833,
      "num_tokens": 2979670.0,
      "step": 6837
    },
    {
      "epoch": 0.5259595415737252,
      "grad_norm": 2.890625,
      "learning_rate": 2.2110606876394126e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2980188.0,
      "step": 6838
    },
    {
      "epoch": 0.5260364587339436,
      "grad_norm": 6.78125,
      "learning_rate": 2.210945311899085e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 2980516.0,
      "step": 6839
    },
    {
      "epoch": 0.526113375894162,
      "grad_norm": 12.0625,
      "learning_rate": 2.210829936158757e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2980822.0,
      "step": 6840
    },
    {
      "epoch": 0.5261902930543805,
      "grad_norm": 6.6875,
      "learning_rate": 2.2107145604184294e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 2981160.0,
      "step": 6841
    },
    {
      "epoch": 0.5262672102145989,
      "grad_norm": 15.0625,
      "learning_rate": 2.2105991846781016e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 2981464.0,
      "step": 6842
    },
    {
      "epoch": 0.5263441273748173,
      "grad_norm": 7.15625,
      "learning_rate": 2.2104838089377743e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 2981812.0,
      "step": 6843
    },
    {
      "epoch": 0.5264210445350358,
      "grad_norm": 3.734375,
      "learning_rate": 2.2103684331974465e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 2982596.0,
      "step": 6844
    },
    {
      "epoch": 0.5264979616952542,
      "grad_norm": 5.25,
      "learning_rate": 2.2102530574571184e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 2982908.0,
      "step": 6845
    },
    {
      "epoch": 0.5265748788554726,
      "grad_norm": 6.5625,
      "learning_rate": 2.210137681716791e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 2983472.0,
      "step": 6846
    },
    {
      "epoch": 0.5266517960156911,
      "grad_norm": 16.5,
      "learning_rate": 2.2100223059764633e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 2983802.0,
      "step": 6847
    },
    {
      "epoch": 0.5267287131759095,
      "grad_norm": 13.1875,
      "learning_rate": 2.2099069302361356e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 2984086.0,
      "step": 6848
    },
    {
      "epoch": 0.526805630336128,
      "grad_norm": 21.25,
      "learning_rate": 2.2097915544958083e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 2984448.0,
      "step": 6849
    },
    {
      "epoch": 0.5268825474963464,
      "grad_norm": 19.375,
      "learning_rate": 2.2096761787554805e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 2984752.0,
      "step": 6850
    },
    {
      "epoch": 0.5269594646565648,
      "grad_norm": 3.59375,
      "learning_rate": 2.2095608030151528e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9971292018890381,
      "num_tokens": 2985800.0,
      "step": 6851
    },
    {
      "epoch": 0.5270363818167834,
      "grad_norm": 4.3125,
      "learning_rate": 2.209445427274825e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 2986320.0,
      "step": 6852
    },
    {
      "epoch": 0.5271132989770018,
      "grad_norm": 29.0,
      "learning_rate": 2.2093300515344973e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 2986622.0,
      "step": 6853
    },
    {
      "epoch": 0.5271902161372202,
      "grad_norm": 20.5,
      "learning_rate": 2.20921467579417e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 2986946.0,
      "step": 6854
    },
    {
      "epoch": 0.5272671332974387,
      "grad_norm": 2.296875,
      "learning_rate": 2.2090993000538422e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 2987726.0,
      "step": 6855
    },
    {
      "epoch": 0.5273440504576571,
      "grad_norm": 7.25,
      "learning_rate": 2.208983924313514e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 2988224.0,
      "step": 6856
    },
    {
      "epoch": 0.5274209676178756,
      "grad_norm": 5.09375,
      "learning_rate": 2.2088685485731868e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 2988512.0,
      "step": 6857
    },
    {
      "epoch": 0.527497884778094,
      "grad_norm": 4.59375,
      "learning_rate": 2.208753172832859e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 2989084.0,
      "step": 6858
    },
    {
      "epoch": 0.5275748019383124,
      "grad_norm": 11.0,
      "learning_rate": 2.2086377970925313e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 2989380.0,
      "step": 6859
    },
    {
      "epoch": 0.5276517190985309,
      "grad_norm": 15.0625,
      "learning_rate": 2.208522421352204e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 2989658.0,
      "step": 6860
    },
    {
      "epoch": 0.5277286362587493,
      "grad_norm": 7.03125,
      "learning_rate": 2.208407045611876e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 2989996.0,
      "step": 6861
    },
    {
      "epoch": 0.5278055534189677,
      "grad_norm": 3.265625,
      "learning_rate": 2.2082916698715485e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 2990542.0,
      "step": 6862
    },
    {
      "epoch": 0.5278824705791862,
      "grad_norm": 5.15625,
      "learning_rate": 2.2081762941312207e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 2990868.0,
      "step": 6863
    },
    {
      "epoch": 0.5279593877394047,
      "grad_norm": 7.25,
      "learning_rate": 2.208060918390893e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 2991182.0,
      "step": 6864
    },
    {
      "epoch": 0.528036304899623,
      "grad_norm": 12.1875,
      "learning_rate": 2.2079455426505657e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 2991488.0,
      "step": 6865
    },
    {
      "epoch": 0.5281132220598416,
      "grad_norm": 6.75,
      "learning_rate": 2.207830166910238e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 2992058.0,
      "step": 6866
    },
    {
      "epoch": 0.52819013922006,
      "grad_norm": 14.0,
      "learning_rate": 2.2077147911699098e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 2992366.0,
      "step": 6867
    },
    {
      "epoch": 0.5282670563802785,
      "grad_norm": 2.703125,
      "learning_rate": 2.2075994154295824e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 2992940.0,
      "step": 6868
    },
    {
      "epoch": 0.5283439735404969,
      "grad_norm": 3.546875,
      "learning_rate": 2.2074840396892547e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 2993502.0,
      "step": 6869
    },
    {
      "epoch": 0.5284208907007153,
      "grad_norm": 11.0,
      "learning_rate": 2.207368663948927e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 2993848.0,
      "step": 6870
    },
    {
      "epoch": 0.5284978078609338,
      "grad_norm": 8.6875,
      "learning_rate": 2.2072532882085996e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 2994652.0,
      "step": 6871
    },
    {
      "epoch": 0.5285747250211522,
      "grad_norm": 9.375,
      "learning_rate": 2.2071379124682715e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 2994988.0,
      "step": 6872
    },
    {
      "epoch": 0.5286516421813706,
      "grad_norm": 16.375,
      "learning_rate": 2.207022536727944e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 2995350.0,
      "step": 6873
    },
    {
      "epoch": 0.5287285593415891,
      "grad_norm": 4.28125,
      "learning_rate": 2.2069071609876164e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 2995906.0,
      "step": 6874
    },
    {
      "epoch": 0.5288054765018075,
      "grad_norm": 4.4375,
      "learning_rate": 2.2067917852472887e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 2996662.0,
      "step": 6875
    },
    {
      "epoch": 0.528882393662026,
      "grad_norm": 5.15625,
      "learning_rate": 2.2066764095069613e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 2997248.0,
      "step": 6876
    },
    {
      "epoch": 0.5289593108222445,
      "grad_norm": 4.96875,
      "learning_rate": 2.2065610337666336e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9915110468864441,
      "num_tokens": 2997838.0,
      "step": 6877
    },
    {
      "epoch": 0.5290362279824629,
      "grad_norm": 19.875,
      "learning_rate": 2.2064456580263054e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 2998120.0,
      "step": 6878
    },
    {
      "epoch": 0.5291131451426814,
      "grad_norm": 7.1875,
      "learning_rate": 2.206330282285978e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 2998638.0,
      "step": 6879
    },
    {
      "epoch": 0.5291900623028998,
      "grad_norm": 3.4375,
      "learning_rate": 2.2062149065456504e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 2999190.0,
      "step": 6880
    },
    {
      "epoch": 0.5292669794631182,
      "grad_norm": 17.25,
      "learning_rate": 2.2060995308053226e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 2999500.0,
      "step": 6881
    },
    {
      "epoch": 0.5293438966233367,
      "grad_norm": 16.5,
      "learning_rate": 2.2059841550649953e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9814241528511047,
      "num_tokens": 2999824.0,
      "step": 6882
    },
    {
      "epoch": 0.5294208137835551,
      "grad_norm": 12.5,
      "learning_rate": 2.205868779324667e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9971988797187805,
      "num_tokens": 3000182.0,
      "step": 6883
    },
    {
      "epoch": 0.5294977309437735,
      "grad_norm": 3.109375,
      "learning_rate": 2.20575340358434e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3000964.0,
      "step": 6884
    },
    {
      "epoch": 0.529574648103992,
      "grad_norm": 11.4375,
      "learning_rate": 2.205638027844012e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9858757257461548,
      "num_tokens": 3001320.0,
      "step": 6885
    },
    {
      "epoch": 0.5296515652642104,
      "grad_norm": 6.0,
      "learning_rate": 2.2055226521036844e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3001670.0,
      "step": 6886
    },
    {
      "epoch": 0.5297284824244289,
      "grad_norm": 11.4375,
      "learning_rate": 2.2054072763633566e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 3001964.0,
      "step": 6887
    },
    {
      "epoch": 0.5298053995846473,
      "grad_norm": 4.78125,
      "learning_rate": 2.2052919006230293e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3002526.0,
      "step": 6888
    },
    {
      "epoch": 0.5298823167448657,
      "grad_norm": 16.0,
      "learning_rate": 2.205176524882701e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 3002866.0,
      "step": 6889
    },
    {
      "epoch": 0.5299592339050843,
      "grad_norm": 3.34375,
      "learning_rate": 2.205061149142374e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 3003650.0,
      "step": 6890
    },
    {
      "epoch": 0.5300361510653027,
      "grad_norm": 5.3125,
      "learning_rate": 2.204945773402046e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3004172.0,
      "step": 6891
    },
    {
      "epoch": 0.5301130682255211,
      "grad_norm": 10.3125,
      "learning_rate": 2.2048303976617183e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9983633160591125,
      "num_tokens": 3004784.0,
      "step": 6892
    },
    {
      "epoch": 0.5301899853857396,
      "grad_norm": 4.875,
      "learning_rate": 2.204715021921391e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9925925731658936,
      "num_tokens": 3005596.0,
      "step": 6893
    },
    {
      "epoch": 0.530266902545958,
      "grad_norm": 9.0,
      "learning_rate": 2.204599646181063e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3005942.0,
      "step": 6894
    },
    {
      "epoch": 0.5303438197061765,
      "grad_norm": 10.125,
      "learning_rate": 2.204484270440735e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 3006230.0,
      "step": 6895
    },
    {
      "epoch": 0.5304207368663949,
      "grad_norm": 5.78125,
      "learning_rate": 2.2043688947004078e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 3006802.0,
      "step": 6896
    },
    {
      "epoch": 0.5304976540266133,
      "grad_norm": 8.375,
      "learning_rate": 2.20425351896008e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 3007118.0,
      "step": 6897
    },
    {
      "epoch": 0.5305745711868318,
      "grad_norm": 18.5,
      "learning_rate": 2.2041381432197523e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3007468.0,
      "step": 6898
    },
    {
      "epoch": 0.5306514883470502,
      "grad_norm": 20.875,
      "learning_rate": 2.2040227674794246e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 3007744.0,
      "step": 6899
    },
    {
      "epoch": 0.5307284055072686,
      "grad_norm": 15.875,
      "learning_rate": 2.203907391739097e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3008090.0,
      "step": 6900
    },
    {
      "epoch": 0.5308053226674871,
      "grad_norm": 7.4375,
      "learning_rate": 2.2037920159987695e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 3008904.0,
      "step": 6901
    },
    {
      "epoch": 0.5308822398277055,
      "grad_norm": 10.6875,
      "learning_rate": 2.2036766402584418e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3009160.0,
      "step": 6902
    },
    {
      "epoch": 0.530959156987924,
      "grad_norm": 8.875,
      "learning_rate": 2.203561264518114e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3009468.0,
      "step": 6903
    },
    {
      "epoch": 0.5310360741481425,
      "grad_norm": 6.3125,
      "learning_rate": 2.2034458887777867e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3010042.0,
      "step": 6904
    },
    {
      "epoch": 0.5311129913083609,
      "grad_norm": 7.625,
      "learning_rate": 2.2033305130374585e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 3010326.0,
      "step": 6905
    },
    {
      "epoch": 0.5311899084685794,
      "grad_norm": 4.59375,
      "learning_rate": 2.2032151372971308e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3010870.0,
      "step": 6906
    },
    {
      "epoch": 0.5312668256287978,
      "grad_norm": 3.046875,
      "learning_rate": 2.2030997615568035e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3011416.0,
      "step": 6907
    },
    {
      "epoch": 0.5313437427890162,
      "grad_norm": 25.75,
      "learning_rate": 2.2029843858164757e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3011694.0,
      "step": 6908
    },
    {
      "epoch": 0.5314206599492347,
      "grad_norm": 3.96875,
      "learning_rate": 2.202869010076148e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 3012478.0,
      "step": 6909
    },
    {
      "epoch": 0.5314975771094531,
      "grad_norm": 9.5625,
      "learning_rate": 2.2027536343358202e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3012828.0,
      "step": 6910
    },
    {
      "epoch": 0.5315744942696715,
      "grad_norm": 11.1875,
      "learning_rate": 2.2026382585954925e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3013140.0,
      "step": 6911
    },
    {
      "epoch": 0.53165141142989,
      "grad_norm": 6.1875,
      "learning_rate": 2.202522882855165e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3013464.0,
      "step": 6912
    },
    {
      "epoch": 0.5317283285901084,
      "grad_norm": 5.53125,
      "learning_rate": 2.2024075071148374e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3014004.0,
      "step": 6913
    },
    {
      "epoch": 0.531805245750327,
      "grad_norm": 6.5625,
      "learning_rate": 2.2022921313745097e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 3014584.0,
      "step": 6914
    },
    {
      "epoch": 0.5318821629105454,
      "grad_norm": 20.0,
      "learning_rate": 2.2021767556341824e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 3014840.0,
      "step": 6915
    },
    {
      "epoch": 0.5319590800707638,
      "grad_norm": 10.625,
      "learning_rate": 2.2020613798938542e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 3015152.0,
      "step": 6916
    },
    {
      "epoch": 0.5320359972309823,
      "grad_norm": 3.59375,
      "learning_rate": 2.2019460041535265e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 3015700.0,
      "step": 6917
    },
    {
      "epoch": 0.5321129143912007,
      "grad_norm": 12.5625,
      "learning_rate": 2.201830628413199e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3016040.0,
      "step": 6918
    },
    {
      "epoch": 0.5321898315514191,
      "grad_norm": 14.375,
      "learning_rate": 2.2017152526728714e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 3016370.0,
      "step": 6919
    },
    {
      "epoch": 0.5322667487116376,
      "grad_norm": 10.4375,
      "learning_rate": 2.2015998769325437e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3016694.0,
      "step": 6920
    },
    {
      "epoch": 0.532343665871856,
      "grad_norm": 12.125,
      "learning_rate": 2.201484501192216e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3017022.0,
      "step": 6921
    },
    {
      "epoch": 0.5324205830320744,
      "grad_norm": 11.6875,
      "learning_rate": 2.201369125451888e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3017338.0,
      "step": 6922
    },
    {
      "epoch": 0.5324975001922929,
      "grad_norm": 15.125,
      "learning_rate": 2.201253749711561e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3017620.0,
      "step": 6923
    },
    {
      "epoch": 0.5325744173525113,
      "grad_norm": 4.59375,
      "learning_rate": 2.201138373971233e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3018188.0,
      "step": 6924
    },
    {
      "epoch": 0.5326513345127298,
      "grad_norm": 15.0625,
      "learning_rate": 2.2010229982309054e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 3018498.0,
      "step": 6925
    },
    {
      "epoch": 0.5327282516729482,
      "grad_norm": 3.703125,
      "learning_rate": 2.200907622490578e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 3018992.0,
      "step": 6926
    },
    {
      "epoch": 0.5328051688331666,
      "grad_norm": 11.0,
      "learning_rate": 2.20079224675025e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 3019748.0,
      "step": 6927
    },
    {
      "epoch": 0.5328820859933852,
      "grad_norm": 11.0625,
      "learning_rate": 2.200676871009922e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3020008.0,
      "step": 6928
    },
    {
      "epoch": 0.5329590031536036,
      "grad_norm": 20.125,
      "learning_rate": 2.200561495269595e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3020330.0,
      "step": 6929
    },
    {
      "epoch": 0.533035920313822,
      "grad_norm": 8.5625,
      "learning_rate": 2.200446119529267e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 3020906.0,
      "step": 6930
    },
    {
      "epoch": 0.5331128374740405,
      "grad_norm": 19.0,
      "learning_rate": 2.2003307437889394e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3021162.0,
      "step": 6931
    },
    {
      "epoch": 0.5331897546342589,
      "grad_norm": 29.0,
      "learning_rate": 2.2002153680486116e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9880478382110596,
      "num_tokens": 3021414.0,
      "step": 6932
    },
    {
      "epoch": 0.5332666717944774,
      "grad_norm": 7.84375,
      "learning_rate": 2.200099992308284e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3021960.0,
      "step": 6933
    },
    {
      "epoch": 0.5333435889546958,
      "grad_norm": 20.125,
      "learning_rate": 2.1999846165679566e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3022278.0,
      "step": 6934
    },
    {
      "epoch": 0.5334205061149142,
      "grad_norm": 11.5625,
      "learning_rate": 2.199869240827629e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3022538.0,
      "step": 6935
    },
    {
      "epoch": 0.5334974232751327,
      "grad_norm": 6.03125,
      "learning_rate": 2.199753865087301e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3023078.0,
      "step": 6936
    },
    {
      "epoch": 0.5335743404353511,
      "grad_norm": 22.875,
      "learning_rate": 2.1996384893469733e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3023414.0,
      "step": 6937
    },
    {
      "epoch": 0.5336512575955695,
      "grad_norm": 10.0,
      "learning_rate": 2.1995231136066456e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 3023730.0,
      "step": 6938
    },
    {
      "epoch": 0.533728174755788,
      "grad_norm": 24.75,
      "learning_rate": 2.199407737866318e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 3024330.0,
      "step": 6939
    },
    {
      "epoch": 0.5338050919160064,
      "grad_norm": 4.625,
      "learning_rate": 2.1992923621259905e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3024852.0,
      "step": 6940
    },
    {
      "epoch": 0.5338820090762249,
      "grad_norm": 8.0625,
      "learning_rate": 2.1991769863856628e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3025196.0,
      "step": 6941
    },
    {
      "epoch": 0.5339589262364434,
      "grad_norm": 7.21875,
      "learning_rate": 2.199061610645335e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3025518.0,
      "step": 6942
    },
    {
      "epoch": 0.5340358433966618,
      "grad_norm": 27.5,
      "learning_rate": 2.1989462349050073e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 3025752.0,
      "step": 6943
    },
    {
      "epoch": 0.5341127605568803,
      "grad_norm": 25.25,
      "learning_rate": 2.1988308591646796e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3026064.0,
      "step": 6944
    },
    {
      "epoch": 0.5341896777170987,
      "grad_norm": 8.625,
      "learning_rate": 2.198715483424352e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3026648.0,
      "step": 6945
    },
    {
      "epoch": 0.5342665948773171,
      "grad_norm": 5.90625,
      "learning_rate": 2.1986001076840245e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 3027198.0,
      "step": 6946
    },
    {
      "epoch": 0.5343435120375356,
      "grad_norm": 4.5625,
      "learning_rate": 2.1984847319436968e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 3027746.0,
      "step": 6947
    },
    {
      "epoch": 0.534420429197754,
      "grad_norm": 20.5,
      "learning_rate": 2.198369356203369e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3028030.0,
      "step": 6948
    },
    {
      "epoch": 0.5344973463579724,
      "grad_norm": 18.75,
      "learning_rate": 2.1982539804630413e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9928315281867981,
      "num_tokens": 3028310.0,
      "step": 6949
    },
    {
      "epoch": 0.5345742635181909,
      "grad_norm": 25.0,
      "learning_rate": 2.1981386047227135e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3028566.0,
      "step": 6950
    },
    {
      "epoch": 0.5346511806784093,
      "grad_norm": 36.25,
      "learning_rate": 2.1980232289823862e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 3028870.0,
      "step": 6951
    },
    {
      "epoch": 0.5347280978386278,
      "grad_norm": 8.5,
      "learning_rate": 2.1979078532420585e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3029216.0,
      "step": 6952
    },
    {
      "epoch": 0.5348050149988463,
      "grad_norm": 20.875,
      "learning_rate": 2.1977924775017307e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9845201373100281,
      "num_tokens": 3029540.0,
      "step": 6953
    },
    {
      "epoch": 0.5348819321590647,
      "grad_norm": 6.3125,
      "learning_rate": 2.197677101761403e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 3029856.0,
      "step": 6954
    },
    {
      "epoch": 0.5349588493192832,
      "grad_norm": 21.0,
      "learning_rate": 2.1975617260210752e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3030418.0,
      "step": 6955
    },
    {
      "epoch": 0.5350357664795016,
      "grad_norm": 18.375,
      "learning_rate": 2.1974463502807475e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 3031004.0,
      "step": 6956
    },
    {
      "epoch": 0.53511268363972,
      "grad_norm": 13.25,
      "learning_rate": 2.19733097454042e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3031784.0,
      "step": 6957
    },
    {
      "epoch": 0.5351896007999385,
      "grad_norm": 10.25,
      "learning_rate": 2.1972155988000924e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3032324.0,
      "step": 6958
    },
    {
      "epoch": 0.5352665179601569,
      "grad_norm": 10.75,
      "learning_rate": 2.1971002230597647e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 3032638.0,
      "step": 6959
    },
    {
      "epoch": 0.5353434351203754,
      "grad_norm": 11.375,
      "learning_rate": 2.196984847319437e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3032956.0,
      "step": 6960
    },
    {
      "epoch": 0.5354203522805938,
      "grad_norm": 8.625,
      "learning_rate": 2.1968694715791092e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3033240.0,
      "step": 6961
    },
    {
      "epoch": 0.5354972694408122,
      "grad_norm": 11.5,
      "learning_rate": 2.196754095838782e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 3033532.0,
      "step": 6962
    },
    {
      "epoch": 0.5355741866010307,
      "grad_norm": 12.625,
      "learning_rate": 2.196638720098454e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3033854.0,
      "step": 6963
    },
    {
      "epoch": 0.5356511037612491,
      "grad_norm": 4.3125,
      "learning_rate": 2.1965233443581264e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3034434.0,
      "step": 6964
    },
    {
      "epoch": 0.5357280209214675,
      "grad_norm": 8.1875,
      "learning_rate": 2.1964079686177987e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3035002.0,
      "step": 6965
    },
    {
      "epoch": 0.535804938081686,
      "grad_norm": 8.375,
      "learning_rate": 2.196292592877471e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 3035338.0,
      "step": 6966
    },
    {
      "epoch": 0.5358818552419045,
      "grad_norm": 8.125,
      "learning_rate": 2.196177217137143e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3035662.0,
      "step": 6967
    },
    {
      "epoch": 0.5359587724021229,
      "grad_norm": 3.046875,
      "learning_rate": 2.196061841396816e-06,
      "loss": 0.0074,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 3036682.0,
      "step": 6968
    },
    {
      "epoch": 0.5360356895623414,
      "grad_norm": 9.125,
      "learning_rate": 2.195946465656488e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3036978.0,
      "step": 6969
    },
    {
      "epoch": 0.5361126067225598,
      "grad_norm": 23.625,
      "learning_rate": 2.1958310899161604e-06,
      "loss": 0.0401,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3037306.0,
      "step": 6970
    },
    {
      "epoch": 0.5361895238827783,
      "grad_norm": 16.875,
      "learning_rate": 2.1957157141758326e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 3037598.0,
      "step": 6971
    },
    {
      "epoch": 0.5362664410429967,
      "grad_norm": 6.75,
      "learning_rate": 2.195600338435505e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3038178.0,
      "step": 6972
    },
    {
      "epoch": 0.5363433582032151,
      "grad_norm": 7.46875,
      "learning_rate": 2.1954849626951776e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 3038994.0,
      "step": 6973
    },
    {
      "epoch": 0.5364202753634336,
      "grad_norm": 5.96875,
      "learning_rate": 2.19536958695485e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3039516.0,
      "step": 6974
    },
    {
      "epoch": 0.536497192523652,
      "grad_norm": 12.125,
      "learning_rate": 2.1952542112145217e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3039822.0,
      "step": 6975
    },
    {
      "epoch": 0.5365741096838704,
      "grad_norm": 12.9375,
      "learning_rate": 2.1951388354741944e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 3040348.0,
      "step": 6976
    },
    {
      "epoch": 0.5366510268440889,
      "grad_norm": 10.4375,
      "learning_rate": 2.1950234597338666e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 3041158.0,
      "step": 6977
    },
    {
      "epoch": 0.5367279440043073,
      "grad_norm": 8.0625,
      "learning_rate": 2.194908083993539e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3041468.0,
      "step": 6978
    },
    {
      "epoch": 0.5368048611645259,
      "grad_norm": 10.875,
      "learning_rate": 2.1947927082532116e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3041804.0,
      "step": 6979
    },
    {
      "epoch": 0.5368817783247443,
      "grad_norm": 19.5,
      "learning_rate": 2.194677332512884e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3042130.0,
      "step": 6980
    },
    {
      "epoch": 0.5369586954849627,
      "grad_norm": 10.25,
      "learning_rate": 2.194561956772556e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3042430.0,
      "step": 6981
    },
    {
      "epoch": 0.5370356126451812,
      "grad_norm": 37.25,
      "learning_rate": 2.1944465810322283e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 3042762.0,
      "step": 6982
    },
    {
      "epoch": 0.5371125298053996,
      "grad_norm": 9.75,
      "learning_rate": 2.1943312052919006e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3043048.0,
      "step": 6983
    },
    {
      "epoch": 0.537189446965618,
      "grad_norm": 26.375,
      "learning_rate": 2.1942158295515733e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 3043376.0,
      "step": 6984
    },
    {
      "epoch": 0.5372663641258365,
      "grad_norm": 20.0,
      "learning_rate": 2.1941004538112455e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3043910.0,
      "step": 6985
    },
    {
      "epoch": 0.5373432812860549,
      "grad_norm": 7.5625,
      "learning_rate": 2.1939850780709174e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 3044240.0,
      "step": 6986
    },
    {
      "epoch": 0.5374201984462733,
      "grad_norm": 27.875,
      "learning_rate": 2.19386970233059e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 3045076.0,
      "step": 6987
    },
    {
      "epoch": 0.5374971156064918,
      "grad_norm": 20.625,
      "learning_rate": 2.1937543265902623e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3045376.0,
      "step": 6988
    },
    {
      "epoch": 0.5375740327667102,
      "grad_norm": 7.9375,
      "learning_rate": 2.1936389508499346e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9967948794364929,
      "num_tokens": 3045690.0,
      "step": 6989
    },
    {
      "epoch": 0.5376509499269287,
      "grad_norm": 27.75,
      "learning_rate": 2.1935235751096072e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3045946.0,
      "step": 6990
    },
    {
      "epoch": 0.5377278670871471,
      "grad_norm": 44.0,
      "learning_rate": 2.1934081993692795e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3046248.0,
      "step": 6991
    },
    {
      "epoch": 0.5378047842473656,
      "grad_norm": 5.1875,
      "learning_rate": 2.1932928236289518e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3046792.0,
      "step": 6992
    },
    {
      "epoch": 0.5378817014075841,
      "grad_norm": 31.125,
      "learning_rate": 2.193177447888624e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3047098.0,
      "step": 6993
    },
    {
      "epoch": 0.5379586185678025,
      "grad_norm": 39.75,
      "learning_rate": 2.1930620721482963e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 3047710.0,
      "step": 6994
    },
    {
      "epoch": 0.5380355357280209,
      "grad_norm": 16.25,
      "learning_rate": 2.1929466964079685e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 3048270.0,
      "step": 6995
    },
    {
      "epoch": 0.5381124528882394,
      "grad_norm": 9.625,
      "learning_rate": 2.1928313206676412e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 3049078.0,
      "step": 6996
    },
    {
      "epoch": 0.5381893700484578,
      "grad_norm": 15.9375,
      "learning_rate": 2.192715944927313e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 3049904.0,
      "step": 6997
    },
    {
      "epoch": 0.5382662872086763,
      "grad_norm": 26.75,
      "learning_rate": 2.1926005691869857e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 3050156.0,
      "step": 6998
    },
    {
      "epoch": 0.5383432043688947,
      "grad_norm": 5.4375,
      "learning_rate": 2.192485193446658e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3050678.0,
      "step": 6999
    },
    {
      "epoch": 0.5384201215291131,
      "grad_norm": 2.5,
      "learning_rate": 2.1923698177063302e-06,
      "loss": 0.0078,
      "step": 7000
    },
    {
      "epoch": 0.5384201215291131,
      "eval_loss": 0.016563162207603455,
      "eval_mean_token_accuracy": 0.9940684390068054,
      "eval_num_tokens": 3051460.0,
      "eval_runtime": 21.0943,
      "eval_samples_per_second": 9.481,
      "eval_steps_per_second": 1.185,
      "step": 7000
    },
    {
      "epoch": 0.5384970386893316,
      "grad_norm": 6.5,
      "learning_rate": 2.192254441966003e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9933747351169586,
      "num_tokens": 3051780.0,
      "step": 7001
    },
    {
      "epoch": 0.53857395584955,
      "grad_norm": 44.25,
      "learning_rate": 2.1921390662256748e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3052120.0,
      "step": 7002
    },
    {
      "epoch": 0.5386508730097684,
      "grad_norm": 47.0,
      "learning_rate": 2.192023690485347e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 3052424.0,
      "step": 7003
    },
    {
      "epoch": 0.538727790169987,
      "grad_norm": 37.75,
      "learning_rate": 2.1919083147450197e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 3052728.0,
      "step": 7004
    },
    {
      "epoch": 0.5388047073302054,
      "grad_norm": 18.375,
      "learning_rate": 2.191792939004692e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 3053066.0,
      "step": 7005
    },
    {
      "epoch": 0.5388816244904238,
      "grad_norm": 4.375,
      "learning_rate": 2.1916775632643642e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3053878.0,
      "step": 7006
    },
    {
      "epoch": 0.5389585416506423,
      "grad_norm": 20.875,
      "learning_rate": 2.191562187524037e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 3054182.0,
      "step": 7007
    },
    {
      "epoch": 0.5390354588108607,
      "grad_norm": 17.875,
      "learning_rate": 2.1914468117837087e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 3055002.0,
      "step": 7008
    },
    {
      "epoch": 0.5391123759710792,
      "grad_norm": 12.0625,
      "learning_rate": 2.1913314360433814e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 3055256.0,
      "step": 7009
    },
    {
      "epoch": 0.5391892931312976,
      "grad_norm": 6.84375,
      "learning_rate": 2.1912160603030537e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3055824.0,
      "step": 7010
    },
    {
      "epoch": 0.539266210291516,
      "grad_norm": 32.75,
      "learning_rate": 2.191100684562726e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3056118.0,
      "step": 7011
    },
    {
      "epoch": 0.5393431274517345,
      "grad_norm": 14.25,
      "learning_rate": 2.1909853088223986e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3056422.0,
      "step": 7012
    },
    {
      "epoch": 0.5394200446119529,
      "grad_norm": 24.375,
      "learning_rate": 2.1908699330820705e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3056762.0,
      "step": 7013
    },
    {
      "epoch": 0.5394969617721713,
      "grad_norm": 23.125,
      "learning_rate": 2.1907545573417427e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3057086.0,
      "step": 7014
    },
    {
      "epoch": 0.5395738789323898,
      "grad_norm": 8.0625,
      "learning_rate": 2.1906391816014154e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3057636.0,
      "step": 7015
    },
    {
      "epoch": 0.5396507960926082,
      "grad_norm": 17.0,
      "learning_rate": 2.1905238058610876e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 3058168.0,
      "step": 7016
    },
    {
      "epoch": 0.5397277132528268,
      "grad_norm": 5.40625,
      "learning_rate": 2.19040843012076e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3058714.0,
      "step": 7017
    },
    {
      "epoch": 0.5398046304130452,
      "grad_norm": 12.8125,
      "learning_rate": 2.1902930543804326e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 3059304.0,
      "step": 7018
    },
    {
      "epoch": 0.5398815475732636,
      "grad_norm": 15.6875,
      "learning_rate": 2.1901776786401044e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3059844.0,
      "step": 7019
    },
    {
      "epoch": 0.5399584647334821,
      "grad_norm": 9.625,
      "learning_rate": 2.190062302899777e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3060144.0,
      "step": 7020
    },
    {
      "epoch": 0.5400353818937005,
      "grad_norm": 4.53125,
      "learning_rate": 2.1899469271594494e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3060696.0,
      "step": 7021
    },
    {
      "epoch": 0.5401122990539189,
      "grad_norm": 9.875,
      "learning_rate": 2.1898315514191216e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9981446862220764,
      "num_tokens": 3061236.0,
      "step": 7022
    },
    {
      "epoch": 0.5401892162141374,
      "grad_norm": 12.8125,
      "learning_rate": 2.1897161756787943e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3061546.0,
      "step": 7023
    },
    {
      "epoch": 0.5402661333743558,
      "grad_norm": 6.53125,
      "learning_rate": 2.189600799938466e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3062116.0,
      "step": 7024
    },
    {
      "epoch": 0.5403430505345742,
      "grad_norm": 20.875,
      "learning_rate": 2.1894854241981384e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3062396.0,
      "step": 7025
    },
    {
      "epoch": 0.5404199676947927,
      "grad_norm": 3.375,
      "learning_rate": 2.189370048457811e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3062952.0,
      "step": 7026
    },
    {
      "epoch": 0.5404968848550111,
      "grad_norm": 7.65625,
      "learning_rate": 2.1892546727174833e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3063302.0,
      "step": 7027
    },
    {
      "epoch": 0.5405738020152296,
      "grad_norm": 10.625,
      "learning_rate": 2.1891392969771556e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 3063628.0,
      "step": 7028
    },
    {
      "epoch": 0.540650719175448,
      "grad_norm": 5.21875,
      "learning_rate": 2.1890239212368283e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3063978.0,
      "step": 7029
    },
    {
      "epoch": 0.5407276363356665,
      "grad_norm": 4.78125,
      "learning_rate": 2.1889085454965e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3064518.0,
      "step": 7030
    },
    {
      "epoch": 0.540804553495885,
      "grad_norm": 8.75,
      "learning_rate": 2.188793169756173e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3065066.0,
      "step": 7031
    },
    {
      "epoch": 0.5408814706561034,
      "grad_norm": 17.625,
      "learning_rate": 2.188677794015845e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3065346.0,
      "step": 7032
    },
    {
      "epoch": 0.5409583878163218,
      "grad_norm": 16.0,
      "learning_rate": 2.1885624182755173e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3065654.0,
      "step": 7033
    },
    {
      "epoch": 0.5410353049765403,
      "grad_norm": 14.0625,
      "learning_rate": 2.18844704253519e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 3065930.0,
      "step": 7034
    },
    {
      "epoch": 0.5411122221367587,
      "grad_norm": 19.5,
      "learning_rate": 2.188331666794862e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3066520.0,
      "step": 7035
    },
    {
      "epoch": 0.5411891392969772,
      "grad_norm": 16.125,
      "learning_rate": 2.188216291054534e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3066840.0,
      "step": 7036
    },
    {
      "epoch": 0.5412660564571956,
      "grad_norm": 17.25,
      "learning_rate": 2.1881009153142068e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3067158.0,
      "step": 7037
    },
    {
      "epoch": 0.541342973617414,
      "grad_norm": 6.0,
      "learning_rate": 2.187985539573879e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 3067462.0,
      "step": 7038
    },
    {
      "epoch": 0.5414198907776325,
      "grad_norm": 5.6875,
      "learning_rate": 2.1878701638335513e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3067744.0,
      "step": 7039
    },
    {
      "epoch": 0.5414968079378509,
      "grad_norm": 5.1875,
      "learning_rate": 2.1877547880932235e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3068050.0,
      "step": 7040
    },
    {
      "epoch": 0.5415737250980693,
      "grad_norm": 18.375,
      "learning_rate": 2.187639412352896e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3068368.0,
      "step": 7041
    },
    {
      "epoch": 0.5416506422582879,
      "grad_norm": 14.875,
      "learning_rate": 2.1875240366125685e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3068666.0,
      "step": 7042
    },
    {
      "epoch": 0.5417275594185063,
      "grad_norm": 7.96875,
      "learning_rate": 2.1874086608722407e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 3068980.0,
      "step": 7043
    },
    {
      "epoch": 0.5418044765787247,
      "grad_norm": 6.125,
      "learning_rate": 2.187293285131913e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3069292.0,
      "step": 7044
    },
    {
      "epoch": 0.5418813937389432,
      "grad_norm": 7.96875,
      "learning_rate": 2.1871779093915853e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 3070342.0,
      "step": 7045
    },
    {
      "epoch": 0.5419583108991616,
      "grad_norm": 7.40625,
      "learning_rate": 2.1870625336512575e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3070652.0,
      "step": 7046
    },
    {
      "epoch": 0.5420352280593801,
      "grad_norm": 7.5625,
      "learning_rate": 2.1869471579109298e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3070986.0,
      "step": 7047
    },
    {
      "epoch": 0.5421121452195985,
      "grad_norm": 7.34375,
      "learning_rate": 2.1868317821706024e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 3071290.0,
      "step": 7048
    },
    {
      "epoch": 0.5421890623798169,
      "grad_norm": 9.1875,
      "learning_rate": 2.1867164064302747e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 3072078.0,
      "step": 7049
    },
    {
      "epoch": 0.5422659795400354,
      "grad_norm": 5.96875,
      "learning_rate": 2.186601030689947e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 3072416.0,
      "step": 7050
    },
    {
      "epoch": 0.5423428967002538,
      "grad_norm": 6.4375,
      "learning_rate": 2.1864856549496192e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3072960.0,
      "step": 7051
    },
    {
      "epoch": 0.5424198138604722,
      "grad_norm": 8.6875,
      "learning_rate": 2.1863702792092915e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 3073454.0,
      "step": 7052
    },
    {
      "epoch": 0.5424967310206907,
      "grad_norm": 9.125,
      "learning_rate": 2.1862549034689637e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3073736.0,
      "step": 7053
    },
    {
      "epoch": 0.5425736481809091,
      "grad_norm": 13.5,
      "learning_rate": 2.1861395277286364e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 3074094.0,
      "step": 7054
    },
    {
      "epoch": 0.5426505653411277,
      "grad_norm": 5.3125,
      "learning_rate": 2.1860241519883087e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3074616.0,
      "step": 7055
    },
    {
      "epoch": 0.5427274825013461,
      "grad_norm": 11.4375,
      "learning_rate": 2.185908776247981e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3074950.0,
      "step": 7056
    },
    {
      "epoch": 0.5428043996615645,
      "grad_norm": 10.6875,
      "learning_rate": 2.185793400507653e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3075296.0,
      "step": 7057
    },
    {
      "epoch": 0.542881316821783,
      "grad_norm": 13.3125,
      "learning_rate": 2.1856780247673255e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3075590.0,
      "step": 7058
    },
    {
      "epoch": 0.5429582339820014,
      "grad_norm": 12.6875,
      "learning_rate": 2.185562649026998e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3075900.0,
      "step": 7059
    },
    {
      "epoch": 0.5430351511422198,
      "grad_norm": 23.125,
      "learning_rate": 2.1854472732866704e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3076208.0,
      "step": 7060
    },
    {
      "epoch": 0.5431120683024383,
      "grad_norm": 21.0,
      "learning_rate": 2.1853318975463427e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 3076566.0,
      "step": 7061
    },
    {
      "epoch": 0.5431889854626567,
      "grad_norm": 33.5,
      "learning_rate": 2.185216521806015e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3076822.0,
      "step": 7062
    },
    {
      "epoch": 0.5432659026228752,
      "grad_norm": 6.5625,
      "learning_rate": 2.185101146065687e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3077356.0,
      "step": 7063
    },
    {
      "epoch": 0.5433428197830936,
      "grad_norm": 17.25,
      "learning_rate": 2.1849857703253594e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3077928.0,
      "step": 7064
    },
    {
      "epoch": 0.543419736943312,
      "grad_norm": 5.5625,
      "learning_rate": 2.184870394585032e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3078492.0,
      "step": 7065
    },
    {
      "epoch": 0.5434966541035305,
      "grad_norm": 22.0,
      "learning_rate": 2.1847550188447044e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3078824.0,
      "step": 7066
    },
    {
      "epoch": 0.543573571263749,
      "grad_norm": 4.9375,
      "learning_rate": 2.1846396431043766e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3079388.0,
      "step": 7067
    },
    {
      "epoch": 0.5436504884239673,
      "grad_norm": 5.1875,
      "learning_rate": 2.184524267364049e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 3080192.0,
      "step": 7068
    },
    {
      "epoch": 0.5437274055841859,
      "grad_norm": 30.625,
      "learning_rate": 2.184408891623721e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3080472.0,
      "step": 7069
    },
    {
      "epoch": 0.5438043227444043,
      "grad_norm": 4.53125,
      "learning_rate": 2.184293515883394e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3081014.0,
      "step": 7070
    },
    {
      "epoch": 0.5438812399046227,
      "grad_norm": 12.875,
      "learning_rate": 2.184178140143066e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 3081356.0,
      "step": 7071
    },
    {
      "epoch": 0.5439581570648412,
      "grad_norm": 7.875,
      "learning_rate": 2.1840627644027383e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3081680.0,
      "step": 7072
    },
    {
      "epoch": 0.5440350742250596,
      "grad_norm": 3.71875,
      "learning_rate": 2.1839473886624106e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 3082248.0,
      "step": 7073
    },
    {
      "epoch": 0.5441119913852781,
      "grad_norm": 26.75,
      "learning_rate": 2.183832012922083e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 3082586.0,
      "step": 7074
    },
    {
      "epoch": 0.5441889085454965,
      "grad_norm": 15.875,
      "learning_rate": 2.183716637181755e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3082882.0,
      "step": 7075
    },
    {
      "epoch": 0.5442658257057149,
      "grad_norm": 21.875,
      "learning_rate": 2.183601261441428e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3083136.0,
      "step": 7076
    },
    {
      "epoch": 0.5443427428659334,
      "grad_norm": 23.0,
      "learning_rate": 2.1834858857011e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3083416.0,
      "step": 7077
    },
    {
      "epoch": 0.5444196600261518,
      "grad_norm": 7.9375,
      "learning_rate": 2.1833705099607723e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 3083750.0,
      "step": 7078
    },
    {
      "epoch": 0.5444965771863702,
      "grad_norm": 9.375,
      "learning_rate": 2.1832551342204446e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3084306.0,
      "step": 7079
    },
    {
      "epoch": 0.5445734943465887,
      "grad_norm": 17.625,
      "learning_rate": 2.183139758480117e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 3084608.0,
      "step": 7080
    },
    {
      "epoch": 0.5446504115068072,
      "grad_norm": 11.5,
      "learning_rate": 2.1830243827397895e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 3085190.0,
      "step": 7081
    },
    {
      "epoch": 0.5447273286670257,
      "grad_norm": 6.1875,
      "learning_rate": 2.1829090069994618e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3085770.0,
      "step": 7082
    },
    {
      "epoch": 0.5448042458272441,
      "grad_norm": 30.875,
      "learning_rate": 2.182793631259134e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3086070.0,
      "step": 7083
    },
    {
      "epoch": 0.5448811629874625,
      "grad_norm": 5.8125,
      "learning_rate": 2.1826782555188063e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3086394.0,
      "step": 7084
    },
    {
      "epoch": 0.544958080147681,
      "grad_norm": 11.0625,
      "learning_rate": 2.1825628797784785e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 3086706.0,
      "step": 7085
    },
    {
      "epoch": 0.5450349973078994,
      "grad_norm": 6.78125,
      "learning_rate": 2.182447504038151e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 3087278.0,
      "step": 7086
    },
    {
      "epoch": 0.5451119144681178,
      "grad_norm": 6.53125,
      "learning_rate": 2.1823321282978235e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3087850.0,
      "step": 7087
    },
    {
      "epoch": 0.5451888316283363,
      "grad_norm": 3.59375,
      "learning_rate": 2.1822167525574957e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3088368.0,
      "step": 7088
    },
    {
      "epoch": 0.5452657487885547,
      "grad_norm": 30.25,
      "learning_rate": 2.182101376817168e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3088650.0,
      "step": 7089
    },
    {
      "epoch": 0.5453426659487731,
      "grad_norm": 17.125,
      "learning_rate": 2.1819860010768403e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3088958.0,
      "step": 7090
    },
    {
      "epoch": 0.5454195831089916,
      "grad_norm": 8.375,
      "learning_rate": 2.1818706253365125e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 3089274.0,
      "step": 7091
    },
    {
      "epoch": 0.54549650026921,
      "grad_norm": 16.125,
      "learning_rate": 2.181755249596185e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 3089618.0,
      "step": 7092
    },
    {
      "epoch": 0.5455734174294286,
      "grad_norm": 26.0,
      "learning_rate": 2.1816398738558575e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 3089890.0,
      "step": 7093
    },
    {
      "epoch": 0.545650334589647,
      "grad_norm": 7.78125,
      "learning_rate": 2.1815244981155297e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 3090386.0,
      "step": 7094
    },
    {
      "epoch": 0.5457272517498654,
      "grad_norm": 5.25,
      "learning_rate": 2.181409122375202e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3090932.0,
      "step": 7095
    },
    {
      "epoch": 0.5458041689100839,
      "grad_norm": 16.125,
      "learning_rate": 2.1812937466348742e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3091496.0,
      "step": 7096
    },
    {
      "epoch": 0.5458810860703023,
      "grad_norm": 5.09375,
      "learning_rate": 2.1811783708945465e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3092020.0,
      "step": 7097
    },
    {
      "epoch": 0.5459580032305207,
      "grad_norm": 4.375,
      "learning_rate": 2.181062995154219e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3092576.0,
      "step": 7098
    },
    {
      "epoch": 0.5460349203907392,
      "grad_norm": 7.46875,
      "learning_rate": 2.1809476194138914e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3093160.0,
      "step": 7099
    },
    {
      "epoch": 0.5461118375509576,
      "grad_norm": 7.75,
      "learning_rate": 2.1808322436735637e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3093734.0,
      "step": 7100
    },
    {
      "epoch": 0.5461887547111761,
      "grad_norm": 25.875,
      "learning_rate": 2.180716867933236e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3094018.0,
      "step": 7101
    },
    {
      "epoch": 0.5462656718713945,
      "grad_norm": 6.25,
      "learning_rate": 2.180601492192908e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 3094374.0,
      "step": 7102
    },
    {
      "epoch": 0.5463425890316129,
      "grad_norm": 14.0625,
      "learning_rate": 2.180486116452581e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 3094664.0,
      "step": 7103
    },
    {
      "epoch": 0.5464195061918314,
      "grad_norm": 24.5,
      "learning_rate": 2.180370740712253e-06,
      "loss": 0.0464,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 3094942.0,
      "step": 7104
    },
    {
      "epoch": 0.5464964233520498,
      "grad_norm": 20.125,
      "learning_rate": 2.1802553649719254e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9810725450515747,
      "num_tokens": 3095260.0,
      "step": 7105
    },
    {
      "epoch": 0.5465733405122682,
      "grad_norm": 3.046875,
      "learning_rate": 2.1801399892315977e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3095822.0,
      "step": 7106
    },
    {
      "epoch": 0.5466502576724868,
      "grad_norm": 12.625,
      "learning_rate": 2.18002461349127e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3096144.0,
      "step": 7107
    },
    {
      "epoch": 0.5467271748327052,
      "grad_norm": 3.5625,
      "learning_rate": 2.179909237750942e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3096696.0,
      "step": 7108
    },
    {
      "epoch": 0.5468040919929236,
      "grad_norm": 12.0,
      "learning_rate": 2.179793862010615e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3097012.0,
      "step": 7109
    },
    {
      "epoch": 0.5468810091531421,
      "grad_norm": 4.40625,
      "learning_rate": 2.179678486270287e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3097586.0,
      "step": 7110
    },
    {
      "epoch": 0.5469579263133605,
      "grad_norm": 8.25,
      "learning_rate": 2.1795631105299594e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3097936.0,
      "step": 7111
    },
    {
      "epoch": 0.547034843473579,
      "grad_norm": 5.1875,
      "learning_rate": 2.1794477347896316e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3098484.0,
      "step": 7112
    },
    {
      "epoch": 0.5471117606337974,
      "grad_norm": 16.0,
      "learning_rate": 2.179332359049304e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3098766.0,
      "step": 7113
    },
    {
      "epoch": 0.5471886777940158,
      "grad_norm": 15.3125,
      "learning_rate": 2.179216983308976e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3099086.0,
      "step": 7114
    },
    {
      "epoch": 0.5472655949542343,
      "grad_norm": 15.3125,
      "learning_rate": 2.179101607568649e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3099672.0,
      "step": 7115
    },
    {
      "epoch": 0.5473425121144527,
      "grad_norm": 18.875,
      "learning_rate": 2.1789862318283207e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3100000.0,
      "step": 7116
    },
    {
      "epoch": 0.5474194292746711,
      "grad_norm": 17.125,
      "learning_rate": 2.1788708560879933e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3100584.0,
      "step": 7117
    },
    {
      "epoch": 0.5474963464348896,
      "grad_norm": 10.3125,
      "learning_rate": 2.1787554803476656e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 3100894.0,
      "step": 7118
    },
    {
      "epoch": 0.547573263595108,
      "grad_norm": 6.34375,
      "learning_rate": 2.178640104607338e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3101438.0,
      "step": 7119
    },
    {
      "epoch": 0.5476501807553266,
      "grad_norm": 9.9375,
      "learning_rate": 2.1785247288670105e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 3101696.0,
      "step": 7120
    },
    {
      "epoch": 0.547727097915545,
      "grad_norm": 16.0,
      "learning_rate": 2.178409353126683e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3102252.0,
      "step": 7121
    },
    {
      "epoch": 0.5478040150757634,
      "grad_norm": 22.25,
      "learning_rate": 2.1782939773863546e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3102552.0,
      "step": 7122
    },
    {
      "epoch": 0.5478809322359819,
      "grad_norm": 6.9375,
      "learning_rate": 2.1781786016460273e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3102836.0,
      "step": 7123
    },
    {
      "epoch": 0.5479578493962003,
      "grad_norm": 15.0,
      "learning_rate": 2.1780632259056996e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3103180.0,
      "step": 7124
    },
    {
      "epoch": 0.5480347665564187,
      "grad_norm": 22.5,
      "learning_rate": 2.177947850165372e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 3103776.0,
      "step": 7125
    },
    {
      "epoch": 0.5481116837166372,
      "grad_norm": 18.0,
      "learning_rate": 2.1778324744250445e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3104128.0,
      "step": 7126
    },
    {
      "epoch": 0.5481886008768556,
      "grad_norm": 2.5,
      "learning_rate": 2.1777170986847163e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3104674.0,
      "step": 7127
    },
    {
      "epoch": 0.548265518037074,
      "grad_norm": 7.75,
      "learning_rate": 2.177601722944389e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 3105006.0,
      "step": 7128
    },
    {
      "epoch": 0.5483424351972925,
      "grad_norm": 8.4375,
      "learning_rate": 2.1774863472040613e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.996835470199585,
      "num_tokens": 3105324.0,
      "step": 7129
    },
    {
      "epoch": 0.5484193523575109,
      "grad_norm": 18.375,
      "learning_rate": 2.1773709714637335e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 3105628.0,
      "step": 7130
    },
    {
      "epoch": 0.5484962695177295,
      "grad_norm": 20.0,
      "learning_rate": 2.1772555957234062e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3105938.0,
      "step": 7131
    },
    {
      "epoch": 0.5485731866779479,
      "grad_norm": 14.5,
      "learning_rate": 2.1771402199830785e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3106208.0,
      "step": 7132
    },
    {
      "epoch": 0.5486501038381663,
      "grad_norm": 2.84375,
      "learning_rate": 2.1770248442427503e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3106766.0,
      "step": 7133
    },
    {
      "epoch": 0.5487270209983848,
      "grad_norm": 12.875,
      "learning_rate": 2.176909468502423e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3107050.0,
      "step": 7134
    },
    {
      "epoch": 0.5488039381586032,
      "grad_norm": 4.59375,
      "learning_rate": 2.1767940927620953e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3107606.0,
      "step": 7135
    },
    {
      "epoch": 0.5488808553188216,
      "grad_norm": 10.4375,
      "learning_rate": 2.1766787170217675e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3107932.0,
      "step": 7136
    },
    {
      "epoch": 0.5489577724790401,
      "grad_norm": 4.125,
      "learning_rate": 2.17656334128144e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3108488.0,
      "step": 7137
    },
    {
      "epoch": 0.5490346896392585,
      "grad_norm": 4.5,
      "learning_rate": 2.176447965541112e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 3109244.0,
      "step": 7138
    },
    {
      "epoch": 0.549111606799477,
      "grad_norm": 6.09375,
      "learning_rate": 2.1763325898007847e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3109818.0,
      "step": 7139
    },
    {
      "epoch": 0.5491885239596954,
      "grad_norm": 6.5,
      "learning_rate": 2.176217214060457e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 3110600.0,
      "step": 7140
    },
    {
      "epoch": 0.5492654411199138,
      "grad_norm": 4.96875,
      "learning_rate": 2.1761018383201292e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3111150.0,
      "step": 7141
    },
    {
      "epoch": 0.5493423582801323,
      "grad_norm": 4.25,
      "learning_rate": 2.175986462579802e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 3112194.0,
      "step": 7142
    },
    {
      "epoch": 0.5494192754403507,
      "grad_norm": 6.5625,
      "learning_rate": 2.1758710868394737e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 3112794.0,
      "step": 7143
    },
    {
      "epoch": 0.5494961926005691,
      "grad_norm": 8.8125,
      "learning_rate": 2.175755711099146e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 3113128.0,
      "step": 7144
    },
    {
      "epoch": 0.5495731097607877,
      "grad_norm": 8.5625,
      "learning_rate": 2.1756403353588187e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3113676.0,
      "step": 7145
    },
    {
      "epoch": 0.5496500269210061,
      "grad_norm": 11.0625,
      "learning_rate": 2.175524959618491e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3114224.0,
      "step": 7146
    },
    {
      "epoch": 0.5497269440812245,
      "grad_norm": 6.8125,
      "learning_rate": 2.175409583878163e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3114774.0,
      "step": 7147
    },
    {
      "epoch": 0.549803861241443,
      "grad_norm": 7.625,
      "learning_rate": 2.175294208137836e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3115102.0,
      "step": 7148
    },
    {
      "epoch": 0.5498807784016614,
      "grad_norm": 4.96875,
      "learning_rate": 2.1751788323975077e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 3115676.0,
      "step": 7149
    },
    {
      "epoch": 0.5499576955618799,
      "grad_norm": 4.0,
      "learning_rate": 2.1750634566571804e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9987684488296509,
      "num_tokens": 3116490.0,
      "step": 7150
    },
    {
      "epoch": 0.5500346127220983,
      "grad_norm": 13.5,
      "learning_rate": 2.1749480809168527e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3116814.0,
      "step": 7151
    },
    {
      "epoch": 0.5501115298823167,
      "grad_norm": 5.125,
      "learning_rate": 2.174832705176525e-06,
      "loss": 0.0042,
      "mean_token_accuracy": 0.999053955078125,
      "num_tokens": 3117874.0,
      "step": 7152
    },
    {
      "epoch": 0.5501884470425352,
      "grad_norm": 51.0,
      "learning_rate": 2.1747173294361976e-06,
      "loss": 0.0404,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3118180.0,
      "step": 7153
    },
    {
      "epoch": 0.5502653642027536,
      "grad_norm": 30.25,
      "learning_rate": 2.1746019536958694e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3118474.0,
      "step": 7154
    },
    {
      "epoch": 0.550342281362972,
      "grad_norm": 12.1875,
      "learning_rate": 2.1744865779555417e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3119070.0,
      "step": 7155
    },
    {
      "epoch": 0.5504191985231905,
      "grad_norm": 7.125,
      "learning_rate": 2.1743712022152144e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 3119646.0,
      "step": 7156
    },
    {
      "epoch": 0.550496115683409,
      "grad_norm": 38.5,
      "learning_rate": 2.1742558264748866e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3119978.0,
      "step": 7157
    },
    {
      "epoch": 0.5505730328436275,
      "grad_norm": 30.25,
      "learning_rate": 2.174140450734559e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3120310.0,
      "step": 7158
    },
    {
      "epoch": 0.5506499500038459,
      "grad_norm": 14.0,
      "learning_rate": 2.1740250749942316e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3120592.0,
      "step": 7159
    },
    {
      "epoch": 0.5507268671640643,
      "grad_norm": 9.125,
      "learning_rate": 2.1739096992539034e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3120922.0,
      "step": 7160
    },
    {
      "epoch": 0.5508037843242828,
      "grad_norm": 14.0,
      "learning_rate": 2.173794323513576e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3121218.0,
      "step": 7161
    },
    {
      "epoch": 0.5508807014845012,
      "grad_norm": 15.5,
      "learning_rate": 2.1736789477732483e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 3121470.0,
      "step": 7162
    },
    {
      "epoch": 0.5509576186447196,
      "grad_norm": 20.375,
      "learning_rate": 2.1735635720329206e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9784172773361206,
      "num_tokens": 3121750.0,
      "step": 7163
    },
    {
      "epoch": 0.5510345358049381,
      "grad_norm": 7.59375,
      "learning_rate": 2.173448196292593e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 3122770.0,
      "step": 7164
    },
    {
      "epoch": 0.5511114529651565,
      "grad_norm": 5.78125,
      "learning_rate": 2.173332820552265e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3123288.0,
      "step": 7165
    },
    {
      "epoch": 0.5511883701253749,
      "grad_norm": 3.84375,
      "learning_rate": 2.1732174448119374e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3123834.0,
      "step": 7166
    },
    {
      "epoch": 0.5512652872855934,
      "grad_norm": 8.0625,
      "learning_rate": 2.17310206907161e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 3124350.0,
      "step": 7167
    },
    {
      "epoch": 0.5513422044458118,
      "grad_norm": 9.25,
      "learning_rate": 2.1729866933312823e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3124632.0,
      "step": 7168
    },
    {
      "epoch": 0.5514191216060303,
      "grad_norm": 6.34375,
      "learning_rate": 2.1728713175909546e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3125178.0,
      "step": 7169
    },
    {
      "epoch": 0.5514960387662488,
      "grad_norm": 20.5,
      "learning_rate": 2.1727559418506273e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3125510.0,
      "step": 7170
    },
    {
      "epoch": 0.5515729559264672,
      "grad_norm": 14.25,
      "learning_rate": 2.172640566110299e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3126072.0,
      "step": 7171
    },
    {
      "epoch": 0.5516498730866857,
      "grad_norm": 18.375,
      "learning_rate": 2.1725251903699713e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3126396.0,
      "step": 7172
    },
    {
      "epoch": 0.5517267902469041,
      "grad_norm": 6.6875,
      "learning_rate": 2.172409814629644e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 3126912.0,
      "step": 7173
    },
    {
      "epoch": 0.5518037074071225,
      "grad_norm": 24.375,
      "learning_rate": 2.1722944388893163e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3127222.0,
      "step": 7174
    },
    {
      "epoch": 0.551880624567341,
      "grad_norm": 18.5,
      "learning_rate": 2.1721790631489885e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 3127572.0,
      "step": 7175
    },
    {
      "epoch": 0.5519575417275594,
      "grad_norm": 3.5,
      "learning_rate": 2.172063687408661e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3128152.0,
      "step": 7176
    },
    {
      "epoch": 0.5520344588877779,
      "grad_norm": 6.03125,
      "learning_rate": 2.171948311668333e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3128708.0,
      "step": 7177
    },
    {
      "epoch": 0.5521113760479963,
      "grad_norm": 56.25,
      "learning_rate": 2.1718329359280057e-06,
      "loss": 0.0481,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 3128960.0,
      "step": 7178
    },
    {
      "epoch": 0.5521882932082147,
      "grad_norm": 11.6875,
      "learning_rate": 2.171717560187678e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3129534.0,
      "step": 7179
    },
    {
      "epoch": 0.5522652103684332,
      "grad_norm": 28.125,
      "learning_rate": 2.1716021844473503e-06,
      "loss": 0.0411,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 3129872.0,
      "step": 7180
    },
    {
      "epoch": 0.5523421275286516,
      "grad_norm": 26.25,
      "learning_rate": 2.1714868087070225e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 3130184.0,
      "step": 7181
    },
    {
      "epoch": 0.55241904468887,
      "grad_norm": 24.375,
      "learning_rate": 2.1713714329666948e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 3130494.0,
      "step": 7182
    },
    {
      "epoch": 0.5524959618490886,
      "grad_norm": 19.875,
      "learning_rate": 2.171256057226367e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3130778.0,
      "step": 7183
    },
    {
      "epoch": 0.552572879009307,
      "grad_norm": 12.125,
      "learning_rate": 2.1711406814860397e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3131072.0,
      "step": 7184
    },
    {
      "epoch": 0.5526497961695255,
      "grad_norm": 5.09375,
      "learning_rate": 2.171025305745712e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3131616.0,
      "step": 7185
    },
    {
      "epoch": 0.5527267133297439,
      "grad_norm": 5.5,
      "learning_rate": 2.1709099300053842e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3132162.0,
      "step": 7186
    },
    {
      "epoch": 0.5528036304899623,
      "grad_norm": 8.4375,
      "learning_rate": 2.1707945542650565e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3132496.0,
      "step": 7187
    },
    {
      "epoch": 0.5528805476501808,
      "grad_norm": 15.5,
      "learning_rate": 2.1706791785247287e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 3133320.0,
      "step": 7188
    },
    {
      "epoch": 0.5529574648103992,
      "grad_norm": 19.5,
      "learning_rate": 2.1705638027844014e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3133646.0,
      "step": 7189
    },
    {
      "epoch": 0.5530343819706176,
      "grad_norm": 13.3125,
      "learning_rate": 2.1704484270440737e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3133948.0,
      "step": 7190
    },
    {
      "epoch": 0.5531112991308361,
      "grad_norm": 13.375,
      "learning_rate": 2.170333051303746e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3134228.0,
      "step": 7191
    },
    {
      "epoch": 0.5531882162910545,
      "grad_norm": 9.6875,
      "learning_rate": 2.170217675563418e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 3134508.0,
      "step": 7192
    },
    {
      "epoch": 0.5532651334512729,
      "grad_norm": 8.8125,
      "learning_rate": 2.1701022998230905e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 3134814.0,
      "step": 7193
    },
    {
      "epoch": 0.5533420506114914,
      "grad_norm": 10.25,
      "learning_rate": 2.1699869240827627e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3135394.0,
      "step": 7194
    },
    {
      "epoch": 0.5534189677717098,
      "grad_norm": 13.5625,
      "learning_rate": 2.1698715483424354e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3135692.0,
      "step": 7195
    },
    {
      "epoch": 0.5534958849319284,
      "grad_norm": 12.6875,
      "learning_rate": 2.1697561726021077e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 3135980.0,
      "step": 7196
    },
    {
      "epoch": 0.5535728020921468,
      "grad_norm": 15.3125,
      "learning_rate": 2.16964079686178e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 3136336.0,
      "step": 7197
    },
    {
      "epoch": 0.5536497192523652,
      "grad_norm": 5.84375,
      "learning_rate": 2.169525421121452e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3136854.0,
      "step": 7198
    },
    {
      "epoch": 0.5537266364125837,
      "grad_norm": 8.9375,
      "learning_rate": 2.1694100453811244e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9966996908187866,
      "num_tokens": 3137158.0,
      "step": 7199
    },
    {
      "epoch": 0.5538035535728021,
      "grad_norm": 4.78125,
      "learning_rate": 2.169294669640797e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3137726.0,
      "step": 7200
    },
    {
      "epoch": 0.5538804707330205,
      "grad_norm": 13.25,
      "learning_rate": 2.1691792939004694e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3138046.0,
      "step": 7201
    },
    {
      "epoch": 0.553957387893239,
      "grad_norm": 7.09375,
      "learning_rate": 2.1690639181601416e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3138306.0,
      "step": 7202
    },
    {
      "epoch": 0.5540343050534574,
      "grad_norm": 30.0,
      "learning_rate": 2.168948542419814e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 3138678.0,
      "step": 7203
    },
    {
      "epoch": 0.5541112222136759,
      "grad_norm": 12.1875,
      "learning_rate": 2.168833166679486e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3138998.0,
      "step": 7204
    },
    {
      "epoch": 0.5541881393738943,
      "grad_norm": 26.875,
      "learning_rate": 2.1687177909391584e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 3139298.0,
      "step": 7205
    },
    {
      "epoch": 0.5542650565341127,
      "grad_norm": 2.9375,
      "learning_rate": 2.168602415198831e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 3140080.0,
      "step": 7206
    },
    {
      "epoch": 0.5543419736943312,
      "grad_norm": 24.625,
      "learning_rate": 2.1684870394585033e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3140388.0,
      "step": 7207
    },
    {
      "epoch": 0.5544188908545497,
      "grad_norm": 20.0,
      "learning_rate": 2.1683716637181756e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3140682.0,
      "step": 7208
    },
    {
      "epoch": 0.554495808014768,
      "grad_norm": 8.25,
      "learning_rate": 2.168256287977848e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3141222.0,
      "step": 7209
    },
    {
      "epoch": 0.5545727251749866,
      "grad_norm": 8.8125,
      "learning_rate": 2.16814091223752e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3141546.0,
      "step": 7210
    },
    {
      "epoch": 0.554649642335205,
      "grad_norm": 39.25,
      "learning_rate": 2.168025536497193e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 3141824.0,
      "step": 7211
    },
    {
      "epoch": 0.5547265594954234,
      "grad_norm": 31.875,
      "learning_rate": 2.167910160756865e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3142126.0,
      "step": 7212
    },
    {
      "epoch": 0.5548034766556419,
      "grad_norm": 7.875,
      "learning_rate": 2.1677947850165373e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 3142440.0,
      "step": 7213
    },
    {
      "epoch": 0.5548803938158603,
      "grad_norm": 31.75,
      "learning_rate": 2.1676794092762096e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3142718.0,
      "step": 7214
    },
    {
      "epoch": 0.5549573109760788,
      "grad_norm": 12.5625,
      "learning_rate": 2.167564033535882e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3142974.0,
      "step": 7215
    },
    {
      "epoch": 0.5550342281362972,
      "grad_norm": 5.4375,
      "learning_rate": 2.167448657795554e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3143282.0,
      "step": 7216
    },
    {
      "epoch": 0.5551111452965156,
      "grad_norm": 4.1875,
      "learning_rate": 2.1673332820552268e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3143802.0,
      "step": 7217
    },
    {
      "epoch": 0.5551880624567341,
      "grad_norm": 4.03125,
      "learning_rate": 2.167217906314899e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3144342.0,
      "step": 7218
    },
    {
      "epoch": 0.5552649796169525,
      "grad_norm": 9.125,
      "learning_rate": 2.1671025305745713e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9915110468864441,
      "num_tokens": 3144932.0,
      "step": 7219
    },
    {
      "epoch": 0.5553418967771709,
      "grad_norm": 8.25,
      "learning_rate": 2.1669871548342435e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3145272.0,
      "step": 7220
    },
    {
      "epoch": 0.5554188139373895,
      "grad_norm": 5.34375,
      "learning_rate": 2.166871779093916e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3145550.0,
      "step": 7221
    },
    {
      "epoch": 0.5554957310976079,
      "grad_norm": 5.625,
      "learning_rate": 2.166756403353588e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3146072.0,
      "step": 7222
    },
    {
      "epoch": 0.5555726482578264,
      "grad_norm": 6.09375,
      "learning_rate": 2.1666410276132607e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 3146880.0,
      "step": 7223
    },
    {
      "epoch": 0.5556495654180448,
      "grad_norm": 4.46875,
      "learning_rate": 2.166525651872933e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3147404.0,
      "step": 7224
    },
    {
      "epoch": 0.5557264825782632,
      "grad_norm": 10.3125,
      "learning_rate": 2.1664102761326053e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 3148006.0,
      "step": 7225
    },
    {
      "epoch": 0.5558033997384817,
      "grad_norm": 13.6875,
      "learning_rate": 2.1662949003922775e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3148286.0,
      "step": 7226
    },
    {
      "epoch": 0.5558803168987001,
      "grad_norm": 4.25,
      "learning_rate": 2.1661795246519498e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3148856.0,
      "step": 7227
    },
    {
      "epoch": 0.5559572340589185,
      "grad_norm": 9.625,
      "learning_rate": 2.1660641489116225e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3149162.0,
      "step": 7228
    },
    {
      "epoch": 0.556034151219137,
      "grad_norm": 5.625,
      "learning_rate": 2.1659487731712947e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9939172863960266,
      "num_tokens": 3149986.0,
      "step": 7229
    },
    {
      "epoch": 0.5561110683793554,
      "grad_norm": 9.125,
      "learning_rate": 2.1658333974309666e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3150292.0,
      "step": 7230
    },
    {
      "epoch": 0.5561879855395738,
      "grad_norm": 7.0,
      "learning_rate": 2.1657180216906392e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 3150642.0,
      "step": 7231
    },
    {
      "epoch": 0.5562649026997923,
      "grad_norm": 2.703125,
      "learning_rate": 2.1656026459503115e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3151190.0,
      "step": 7232
    },
    {
      "epoch": 0.5563418198600107,
      "grad_norm": 4.34375,
      "learning_rate": 2.1654872702099837e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 3152014.0,
      "step": 7233
    },
    {
      "epoch": 0.5564187370202293,
      "grad_norm": 10.0,
      "learning_rate": 2.1653718944696564e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3152346.0,
      "step": 7234
    },
    {
      "epoch": 0.5564956541804477,
      "grad_norm": 8.6875,
      "learning_rate": 2.1652565187293287e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3152690.0,
      "step": 7235
    },
    {
      "epoch": 0.5565725713406661,
      "grad_norm": 13.1875,
      "learning_rate": 2.165141142989001e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 3152998.0,
      "step": 7236
    },
    {
      "epoch": 0.5566494885008846,
      "grad_norm": 14.375,
      "learning_rate": 2.165025767248673e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3153280.0,
      "step": 7237
    },
    {
      "epoch": 0.556726405661103,
      "grad_norm": 5.46875,
      "learning_rate": 2.1649103915083455e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3153876.0,
      "step": 7238
    },
    {
      "epoch": 0.5568033228213214,
      "grad_norm": 7.3125,
      "learning_rate": 2.164795015768018e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 3154214.0,
      "step": 7239
    },
    {
      "epoch": 0.5568802399815399,
      "grad_norm": 5.40625,
      "learning_rate": 2.1646796400276904e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3154542.0,
      "step": 7240
    },
    {
      "epoch": 0.5569571571417583,
      "grad_norm": 2.71875,
      "learning_rate": 2.1645642642873622e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 3155328.0,
      "step": 7241
    },
    {
      "epoch": 0.5570340743019768,
      "grad_norm": 5.9375,
      "learning_rate": 2.164448888547035e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3155650.0,
      "step": 7242
    },
    {
      "epoch": 0.5571109914621952,
      "grad_norm": 5.46875,
      "learning_rate": 2.164333512806707e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3155972.0,
      "step": 7243
    },
    {
      "epoch": 0.5571879086224136,
      "grad_norm": 13.9375,
      "learning_rate": 2.1642181370663794e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3156280.0,
      "step": 7244
    },
    {
      "epoch": 0.5572648257826321,
      "grad_norm": 4.6875,
      "learning_rate": 2.164102761326052e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3156822.0,
      "step": 7245
    },
    {
      "epoch": 0.5573417429428505,
      "grad_norm": 4.46875,
      "learning_rate": 2.163987385585724e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3157364.0,
      "step": 7246
    },
    {
      "epoch": 0.557418660103069,
      "grad_norm": 8.3125,
      "learning_rate": 2.1638720098453966e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 3157720.0,
      "step": 7247
    },
    {
      "epoch": 0.5574955772632875,
      "grad_norm": 8.875,
      "learning_rate": 2.163756634105069e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3158064.0,
      "step": 7248
    },
    {
      "epoch": 0.5575724944235059,
      "grad_norm": 15.5625,
      "learning_rate": 2.163641258364741e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3158364.0,
      "step": 7249
    },
    {
      "epoch": 0.5576494115837243,
      "grad_norm": 9.8125,
      "learning_rate": 2.163525882624414e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3158682.0,
      "step": 7250
    },
    {
      "epoch": 0.5577263287439428,
      "grad_norm": 6.25,
      "learning_rate": 2.163410506884086e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3159268.0,
      "step": 7251
    },
    {
      "epoch": 0.5578032459041612,
      "grad_norm": 6.53125,
      "learning_rate": 2.163295131143758e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3159564.0,
      "step": 7252
    },
    {
      "epoch": 0.5578801630643797,
      "grad_norm": 6.46875,
      "learning_rate": 2.1631797554034306e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3159870.0,
      "step": 7253
    },
    {
      "epoch": 0.5579570802245981,
      "grad_norm": 7.625,
      "learning_rate": 2.163064379663103e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3160210.0,
      "step": 7254
    },
    {
      "epoch": 0.5580339973848165,
      "grad_norm": 9.4375,
      "learning_rate": 2.162949003922775e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3160538.0,
      "step": 7255
    },
    {
      "epoch": 0.558110914545035,
      "grad_norm": 10.0625,
      "learning_rate": 2.162833628182448e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3160890.0,
      "step": 7256
    },
    {
      "epoch": 0.5581878317052534,
      "grad_norm": 5.78125,
      "learning_rate": 2.1627182524421196e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 3161238.0,
      "step": 7257
    },
    {
      "epoch": 0.5582647488654718,
      "grad_norm": 6.59375,
      "learning_rate": 2.1626028767017923e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3161566.0,
      "step": 7258
    },
    {
      "epoch": 0.5583416660256904,
      "grad_norm": 17.75,
      "learning_rate": 2.1624875009614646e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.978723406791687,
      "num_tokens": 3161850.0,
      "step": 7259
    },
    {
      "epoch": 0.5584185831859088,
      "grad_norm": 14.875,
      "learning_rate": 2.162372125221137e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3162134.0,
      "step": 7260
    },
    {
      "epoch": 0.5584955003461273,
      "grad_norm": 5.21875,
      "learning_rate": 2.1622567494808095e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9983193278312683,
      "num_tokens": 3162730.0,
      "step": 7261
    },
    {
      "epoch": 0.5585724175063457,
      "grad_norm": 5.75,
      "learning_rate": 2.1621413737404818e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3163080.0,
      "step": 7262
    },
    {
      "epoch": 0.5586493346665641,
      "grad_norm": 20.625,
      "learning_rate": 2.1620259980001536e-06,
      "loss": 0.0391,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 3163356.0,
      "step": 7263
    },
    {
      "epoch": 0.5587262518267826,
      "grad_norm": 3.78125,
      "learning_rate": 2.1619106222598263e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3163706.0,
      "step": 7264
    },
    {
      "epoch": 0.558803168987001,
      "grad_norm": 8.0625,
      "learning_rate": 2.1617952465194985e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9973190426826477,
      "num_tokens": 3164082.0,
      "step": 7265
    },
    {
      "epoch": 0.5588800861472194,
      "grad_norm": 5.3125,
      "learning_rate": 2.161679870779171e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3164672.0,
      "step": 7266
    },
    {
      "epoch": 0.5589570033074379,
      "grad_norm": 5.09375,
      "learning_rate": 2.1615644950388435e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3165226.0,
      "step": 7267
    },
    {
      "epoch": 0.5590339204676563,
      "grad_norm": 7.125,
      "learning_rate": 2.1614491192985153e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 3165798.0,
      "step": 7268
    },
    {
      "epoch": 0.5591108376278747,
      "grad_norm": 15.25,
      "learning_rate": 2.161333743558188e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9889298677444458,
      "num_tokens": 3166070.0,
      "step": 7269
    },
    {
      "epoch": 0.5591877547880932,
      "grad_norm": 25.5,
      "learning_rate": 2.1612183678178603e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3166348.0,
      "step": 7270
    },
    {
      "epoch": 0.5592646719483116,
      "grad_norm": 3.953125,
      "learning_rate": 2.1611029920775325e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 3166924.0,
      "step": 7271
    },
    {
      "epoch": 0.5593415891085302,
      "grad_norm": 26.125,
      "learning_rate": 2.1609876163372048e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3167226.0,
      "step": 7272
    },
    {
      "epoch": 0.5594185062687486,
      "grad_norm": 6.40625,
      "learning_rate": 2.1608722405968775e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3167548.0,
      "step": 7273
    },
    {
      "epoch": 0.559495423428967,
      "grad_norm": 8.5,
      "learning_rate": 2.1607568648565493e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 3168384.0,
      "step": 7274
    },
    {
      "epoch": 0.5595723405891855,
      "grad_norm": 12.0,
      "learning_rate": 2.160641489116222e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3168670.0,
      "step": 7275
    },
    {
      "epoch": 0.5596492577494039,
      "grad_norm": 7.875,
      "learning_rate": 2.1605261133758942e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 3169022.0,
      "step": 7276
    },
    {
      "epoch": 0.5597261749096223,
      "grad_norm": 8.125,
      "learning_rate": 2.1604107376355665e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 3169352.0,
      "step": 7277
    },
    {
      "epoch": 0.5598030920698408,
      "grad_norm": 4.78125,
      "learning_rate": 2.160295361895239e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3169892.0,
      "step": 7278
    },
    {
      "epoch": 0.5598800092300592,
      "grad_norm": 3.9375,
      "learning_rate": 2.160179986154911e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3170460.0,
      "step": 7279
    },
    {
      "epoch": 0.5599569263902777,
      "grad_norm": 10.1875,
      "learning_rate": 2.1600646104145833e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 3171286.0,
      "step": 7280
    },
    {
      "epoch": 0.5600338435504961,
      "grad_norm": 8.4375,
      "learning_rate": 2.159949234674256e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3171596.0,
      "step": 7281
    },
    {
      "epoch": 0.5601107607107145,
      "grad_norm": 5.03125,
      "learning_rate": 2.159833858933928e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 3172384.0,
      "step": 7282
    },
    {
      "epoch": 0.560187677870933,
      "grad_norm": 6.625,
      "learning_rate": 2.1597184831936005e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3172730.0,
      "step": 7283
    },
    {
      "epoch": 0.5602645950311514,
      "grad_norm": 5.71875,
      "learning_rate": 2.1596031074532727e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 3173080.0,
      "step": 7284
    },
    {
      "epoch": 0.5603415121913699,
      "grad_norm": 10.1875,
      "learning_rate": 2.159487731712945e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3173666.0,
      "step": 7285
    },
    {
      "epoch": 0.5604184293515884,
      "grad_norm": 4.25,
      "learning_rate": 2.1593723559726177e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 3174710.0,
      "step": 7286
    },
    {
      "epoch": 0.5604953465118068,
      "grad_norm": 7.96875,
      "learning_rate": 2.15925698023229e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 3175044.0,
      "step": 7287
    },
    {
      "epoch": 0.5605722636720253,
      "grad_norm": 5.71875,
      "learning_rate": 2.159141604491962e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 3175396.0,
      "step": 7288
    },
    {
      "epoch": 0.5606491808322437,
      "grad_norm": 26.75,
      "learning_rate": 2.159026228751635e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3175678.0,
      "step": 7289
    },
    {
      "epoch": 0.5607260979924621,
      "grad_norm": 9.5,
      "learning_rate": 2.1589108530113067e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 3176034.0,
      "step": 7290
    },
    {
      "epoch": 0.5608030151526806,
      "grad_norm": 18.25,
      "learning_rate": 2.158795477270979e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 3176322.0,
      "step": 7291
    },
    {
      "epoch": 0.560879932312899,
      "grad_norm": 3.765625,
      "learning_rate": 2.1586801015306516e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9963833689689636,
      "num_tokens": 3176876.0,
      "step": 7292
    },
    {
      "epoch": 0.5609568494731174,
      "grad_norm": 23.5,
      "learning_rate": 2.158564725790324e-06,
      "loss": 0.044,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3177136.0,
      "step": 7293
    },
    {
      "epoch": 0.5610337666333359,
      "grad_norm": 5.9375,
      "learning_rate": 2.158449350049996e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 3177672.0,
      "step": 7294
    },
    {
      "epoch": 0.5611106837935543,
      "grad_norm": 5.21875,
      "learning_rate": 2.1583339743096684e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9971510171890259,
      "num_tokens": 3178024.0,
      "step": 7295
    },
    {
      "epoch": 0.5611876009537727,
      "grad_norm": 9.875,
      "learning_rate": 2.1582185985693407e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3178542.0,
      "step": 7296
    },
    {
      "epoch": 0.5612645181139913,
      "grad_norm": 9.0,
      "learning_rate": 2.1581032228290134e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3178836.0,
      "step": 7297
    },
    {
      "epoch": 0.5613414352742097,
      "grad_norm": 5.8125,
      "learning_rate": 2.1579878470886856e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 3179136.0,
      "step": 7298
    },
    {
      "epoch": 0.5614183524344282,
      "grad_norm": 6.28125,
      "learning_rate": 2.157872471348358e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3179454.0,
      "step": 7299
    },
    {
      "epoch": 0.5614952695946466,
      "grad_norm": 4.90625,
      "learning_rate": 2.1577570956080305e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3179976.0,
      "step": 7300
    },
    {
      "epoch": 0.561572186754865,
      "grad_norm": 4.375,
      "learning_rate": 2.1576417198677024e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3180494.0,
      "step": 7301
    },
    {
      "epoch": 0.5616491039150835,
      "grad_norm": 7.40625,
      "learning_rate": 2.1575263441273746e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3180814.0,
      "step": 7302
    },
    {
      "epoch": 0.5617260210753019,
      "grad_norm": 5.75,
      "learning_rate": 2.1574109683870473e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9969696998596191,
      "num_tokens": 3181146.0,
      "step": 7303
    },
    {
      "epoch": 0.5618029382355203,
      "grad_norm": 7.75,
      "learning_rate": 2.1572955926467196e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3181704.0,
      "step": 7304
    },
    {
      "epoch": 0.5618798553957388,
      "grad_norm": 10.0625,
      "learning_rate": 2.157180216906392e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3182038.0,
      "step": 7305
    },
    {
      "epoch": 0.5619567725559572,
      "grad_norm": 6.84375,
      "learning_rate": 2.157064841166064e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3182594.0,
      "step": 7306
    },
    {
      "epoch": 0.5620336897161757,
      "grad_norm": 4.28125,
      "learning_rate": 2.1569494654257364e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3183114.0,
      "step": 7307
    },
    {
      "epoch": 0.5621106068763941,
      "grad_norm": 10.4375,
      "learning_rate": 2.156834089685409e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3183426.0,
      "step": 7308
    },
    {
      "epoch": 0.5621875240366125,
      "grad_norm": 20.375,
      "learning_rate": 2.1567187139450813e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3183756.0,
      "step": 7309
    },
    {
      "epoch": 0.562264441196831,
      "grad_norm": 4.25,
      "learning_rate": 2.1566033382047536e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 3184056.0,
      "step": 7310
    },
    {
      "epoch": 0.5623413583570495,
      "grad_norm": 15.0625,
      "learning_rate": 2.1564879624644262e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9821958541870117,
      "num_tokens": 3184394.0,
      "step": 7311
    },
    {
      "epoch": 0.5624182755172679,
      "grad_norm": 5.46875,
      "learning_rate": 2.156372586724098e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.993914783000946,
      "num_tokens": 3184888.0,
      "step": 7312
    },
    {
      "epoch": 0.5624951926774864,
      "grad_norm": 3.234375,
      "learning_rate": 2.1562572109837703e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3185458.0,
      "step": 7313
    },
    {
      "epoch": 0.5625721098377048,
      "grad_norm": 3.96875,
      "learning_rate": 2.156141835243443e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 3186026.0,
      "step": 7314
    },
    {
      "epoch": 0.5626490269979232,
      "grad_norm": 18.375,
      "learning_rate": 2.1560264595031153e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3186354.0,
      "step": 7315
    },
    {
      "epoch": 0.5627259441581417,
      "grad_norm": 5.25,
      "learning_rate": 2.1559110837627875e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9909256100654602,
      "num_tokens": 3186906.0,
      "step": 7316
    },
    {
      "epoch": 0.5628028613183601,
      "grad_norm": 10.625,
      "learning_rate": 2.1557957080224598e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3187246.0,
      "step": 7317
    },
    {
      "epoch": 0.5628797784785786,
      "grad_norm": 25.125,
      "learning_rate": 2.155680332282132e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3187524.0,
      "step": 7318
    },
    {
      "epoch": 0.562956695638797,
      "grad_norm": 22.0,
      "learning_rate": 2.1555649565418047e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9868995547294617,
      "num_tokens": 3187754.0,
      "step": 7319
    },
    {
      "epoch": 0.5630336127990154,
      "grad_norm": 12.6875,
      "learning_rate": 2.155449580801477e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3188344.0,
      "step": 7320
    },
    {
      "epoch": 0.5631105299592339,
      "grad_norm": 46.75,
      "learning_rate": 2.1553342050611492e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3188684.0,
      "step": 7321
    },
    {
      "epoch": 0.5631874471194523,
      "grad_norm": 11.1875,
      "learning_rate": 2.1552188293208215e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.996610164642334,
      "num_tokens": 3188980.0,
      "step": 7322
    },
    {
      "epoch": 0.5632643642796707,
      "grad_norm": 5.25,
      "learning_rate": 2.1551034535804938e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9980879426002502,
      "num_tokens": 3189504.0,
      "step": 7323
    },
    {
      "epoch": 0.5633412814398893,
      "grad_norm": 6.1875,
      "learning_rate": 2.154988077840166e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 3190068.0,
      "step": 7324
    },
    {
      "epoch": 0.5634181986001077,
      "grad_norm": 7.25,
      "learning_rate": 2.1548727020998387e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 3190382.0,
      "step": 7325
    },
    {
      "epoch": 0.5634951157603262,
      "grad_norm": 8.1875,
      "learning_rate": 2.154757326359511e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 3190722.0,
      "step": 7326
    },
    {
      "epoch": 0.5635720329205446,
      "grad_norm": 6.125,
      "learning_rate": 2.154641950619183e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 3191280.0,
      "step": 7327
    },
    {
      "epoch": 0.563648950080763,
      "grad_norm": 7.28125,
      "learning_rate": 2.1545265748788555e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3191798.0,
      "step": 7328
    },
    {
      "epoch": 0.5637258672409815,
      "grad_norm": 23.5,
      "learning_rate": 2.1544111991385277e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 3192394.0,
      "step": 7329
    },
    {
      "epoch": 0.5638027844011999,
      "grad_norm": 8.6875,
      "learning_rate": 2.1542958233982e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3192722.0,
      "step": 7330
    },
    {
      "epoch": 0.5638797015614183,
      "grad_norm": 24.0,
      "learning_rate": 2.1541804476578727e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3192980.0,
      "step": 7331
    },
    {
      "epoch": 0.5639566187216368,
      "grad_norm": 23.625,
      "learning_rate": 2.154065071917545e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3193300.0,
      "step": 7332
    },
    {
      "epoch": 0.5640335358818552,
      "grad_norm": 5.75,
      "learning_rate": 2.153949696177217e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 3193628.0,
      "step": 7333
    },
    {
      "epoch": 0.5641104530420736,
      "grad_norm": 8.875,
      "learning_rate": 2.1538343204368894e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 3194120.0,
      "step": 7334
    },
    {
      "epoch": 0.5641873702022921,
      "grad_norm": 11.4375,
      "learning_rate": 2.1537189446965617e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 3194472.0,
      "step": 7335
    },
    {
      "epoch": 0.5642642873625106,
      "grad_norm": 19.75,
      "learning_rate": 2.1536035689562344e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3194812.0,
      "step": 7336
    },
    {
      "epoch": 0.5643412045227291,
      "grad_norm": 3.453125,
      "learning_rate": 2.1534881932159066e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3195570.0,
      "step": 7337
    },
    {
      "epoch": 0.5644181216829475,
      "grad_norm": 6.875,
      "learning_rate": 2.153372817475579e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3195916.0,
      "step": 7338
    },
    {
      "epoch": 0.5644950388431659,
      "grad_norm": 6.125,
      "learning_rate": 2.153257441735251e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 3196214.0,
      "step": 7339
    },
    {
      "epoch": 0.5645719560033844,
      "grad_norm": 10.375,
      "learning_rate": 2.1531420659949234e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3196770.0,
      "step": 7340
    },
    {
      "epoch": 0.5646488731636028,
      "grad_norm": 19.25,
      "learning_rate": 2.1530266902545957e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3197078.0,
      "step": 7341
    },
    {
      "epoch": 0.5647257903238212,
      "grad_norm": 6.75,
      "learning_rate": 2.1529113145142684e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3197428.0,
      "step": 7342
    },
    {
      "epoch": 0.5648027074840397,
      "grad_norm": 35.5,
      "learning_rate": 2.1527959387739406e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 3197768.0,
      "step": 7343
    },
    {
      "epoch": 0.5648796246442581,
      "grad_norm": 31.75,
      "learning_rate": 2.152680563033613e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3198048.0,
      "step": 7344
    },
    {
      "epoch": 0.5649565418044766,
      "grad_norm": 5.625,
      "learning_rate": 2.152565187293285e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 3198620.0,
      "step": 7345
    },
    {
      "epoch": 0.565033458964695,
      "grad_norm": 3.875,
      "learning_rate": 2.1524498115529574e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3199402.0,
      "step": 7346
    },
    {
      "epoch": 0.5651103761249134,
      "grad_norm": 13.375,
      "learning_rate": 2.15233443581263e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3199982.0,
      "step": 7347
    },
    {
      "epoch": 0.565187293285132,
      "grad_norm": 21.375,
      "learning_rate": 2.1522190600723023e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9983193278312683,
      "num_tokens": 3200578.0,
      "step": 7348
    },
    {
      "epoch": 0.5652642104453504,
      "grad_norm": 4.34375,
      "learning_rate": 2.152103684331974e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3201124.0,
      "step": 7349
    },
    {
      "epoch": 0.5653411276055688,
      "grad_norm": 41.0,
      "learning_rate": 2.151988308591647e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3201382.0,
      "step": 7350
    },
    {
      "epoch": 0.5654180447657873,
      "grad_norm": 5.21875,
      "learning_rate": 2.151872932851319e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3201708.0,
      "step": 7351
    },
    {
      "epoch": 0.5654949619260057,
      "grad_norm": 12.0625,
      "learning_rate": 2.1517575571109914e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3202028.0,
      "step": 7352
    },
    {
      "epoch": 0.5655718790862241,
      "grad_norm": 11.5,
      "learning_rate": 2.151642181370664e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3202562.0,
      "step": 7353
    },
    {
      "epoch": 0.5656487962464426,
      "grad_norm": 19.375,
      "learning_rate": 2.1515268056303363e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3202896.0,
      "step": 7354
    },
    {
      "epoch": 0.565725713406661,
      "grad_norm": 14.9375,
      "learning_rate": 2.1514114298900086e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3203452.0,
      "step": 7355
    },
    {
      "epoch": 0.5658026305668795,
      "grad_norm": 9.5,
      "learning_rate": 2.151296054149681e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9972066879272461,
      "num_tokens": 3203812.0,
      "step": 7356
    },
    {
      "epoch": 0.5658795477270979,
      "grad_norm": 4.21875,
      "learning_rate": 2.151180678409353e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 3204380.0,
      "step": 7357
    },
    {
      "epoch": 0.5659564648873163,
      "grad_norm": 22.625,
      "learning_rate": 2.1510653026690258e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3204726.0,
      "step": 7358
    },
    {
      "epoch": 0.5660333820475348,
      "grad_norm": 20.875,
      "learning_rate": 2.150949926928698e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3204984.0,
      "step": 7359
    },
    {
      "epoch": 0.5661102992077532,
      "grad_norm": 18.5,
      "learning_rate": 2.15083455118837e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3205324.0,
      "step": 7360
    },
    {
      "epoch": 0.5661872163679716,
      "grad_norm": 5.21875,
      "learning_rate": 2.1507191754480425e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3205894.0,
      "step": 7361
    },
    {
      "epoch": 0.5662641335281902,
      "grad_norm": 9.8125,
      "learning_rate": 2.1506037997077148e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 3206708.0,
      "step": 7362
    },
    {
      "epoch": 0.5663410506884086,
      "grad_norm": 6.78125,
      "learning_rate": 2.150488423967387e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3207252.0,
      "step": 7363
    },
    {
      "epoch": 0.5664179678486271,
      "grad_norm": 30.75,
      "learning_rate": 2.1503730482270597e-06,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9784172773361206,
      "num_tokens": 3207532.0,
      "step": 7364
    },
    {
      "epoch": 0.5664948850088455,
      "grad_norm": 33.75,
      "learning_rate": 2.150257672486732e-06,
      "loss": 0.0399,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 3207810.0,
      "step": 7365
    },
    {
      "epoch": 0.5665718021690639,
      "grad_norm": 42.0,
      "learning_rate": 2.1501422967464042e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 3208412.0,
      "step": 7366
    },
    {
      "epoch": 0.5666487193292824,
      "grad_norm": 58.25,
      "learning_rate": 2.1500269210060765e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3208718.0,
      "step": 7367
    },
    {
      "epoch": 0.5667256364895008,
      "grad_norm": 12.875,
      "learning_rate": 2.1499115452657488e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3209500.0,
      "step": 7368
    },
    {
      "epoch": 0.5668025536497192,
      "grad_norm": 30.875,
      "learning_rate": 2.1497961695254214e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 3210320.0,
      "step": 7369
    },
    {
      "epoch": 0.5668794708099377,
      "grad_norm": 13.25,
      "learning_rate": 2.1496807937850937e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3210862.0,
      "step": 7370
    },
    {
      "epoch": 0.5669563879701561,
      "grad_norm": 15.625,
      "learning_rate": 2.1495654180447655e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 3211358.0,
      "step": 7371
    },
    {
      "epoch": 0.5670333051303745,
      "grad_norm": 24.0,
      "learning_rate": 2.1494500423044382e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3211920.0,
      "step": 7372
    },
    {
      "epoch": 0.567110222290593,
      "grad_norm": 5.40625,
      "learning_rate": 2.1493346665641105e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 3212490.0,
      "step": 7373
    },
    {
      "epoch": 0.5671871394508115,
      "grad_norm": 6.96875,
      "learning_rate": 2.1492192908237827e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3213034.0,
      "step": 7374
    },
    {
      "epoch": 0.56726405661103,
      "grad_norm": 43.0,
      "learning_rate": 2.1491039150834554e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3213368.0,
      "step": 7375
    },
    {
      "epoch": 0.5673409737712484,
      "grad_norm": 3.15625,
      "learning_rate": 2.1489885393431277e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3213932.0,
      "step": 7376
    },
    {
      "epoch": 0.5674178909314668,
      "grad_norm": 10.8125,
      "learning_rate": 2.1488731636028e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 3214224.0,
      "step": 7377
    },
    {
      "epoch": 0.5674948080916853,
      "grad_norm": 31.5,
      "learning_rate": 2.148757787862472e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3214540.0,
      "step": 7378
    },
    {
      "epoch": 0.5675717252519037,
      "grad_norm": 7.28125,
      "learning_rate": 2.1486424121221444e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3215096.0,
      "step": 7379
    },
    {
      "epoch": 0.5676486424121221,
      "grad_norm": 30.375,
      "learning_rate": 2.1485270363818167e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3215402.0,
      "step": 7380
    },
    {
      "epoch": 0.5677255595723406,
      "grad_norm": 34.5,
      "learning_rate": 2.1484116606414894e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3215710.0,
      "step": 7381
    },
    {
      "epoch": 0.567802476732559,
      "grad_norm": 19.375,
      "learning_rate": 2.1482962849011612e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 3216290.0,
      "step": 7382
    },
    {
      "epoch": 0.5678793938927775,
      "grad_norm": 45.75,
      "learning_rate": 2.148180909160834e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3216596.0,
      "step": 7383
    },
    {
      "epoch": 0.5679563110529959,
      "grad_norm": 7.21875,
      "learning_rate": 2.148065533420506e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 3217176.0,
      "step": 7384
    },
    {
      "epoch": 0.5680332282132143,
      "grad_norm": 9.0625,
      "learning_rate": 2.1479501576801784e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3217488.0,
      "step": 7385
    },
    {
      "epoch": 0.5681101453734329,
      "grad_norm": 21.125,
      "learning_rate": 2.147834781939851e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 3217816.0,
      "step": 7386
    },
    {
      "epoch": 0.5681870625336513,
      "grad_norm": 27.5,
      "learning_rate": 2.147719406199523e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3218148.0,
      "step": 7387
    },
    {
      "epoch": 0.5682639796938697,
      "grad_norm": 10.1875,
      "learning_rate": 2.147604030459195e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3218696.0,
      "step": 7388
    },
    {
      "epoch": 0.5683408968540882,
      "grad_norm": 7.34375,
      "learning_rate": 2.147488654718868e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3219248.0,
      "step": 7389
    },
    {
      "epoch": 0.5684178140143066,
      "grad_norm": 9.0625,
      "learning_rate": 2.14737327897854e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3220028.0,
      "step": 7390
    },
    {
      "epoch": 0.5684947311745251,
      "grad_norm": 11.125,
      "learning_rate": 2.1472579032382124e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3220356.0,
      "step": 7391
    },
    {
      "epoch": 0.5685716483347435,
      "grad_norm": 6.96875,
      "learning_rate": 2.147142527497885e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3220904.0,
      "step": 7392
    },
    {
      "epoch": 0.5686485654949619,
      "grad_norm": 12.6875,
      "learning_rate": 2.147027151757557e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3221224.0,
      "step": 7393
    },
    {
      "epoch": 0.5687254826551804,
      "grad_norm": 9.3125,
      "learning_rate": 2.1469117760172296e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3221520.0,
      "step": 7394
    },
    {
      "epoch": 0.5688023998153988,
      "grad_norm": 4.90625,
      "learning_rate": 2.146796400276902e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 3222088.0,
      "step": 7395
    },
    {
      "epoch": 0.5688793169756172,
      "grad_norm": 5.0,
      "learning_rate": 2.146681024536574e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3222634.0,
      "step": 7396
    },
    {
      "epoch": 0.5689562341358357,
      "grad_norm": 5.625,
      "learning_rate": 2.1465656487962468e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3223152.0,
      "step": 7397
    },
    {
      "epoch": 0.5690331512960541,
      "grad_norm": 7.3125,
      "learning_rate": 2.1464502730559186e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3223714.0,
      "step": 7398
    },
    {
      "epoch": 0.5691100684562725,
      "grad_norm": 18.625,
      "learning_rate": 2.146334897315591e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3224038.0,
      "step": 7399
    },
    {
      "epoch": 0.5691869856164911,
      "grad_norm": 16.75,
      "learning_rate": 2.1462195215752636e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 3224858.0,
      "step": 7400
    },
    {
      "epoch": 0.5692639027767095,
      "grad_norm": 17.625,
      "learning_rate": 2.146104145834936e-06,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3225178.0,
      "step": 7401
    },
    {
      "epoch": 0.569340819936928,
      "grad_norm": 61.25,
      "learning_rate": 2.145988770094608e-06,
      "loss": 0.0442,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3225436.0,
      "step": 7402
    },
    {
      "epoch": 0.5694177370971464,
      "grad_norm": 22.75,
      "learning_rate": 2.1458733943542808e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3225756.0,
      "step": 7403
    },
    {
      "epoch": 0.5694946542573648,
      "grad_norm": 8.5,
      "learning_rate": 2.1457580186139526e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3226298.0,
      "step": 7404
    },
    {
      "epoch": 0.5695715714175833,
      "grad_norm": 1.640625,
      "learning_rate": 2.1456426428736253e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9974874258041382,
      "num_tokens": 3227096.0,
      "step": 7405
    },
    {
      "epoch": 0.5696484885778017,
      "grad_norm": 15.375,
      "learning_rate": 2.1455272671332975e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3227592.0,
      "step": 7406
    },
    {
      "epoch": 0.5697254057380201,
      "grad_norm": 24.0,
      "learning_rate": 2.1454118913929698e-06,
      "loss": 0.0488,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3227876.0,
      "step": 7407
    },
    {
      "epoch": 0.5698023228982386,
      "grad_norm": 42.75,
      "learning_rate": 2.1452965156526425e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9850187301635742,
      "num_tokens": 3228144.0,
      "step": 7408
    },
    {
      "epoch": 0.569879240058457,
      "grad_norm": 13.75,
      "learning_rate": 2.1451811399123143e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 3228434.0,
      "step": 7409
    },
    {
      "epoch": 0.5699561572186755,
      "grad_norm": 3.390625,
      "learning_rate": 2.1450657641719866e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9974424839019775,
      "num_tokens": 3229218.0,
      "step": 7410
    },
    {
      "epoch": 0.570033074378894,
      "grad_norm": 5.40625,
      "learning_rate": 2.1449503884316592e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9987623691558838,
      "num_tokens": 3230028.0,
      "step": 7411
    },
    {
      "epoch": 0.5701099915391123,
      "grad_norm": 7.34375,
      "learning_rate": 2.1448350126913315e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3230312.0,
      "step": 7412
    },
    {
      "epoch": 0.5701869086993309,
      "grad_norm": 5.34375,
      "learning_rate": 2.1447196369510038e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3230886.0,
      "step": 7413
    },
    {
      "epoch": 0.5702638258595493,
      "grad_norm": 24.625,
      "learning_rate": 2.1446042612106764e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 3231196.0,
      "step": 7414
    },
    {
      "epoch": 0.5703407430197677,
      "grad_norm": 7.375,
      "learning_rate": 2.1444888854703483e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9828178882598877,
      "num_tokens": 3231488.0,
      "step": 7415
    },
    {
      "epoch": 0.5704176601799862,
      "grad_norm": 9.8125,
      "learning_rate": 2.144373509730021e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3231818.0,
      "step": 7416
    },
    {
      "epoch": 0.5704945773402046,
      "grad_norm": 10.125,
      "learning_rate": 2.1442581339896932e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 3232400.0,
      "step": 7417
    },
    {
      "epoch": 0.570571494500423,
      "grad_norm": 5.875,
      "learning_rate": 2.1441427582493655e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9937655925750732,
      "num_tokens": 3233204.0,
      "step": 7418
    },
    {
      "epoch": 0.5706484116606415,
      "grad_norm": 6.0625,
      "learning_rate": 2.144027382509038e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3233550.0,
      "step": 7419
    },
    {
      "epoch": 0.5707253288208599,
      "grad_norm": 9.25,
      "learning_rate": 2.14391200676871e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3233884.0,
      "step": 7420
    },
    {
      "epoch": 0.5708022459810784,
      "grad_norm": 5.1875,
      "learning_rate": 2.1437966310283822e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3234452.0,
      "step": 7421
    },
    {
      "epoch": 0.5708791631412968,
      "grad_norm": 14.4375,
      "learning_rate": 2.143681255288055e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 3234784.0,
      "step": 7422
    },
    {
      "epoch": 0.5709560803015152,
      "grad_norm": 27.375,
      "learning_rate": 2.143565879547727e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3235092.0,
      "step": 7423
    },
    {
      "epoch": 0.5710329974617337,
      "grad_norm": 4.0,
      "learning_rate": 2.1434505038073994e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3235672.0,
      "step": 7424
    },
    {
      "epoch": 0.5711099146219522,
      "grad_norm": 29.625,
      "learning_rate": 2.1433351280670717e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3235978.0,
      "step": 7425
    },
    {
      "epoch": 0.5711868317821706,
      "grad_norm": 45.75,
      "learning_rate": 2.143219752326744e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3236284.0,
      "step": 7426
    },
    {
      "epoch": 0.5712637489423891,
      "grad_norm": 14.4375,
      "learning_rate": 2.1431043765864166e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 3236600.0,
      "step": 7427
    },
    {
      "epoch": 0.5713406661026075,
      "grad_norm": 7.0,
      "learning_rate": 2.142989000846089e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3236908.0,
      "step": 7428
    },
    {
      "epoch": 0.571417583262826,
      "grad_norm": 30.375,
      "learning_rate": 2.142873625105761e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 3237208.0,
      "step": 7429
    },
    {
      "epoch": 0.5714945004230444,
      "grad_norm": 14.875,
      "learning_rate": 2.142758249365434e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3237764.0,
      "step": 7430
    },
    {
      "epoch": 0.5715714175832628,
      "grad_norm": 7.8125,
      "learning_rate": 2.1426428736251057e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 3238336.0,
      "step": 7431
    },
    {
      "epoch": 0.5716483347434813,
      "grad_norm": 12.8125,
      "learning_rate": 2.142527497884778e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 3238920.0,
      "step": 7432
    },
    {
      "epoch": 0.5717252519036997,
      "grad_norm": 11.875,
      "learning_rate": 2.1424121221444506e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9815384745597839,
      "num_tokens": 3239246.0,
      "step": 7433
    },
    {
      "epoch": 0.5718021690639181,
      "grad_norm": 11.5,
      "learning_rate": 2.142296746404123e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 3239564.0,
      "step": 7434
    },
    {
      "epoch": 0.5718790862241366,
      "grad_norm": 11.6875,
      "learning_rate": 2.142181370663795e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3239916.0,
      "step": 7435
    },
    {
      "epoch": 0.571956003384355,
      "grad_norm": 11.875,
      "learning_rate": 2.1420659949234674e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 3240702.0,
      "step": 7436
    },
    {
      "epoch": 0.5720329205445734,
      "grad_norm": 13.8125,
      "learning_rate": 2.1419506191831396e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3241048.0,
      "step": 7437
    },
    {
      "epoch": 0.572109837704792,
      "grad_norm": 11.0,
      "learning_rate": 2.1418352434428123e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 3241606.0,
      "step": 7438
    },
    {
      "epoch": 0.5721867548650104,
      "grad_norm": 8.1875,
      "learning_rate": 2.1417198677024846e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9969879388809204,
      "num_tokens": 3241940.0,
      "step": 7439
    },
    {
      "epoch": 0.5722636720252289,
      "grad_norm": 4.6875,
      "learning_rate": 2.141604491962157e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 3242748.0,
      "step": 7440
    },
    {
      "epoch": 0.5723405891854473,
      "grad_norm": 6.15625,
      "learning_rate": 2.141489116221829e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 3243320.0,
      "step": 7441
    },
    {
      "epoch": 0.5724175063456657,
      "grad_norm": 12.0,
      "learning_rate": 2.1413737404815014e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 3243632.0,
      "step": 7442
    },
    {
      "epoch": 0.5724944235058842,
      "grad_norm": 12.625,
      "learning_rate": 2.1412583647411736e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 3243968.0,
      "step": 7443
    },
    {
      "epoch": 0.5725713406661026,
      "grad_norm": 8.75,
      "learning_rate": 2.1411429890008463e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 3244772.0,
      "step": 7444
    },
    {
      "epoch": 0.572648257826321,
      "grad_norm": 23.75,
      "learning_rate": 2.1410276132605186e-06,
      "loss": 0.0501,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 3245026.0,
      "step": 7445
    },
    {
      "epoch": 0.5727251749865395,
      "grad_norm": 25.5,
      "learning_rate": 2.140912237520191e-06,
      "loss": 0.0427,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3245308.0,
      "step": 7446
    },
    {
      "epoch": 0.5728020921467579,
      "grad_norm": 3.8125,
      "learning_rate": 2.140796861779863e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3245658.0,
      "step": 7447
    },
    {
      "epoch": 0.5728790093069764,
      "grad_norm": 16.0,
      "learning_rate": 2.1406814860395353e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9969040155410767,
      "num_tokens": 3245982.0,
      "step": 7448
    },
    {
      "epoch": 0.5729559264671948,
      "grad_norm": 4.125,
      "learning_rate": 2.1405661102992076e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3246550.0,
      "step": 7449
    },
    {
      "epoch": 0.5730328436274132,
      "grad_norm": 3.71875,
      "learning_rate": 2.1404507345588803e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3247134.0,
      "step": 7450
    },
    {
      "epoch": 0.5731097607876318,
      "grad_norm": 11.9375,
      "learning_rate": 2.1403353588185525e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3247390.0,
      "step": 7451
    },
    {
      "epoch": 0.5731866779478502,
      "grad_norm": 13.1875,
      "learning_rate": 2.140219983078225e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3247952.0,
      "step": 7452
    },
    {
      "epoch": 0.5732635951080686,
      "grad_norm": 15.8125,
      "learning_rate": 2.140104607337897e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3248486.0,
      "step": 7453
    },
    {
      "epoch": 0.5733405122682871,
      "grad_norm": 25.375,
      "learning_rate": 2.1399892315975693e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 3248796.0,
      "step": 7454
    },
    {
      "epoch": 0.5734174294285055,
      "grad_norm": 5.46875,
      "learning_rate": 2.139873855857242e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 3249606.0,
      "step": 7455
    },
    {
      "epoch": 0.5734943465887239,
      "grad_norm": 9.6875,
      "learning_rate": 2.1397584801169142e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3249936.0,
      "step": 7456
    },
    {
      "epoch": 0.5735712637489424,
      "grad_norm": 5.8125,
      "learning_rate": 2.1396431043765865e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3250246.0,
      "step": 7457
    },
    {
      "epoch": 0.5736481809091608,
      "grad_norm": 9.0,
      "learning_rate": 2.1395277286362588e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3250500.0,
      "step": 7458
    },
    {
      "epoch": 0.5737250980693793,
      "grad_norm": 4.90625,
      "learning_rate": 2.139412352895931e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3250806.0,
      "step": 7459
    },
    {
      "epoch": 0.5738020152295977,
      "grad_norm": 21.25,
      "learning_rate": 2.1392969771556033e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3251130.0,
      "step": 7460
    },
    {
      "epoch": 0.5738789323898161,
      "grad_norm": 17.0,
      "learning_rate": 2.139181601415276e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3251448.0,
      "step": 7461
    },
    {
      "epoch": 0.5739558495500346,
      "grad_norm": 36.5,
      "learning_rate": 2.1390662256749482e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3251758.0,
      "step": 7462
    },
    {
      "epoch": 0.574032766710253,
      "grad_norm": 16.875,
      "learning_rate": 2.1389508499346205e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3252110.0,
      "step": 7463
    },
    {
      "epoch": 0.5741096838704715,
      "grad_norm": 10.8125,
      "learning_rate": 2.1388354741942927e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 3252418.0,
      "step": 7464
    },
    {
      "epoch": 0.57418660103069,
      "grad_norm": 7.34375,
      "learning_rate": 2.138720098453965e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3253198.0,
      "step": 7465
    },
    {
      "epoch": 0.5742635181909084,
      "grad_norm": 11.1875,
      "learning_rate": 2.1386047227136377e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3253508.0,
      "step": 7466
    },
    {
      "epoch": 0.5743404353511269,
      "grad_norm": 9.25,
      "learning_rate": 2.13848934697331e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3253778.0,
      "step": 7467
    },
    {
      "epoch": 0.5744173525113453,
      "grad_norm": 10.125,
      "learning_rate": 2.138373971232982e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 3254050.0,
      "step": 7468
    },
    {
      "epoch": 0.5744942696715637,
      "grad_norm": 3.078125,
      "learning_rate": 2.1382585954926544e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3254634.0,
      "step": 7469
    },
    {
      "epoch": 0.5745711868317822,
      "grad_norm": 3.953125,
      "learning_rate": 2.1381432197523267e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3255182.0,
      "step": 7470
    },
    {
      "epoch": 0.5746481039920006,
      "grad_norm": 13.1875,
      "learning_rate": 2.138027844011999e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3255722.0,
      "step": 7471
    },
    {
      "epoch": 0.574725021152219,
      "grad_norm": 2.796875,
      "learning_rate": 2.1379124682716716e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3256274.0,
      "step": 7472
    },
    {
      "epoch": 0.5748019383124375,
      "grad_norm": 25.25,
      "learning_rate": 2.137797092531344e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 3256550.0,
      "step": 7473
    },
    {
      "epoch": 0.5748788554726559,
      "grad_norm": 6.59375,
      "learning_rate": 2.137681716791016e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 3257114.0,
      "step": 7474
    },
    {
      "epoch": 0.5749557726328743,
      "grad_norm": 4.03125,
      "learning_rate": 2.1375663410506884e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 3257412.0,
      "step": 7475
    },
    {
      "epoch": 0.5750326897930929,
      "grad_norm": 9.125,
      "learning_rate": 2.1374509653103607e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3257958.0,
      "step": 7476
    },
    {
      "epoch": 0.5751096069533113,
      "grad_norm": 9.125,
      "learning_rate": 2.1373355895700334e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 3258284.0,
      "step": 7477
    },
    {
      "epoch": 0.5751865241135298,
      "grad_norm": 5.4375,
      "learning_rate": 2.1372202138297056e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3258808.0,
      "step": 7478
    },
    {
      "epoch": 0.5752634412737482,
      "grad_norm": 23.75,
      "learning_rate": 2.137104838089378e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3259090.0,
      "step": 7479
    },
    {
      "epoch": 0.5753403584339666,
      "grad_norm": 15.9375,
      "learning_rate": 2.13698946234905e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3259400.0,
      "step": 7480
    },
    {
      "epoch": 0.5754172755941851,
      "grad_norm": 10.875,
      "learning_rate": 2.1368740866087224e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 3259986.0,
      "step": 7481
    },
    {
      "epoch": 0.5754941927544035,
      "grad_norm": 8.1875,
      "learning_rate": 2.1367587108683947e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 3260772.0,
      "step": 7482
    },
    {
      "epoch": 0.5755711099146219,
      "grad_norm": 13.5,
      "learning_rate": 2.1366433351280673e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9940476417541504,
      "num_tokens": 3261110.0,
      "step": 7483
    },
    {
      "epoch": 0.5756480270748404,
      "grad_norm": 12.5,
      "learning_rate": 2.1365279593877396e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3261442.0,
      "step": 7484
    },
    {
      "epoch": 0.5757249442350588,
      "grad_norm": 16.375,
      "learning_rate": 2.136412583647412e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3261774.0,
      "step": 7485
    },
    {
      "epoch": 0.5758018613952773,
      "grad_norm": 16.625,
      "learning_rate": 2.136297207907084e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3262084.0,
      "step": 7486
    },
    {
      "epoch": 0.5758787785554957,
      "grad_norm": 10.1875,
      "learning_rate": 2.1361818321667564e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 3262342.0,
      "step": 7487
    },
    {
      "epoch": 0.5759556957157141,
      "grad_norm": 4.6875,
      "learning_rate": 2.136066456426429e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3262664.0,
      "step": 7488
    },
    {
      "epoch": 0.5760326128759327,
      "grad_norm": 3.828125,
      "learning_rate": 2.1359510806861013e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3263206.0,
      "step": 7489
    },
    {
      "epoch": 0.5761095300361511,
      "grad_norm": 16.125,
      "learning_rate": 2.135835704945773e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3263538.0,
      "step": 7490
    },
    {
      "epoch": 0.5761864471963695,
      "grad_norm": 11.5625,
      "learning_rate": 2.135720329205446e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3263858.0,
      "step": 7491
    },
    {
      "epoch": 0.576263364356588,
      "grad_norm": 33.25,
      "learning_rate": 2.135604953465118e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3264174.0,
      "step": 7492
    },
    {
      "epoch": 0.5763402815168064,
      "grad_norm": 32.0,
      "learning_rate": 2.1354895777247903e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3264502.0,
      "step": 7493
    },
    {
      "epoch": 0.5764171986770248,
      "grad_norm": 28.0,
      "learning_rate": 2.135374201984463e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3264832.0,
      "step": 7494
    },
    {
      "epoch": 0.5764941158372433,
      "grad_norm": 2.84375,
      "learning_rate": 2.1352588262441353e-06,
      "loss": 0.0061,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 3265876.0,
      "step": 7495
    },
    {
      "epoch": 0.5765710329974617,
      "grad_norm": 11.3125,
      "learning_rate": 2.1351434505038075e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3266394.0,
      "step": 7496
    },
    {
      "epoch": 0.5766479501576802,
      "grad_norm": 6.0625,
      "learning_rate": 2.13502807476348e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9969230890274048,
      "num_tokens": 3266720.0,
      "step": 7497
    },
    {
      "epoch": 0.5767248673178986,
      "grad_norm": 4.15625,
      "learning_rate": 2.134912699023152e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3267266.0,
      "step": 7498
    },
    {
      "epoch": 0.576801784478117,
      "grad_norm": 9.75,
      "learning_rate": 2.1347973232828243e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3267588.0,
      "step": 7499
    },
    {
      "epoch": 0.5768787016383355,
      "grad_norm": 7.8125,
      "learning_rate": 2.134681947542497e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 3267878.0,
      "step": 7500
    },
    {
      "epoch": 0.576955618798554,
      "grad_norm": 17.375,
      "learning_rate": 2.134566571802169e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 3268186.0,
      "step": 7501
    },
    {
      "epoch": 0.5770325359587724,
      "grad_norm": 15.1875,
      "learning_rate": 2.1344511960618415e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3268726.0,
      "step": 7502
    },
    {
      "epoch": 0.5771094531189909,
      "grad_norm": 22.125,
      "learning_rate": 2.1343358203215138e-06,
      "loss": 0.0398,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3268984.0,
      "step": 7503
    },
    {
      "epoch": 0.5771863702792093,
      "grad_norm": 41.5,
      "learning_rate": 2.134220444581186e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3269304.0,
      "step": 7504
    },
    {
      "epoch": 0.5772632874394278,
      "grad_norm": 6.3125,
      "learning_rate": 2.1341050688408587e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3269600.0,
      "step": 7505
    },
    {
      "epoch": 0.5773402045996462,
      "grad_norm": 13.75,
      "learning_rate": 2.133989693100531e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 3269888.0,
      "step": 7506
    },
    {
      "epoch": 0.5774171217598646,
      "grad_norm": 6.1875,
      "learning_rate": 2.133874317360203e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 3270468.0,
      "step": 7507
    },
    {
      "epoch": 0.5774940389200831,
      "grad_norm": 6.375,
      "learning_rate": 2.1337589416198755e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3270776.0,
      "step": 7508
    },
    {
      "epoch": 0.5775709560803015,
      "grad_norm": 12.1875,
      "learning_rate": 2.1336435658795477e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 3271110.0,
      "step": 7509
    },
    {
      "epoch": 0.5776478732405199,
      "grad_norm": 11.9375,
      "learning_rate": 2.13352819013922e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3271452.0,
      "step": 7510
    },
    {
      "epoch": 0.5777247904007384,
      "grad_norm": 48.75,
      "learning_rate": 2.1334128143988927e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3271722.0,
      "step": 7511
    },
    {
      "epoch": 0.5778017075609568,
      "grad_norm": 20.375,
      "learning_rate": 2.1332974386585645e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 3272510.0,
      "step": 7512
    },
    {
      "epoch": 0.5778786247211753,
      "grad_norm": 28.5,
      "learning_rate": 2.133182062918237e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 3272810.0,
      "step": 7513
    },
    {
      "epoch": 0.5779555418813938,
      "grad_norm": 8.875,
      "learning_rate": 2.1330666871779095e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3273120.0,
      "step": 7514
    },
    {
      "epoch": 0.5780324590416122,
      "grad_norm": 22.5,
      "learning_rate": 2.1329513114375817e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3273438.0,
      "step": 7515
    },
    {
      "epoch": 0.5781093762018307,
      "grad_norm": 10.8125,
      "learning_rate": 2.1328359356972544e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 3273786.0,
      "step": 7516
    },
    {
      "epoch": 0.5781862933620491,
      "grad_norm": 7.9375,
      "learning_rate": 2.1327205599569266e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3274332.0,
      "step": 7517
    },
    {
      "epoch": 0.5782632105222675,
      "grad_norm": 23.375,
      "learning_rate": 2.1326051842165985e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3274866.0,
      "step": 7518
    },
    {
      "epoch": 0.578340127682486,
      "grad_norm": 11.8125,
      "learning_rate": 2.132489808476271e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3275152.0,
      "step": 7519
    },
    {
      "epoch": 0.5784170448427044,
      "grad_norm": 13.6875,
      "learning_rate": 2.1323744327359434e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3275932.0,
      "step": 7520
    },
    {
      "epoch": 0.5784939620029228,
      "grad_norm": 8.625,
      "learning_rate": 2.1322590569956157e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3276518.0,
      "step": 7521
    },
    {
      "epoch": 0.5785708791631413,
      "grad_norm": 29.375,
      "learning_rate": 2.1321436812552884e-06,
      "loss": 0.0369,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3276774.0,
      "step": 7522
    },
    {
      "epoch": 0.5786477963233597,
      "grad_norm": 39.25,
      "learning_rate": 2.13202830551496e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 3277070.0,
      "step": 7523
    },
    {
      "epoch": 0.5787247134835782,
      "grad_norm": 23.25,
      "learning_rate": 2.131912929774633e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.9771241545677185,
      "num_tokens": 3277378.0,
      "step": 7524
    },
    {
      "epoch": 0.5788016306437966,
      "grad_norm": 5.0625,
      "learning_rate": 2.131797554034305e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3277952.0,
      "step": 7525
    },
    {
      "epoch": 0.578878547804015,
      "grad_norm": 69.5,
      "learning_rate": 2.1316821782939774e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3278282.0,
      "step": 7526
    },
    {
      "epoch": 0.5789554649642336,
      "grad_norm": 6.21875,
      "learning_rate": 2.13156680255365e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3278856.0,
      "step": 7527
    },
    {
      "epoch": 0.579032382124452,
      "grad_norm": 14.75,
      "learning_rate": 2.131451426813322e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 3279372.0,
      "step": 7528
    },
    {
      "epoch": 0.5791092992846704,
      "grad_norm": 52.25,
      "learning_rate": 2.131336051072994e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 3279740.0,
      "step": 7529
    },
    {
      "epoch": 0.5791862164448889,
      "grad_norm": 81.5,
      "learning_rate": 2.131220675332667e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3280042.0,
      "step": 7530
    },
    {
      "epoch": 0.5792631336051073,
      "grad_norm": 38.75,
      "learning_rate": 2.131105299592339e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3280370.0,
      "step": 7531
    },
    {
      "epoch": 0.5793400507653258,
      "grad_norm": 12.75,
      "learning_rate": 2.1309899238520114e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3280682.0,
      "step": 7532
    },
    {
      "epoch": 0.5794169679255442,
      "grad_norm": 34.75,
      "learning_rate": 2.130874548111684e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3281028.0,
      "step": 7533
    },
    {
      "epoch": 0.5794938850857626,
      "grad_norm": 31.625,
      "learning_rate": 2.130759172371356e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 3281384.0,
      "step": 7534
    },
    {
      "epoch": 0.5795708022459811,
      "grad_norm": 16.625,
      "learning_rate": 2.1306437966310286e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3281924.0,
      "step": 7535
    },
    {
      "epoch": 0.5796477194061995,
      "grad_norm": 68.5,
      "learning_rate": 2.130528420890701e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3282224.0,
      "step": 7536
    },
    {
      "epoch": 0.5797246365664179,
      "grad_norm": 42.5,
      "learning_rate": 2.130413045150373e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 3282526.0,
      "step": 7537
    },
    {
      "epoch": 0.5798015537266364,
      "grad_norm": 21.5,
      "learning_rate": 2.1302976694100458e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3282876.0,
      "step": 7538
    },
    {
      "epoch": 0.5798784708868548,
      "grad_norm": 29.125,
      "learning_rate": 2.1301822936697176e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3283212.0,
      "step": 7539
    },
    {
      "epoch": 0.5799553880470733,
      "grad_norm": 12.4375,
      "learning_rate": 2.13006691792939e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3283760.0,
      "step": 7540
    },
    {
      "epoch": 0.5800323052072918,
      "grad_norm": 28.875,
      "learning_rate": 2.1299515421890625e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3284044.0,
      "step": 7541
    },
    {
      "epoch": 0.5801092223675102,
      "grad_norm": 18.875,
      "learning_rate": 2.129836166448735e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 3284384.0,
      "step": 7542
    },
    {
      "epoch": 0.5801861395277287,
      "grad_norm": 30.25,
      "learning_rate": 2.129720790708407e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3284666.0,
      "step": 7543
    },
    {
      "epoch": 0.5802630566879471,
      "grad_norm": 14.4375,
      "learning_rate": 2.1296054149680797e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 3285208.0,
      "step": 7544
    },
    {
      "epoch": 0.5803399738481655,
      "grad_norm": 7.71875,
      "learning_rate": 2.1294900392277516e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 3285516.0,
      "step": 7545
    },
    {
      "epoch": 0.580416891008384,
      "grad_norm": 23.375,
      "learning_rate": 2.1293746634874243e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3286050.0,
      "step": 7546
    },
    {
      "epoch": 0.5804938081686024,
      "grad_norm": 34.0,
      "learning_rate": 2.1292592877470965e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3286372.0,
      "step": 7547
    },
    {
      "epoch": 0.5805707253288208,
      "grad_norm": 3.5,
      "learning_rate": 2.1291439120067688e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3286914.0,
      "step": 7548
    },
    {
      "epoch": 0.5806476424890393,
      "grad_norm": 44.75,
      "learning_rate": 2.129028536266441e-06,
      "loss": 0.039,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3287196.0,
      "step": 7549
    },
    {
      "epoch": 0.5807245596492577,
      "grad_norm": 8.5625,
      "learning_rate": 2.1289131605261133e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 3288004.0,
      "step": 7550
    },
    {
      "epoch": 0.5808014768094762,
      "grad_norm": 17.75,
      "learning_rate": 2.1287977847857855e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9788732528686523,
      "num_tokens": 3288290.0,
      "step": 7551
    },
    {
      "epoch": 0.5808783939696947,
      "grad_norm": 52.25,
      "learning_rate": 2.1286824090454582e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3288606.0,
      "step": 7552
    },
    {
      "epoch": 0.580955311129913,
      "grad_norm": 7.34375,
      "learning_rate": 2.1285670333051305e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 3289166.0,
      "step": 7553
    },
    {
      "epoch": 0.5810322282901316,
      "grad_norm": 26.25,
      "learning_rate": 2.1284516575648027e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3289476.0,
      "step": 7554
    },
    {
      "epoch": 0.58110914545035,
      "grad_norm": 5.90625,
      "learning_rate": 2.1283362818244754e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3289800.0,
      "step": 7555
    },
    {
      "epoch": 0.5811860626105684,
      "grad_norm": 2.8125,
      "learning_rate": 2.1282209060841473e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 3290324.0,
      "step": 7556
    },
    {
      "epoch": 0.5812629797707869,
      "grad_norm": 4.59375,
      "learning_rate": 2.1281055303438195e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9962335228919983,
      "num_tokens": 3290856.0,
      "step": 7557
    },
    {
      "epoch": 0.5813398969310053,
      "grad_norm": 9.1875,
      "learning_rate": 2.127990154603492e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9971346855163574,
      "num_tokens": 3291206.0,
      "step": 7558
    },
    {
      "epoch": 0.5814168140912237,
      "grad_norm": 42.5,
      "learning_rate": 2.1278747788631645e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 3291546.0,
      "step": 7559
    },
    {
      "epoch": 0.5814937312514422,
      "grad_norm": 12.1875,
      "learning_rate": 2.1277594031228367e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3292086.0,
      "step": 7560
    },
    {
      "epoch": 0.5815706484116606,
      "grad_norm": 7.65625,
      "learning_rate": 2.127644027382509e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 3292434.0,
      "step": 7561
    },
    {
      "epoch": 0.5816475655718791,
      "grad_norm": 15.6875,
      "learning_rate": 2.1275286516421812e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 3292812.0,
      "step": 7562
    },
    {
      "epoch": 0.5817244827320975,
      "grad_norm": 13.5625,
      "learning_rate": 2.127413275901854e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3293134.0,
      "step": 7563
    },
    {
      "epoch": 0.5818013998923159,
      "grad_norm": 4.625,
      "learning_rate": 2.127297900161526e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3293708.0,
      "step": 7564
    },
    {
      "epoch": 0.5818783170525345,
      "grad_norm": 11.4375,
      "learning_rate": 2.1271825244211984e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 3294016.0,
      "step": 7565
    },
    {
      "epoch": 0.5819552342127529,
      "grad_norm": 9.3125,
      "learning_rate": 2.1270671486808707e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3294366.0,
      "step": 7566
    },
    {
      "epoch": 0.5820321513729713,
      "grad_norm": 22.375,
      "learning_rate": 2.126951772940543e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3294686.0,
      "step": 7567
    },
    {
      "epoch": 0.5821090685331898,
      "grad_norm": 52.0,
      "learning_rate": 2.126836397200215e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3294956.0,
      "step": 7568
    },
    {
      "epoch": 0.5821859856934082,
      "grad_norm": 21.875,
      "learning_rate": 2.126721021459888e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3295298.0,
      "step": 7569
    },
    {
      "epoch": 0.5822629028536267,
      "grad_norm": 10.6875,
      "learning_rate": 2.12660564571956e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 3295604.0,
      "step": 7570
    },
    {
      "epoch": 0.5823398200138451,
      "grad_norm": 4.46875,
      "learning_rate": 2.1264902699792324e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3296148.0,
      "step": 7571
    },
    {
      "epoch": 0.5824167371740635,
      "grad_norm": 38.75,
      "learning_rate": 2.1263748942389047e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3296478.0,
      "step": 7572
    },
    {
      "epoch": 0.582493654334282,
      "grad_norm": 8.875,
      "learning_rate": 2.126259518498577e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9966044425964355,
      "num_tokens": 3297068.0,
      "step": 7573
    },
    {
      "epoch": 0.5825705714945004,
      "grad_norm": 11.3125,
      "learning_rate": 2.1261441427582496e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3297390.0,
      "step": 7574
    },
    {
      "epoch": 0.5826474886547188,
      "grad_norm": 18.875,
      "learning_rate": 2.126028767017922e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3297912.0,
      "step": 7575
    },
    {
      "epoch": 0.5827244058149373,
      "grad_norm": 7.125,
      "learning_rate": 2.125913391277594e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3298498.0,
      "step": 7576
    },
    {
      "epoch": 0.5828013229751557,
      "grad_norm": 16.375,
      "learning_rate": 2.1257980155372664e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3299042.0,
      "step": 7577
    },
    {
      "epoch": 0.5828782401353741,
      "grad_norm": 4.75,
      "learning_rate": 2.1256826397969386e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3299628.0,
      "step": 7578
    },
    {
      "epoch": 0.5829551572955927,
      "grad_norm": 17.0,
      "learning_rate": 2.125567264056611e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 3300188.0,
      "step": 7579
    },
    {
      "epoch": 0.5830320744558111,
      "grad_norm": 2.78125,
      "learning_rate": 2.1254518883162836e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3300774.0,
      "step": 7580
    },
    {
      "epoch": 0.5831089916160296,
      "grad_norm": 28.125,
      "learning_rate": 2.125336512575956e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 3301114.0,
      "step": 7581
    },
    {
      "epoch": 0.583185908776248,
      "grad_norm": 3.546875,
      "learning_rate": 2.125221136835628e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3301894.0,
      "step": 7582
    },
    {
      "epoch": 0.5832628259364664,
      "grad_norm": 38.75,
      "learning_rate": 2.1251057610953003e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3302214.0,
      "step": 7583
    },
    {
      "epoch": 0.5833397430966849,
      "grad_norm": 3.40625,
      "learning_rate": 2.1249903853549726e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 3302998.0,
      "step": 7584
    },
    {
      "epoch": 0.5834166602569033,
      "grad_norm": 4.25,
      "learning_rate": 2.1248750096146453e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3303756.0,
      "step": 7585
    },
    {
      "epoch": 0.5834935774171217,
      "grad_norm": 5.8125,
      "learning_rate": 2.1247596338743175e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3304300.0,
      "step": 7586
    },
    {
      "epoch": 0.5835704945773402,
      "grad_norm": 22.375,
      "learning_rate": 2.12464425813399e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3304618.0,
      "step": 7587
    },
    {
      "epoch": 0.5836474117375586,
      "grad_norm": 25.5,
      "learning_rate": 2.124528882393662e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 3305398.0,
      "step": 7588
    },
    {
      "epoch": 0.5837243288977771,
      "grad_norm": 8.75,
      "learning_rate": 2.1244135066533343e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3305720.0,
      "step": 7589
    },
    {
      "epoch": 0.5838012460579955,
      "grad_norm": 25.5,
      "learning_rate": 2.1242981309130066e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3306002.0,
      "step": 7590
    },
    {
      "epoch": 0.583878163218214,
      "grad_norm": 4.53125,
      "learning_rate": 2.1241827551726793e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3306564.0,
      "step": 7591
    },
    {
      "epoch": 0.5839550803784325,
      "grad_norm": 21.875,
      "learning_rate": 2.1240673794323515e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 3306882.0,
      "step": 7592
    },
    {
      "epoch": 0.5840319975386509,
      "grad_norm": 9.5,
      "learning_rate": 2.1239520036920238e-06,
      "loss": 0.0054,
      "mean_token_accuracy": 0.9974293112754822,
      "num_tokens": 3307662.0,
      "step": 7593
    },
    {
      "epoch": 0.5841089146988693,
      "grad_norm": 8.875,
      "learning_rate": 2.123836627951696e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 3308012.0,
      "step": 7594
    },
    {
      "epoch": 0.5841858318590878,
      "grad_norm": 121.5,
      "learning_rate": 2.1237212522113683e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3308318.0,
      "step": 7595
    },
    {
      "epoch": 0.5842627490193062,
      "grad_norm": 35.75,
      "learning_rate": 2.123605876471041e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3308624.0,
      "step": 7596
    },
    {
      "epoch": 0.5843396661795246,
      "grad_norm": 3.25,
      "learning_rate": 2.1234905007307132e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3309180.0,
      "step": 7597
    },
    {
      "epoch": 0.5844165833397431,
      "grad_norm": 5.09375,
      "learning_rate": 2.1233751249903855e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9981651306152344,
      "num_tokens": 3309726.0,
      "step": 7598
    },
    {
      "epoch": 0.5844935004999615,
      "grad_norm": 43.25,
      "learning_rate": 2.1232597492500577e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3310260.0,
      "step": 7599
    },
    {
      "epoch": 0.58457041766018,
      "grad_norm": 25.875,
      "learning_rate": 2.12314437350973e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3310824.0,
      "step": 7600
    },
    {
      "epoch": 0.5846473348203984,
      "grad_norm": 47.5,
      "learning_rate": 2.1230289977694023e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 3311134.0,
      "step": 7601
    },
    {
      "epoch": 0.5847242519806168,
      "grad_norm": 48.75,
      "learning_rate": 2.122913622029075e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3311430.0,
      "step": 7602
    },
    {
      "epoch": 0.5848011691408354,
      "grad_norm": 19.0,
      "learning_rate": 2.122798246288747e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3311716.0,
      "step": 7603
    },
    {
      "epoch": 0.5848780863010538,
      "grad_norm": 30.625,
      "learning_rate": 2.1226828705484195e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3312250.0,
      "step": 7604
    },
    {
      "epoch": 0.5849550034612722,
      "grad_norm": 8.375,
      "learning_rate": 2.1225674948080917e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 3312566.0,
      "step": 7605
    },
    {
      "epoch": 0.5850319206214907,
      "grad_norm": 18.0,
      "learning_rate": 2.122452119067764e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3312896.0,
      "step": 7606
    },
    {
      "epoch": 0.5851088377817091,
      "grad_norm": 56.5,
      "learning_rate": 2.1223367433274362e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3313190.0,
      "step": 7607
    },
    {
      "epoch": 0.5851857549419276,
      "grad_norm": 10.75,
      "learning_rate": 2.122221367587109e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 3313530.0,
      "step": 7608
    },
    {
      "epoch": 0.585262672102146,
      "grad_norm": 16.625,
      "learning_rate": 2.122105991846781e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3313846.0,
      "step": 7609
    },
    {
      "epoch": 0.5853395892623644,
      "grad_norm": 14.0625,
      "learning_rate": 2.1219906161064534e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3314392.0,
      "step": 7610
    },
    {
      "epoch": 0.5854165064225829,
      "grad_norm": 17.125,
      "learning_rate": 2.1218752403661257e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3314674.0,
      "step": 7611
    },
    {
      "epoch": 0.5854934235828013,
      "grad_norm": 34.25,
      "learning_rate": 2.121759864625798e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3315006.0,
      "step": 7612
    },
    {
      "epoch": 0.5855703407430197,
      "grad_norm": 21.125,
      "learning_rate": 2.1216444888854706e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 3315366.0,
      "step": 7613
    },
    {
      "epoch": 0.5856472579032382,
      "grad_norm": 65.0,
      "learning_rate": 2.121529113145143e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3315648.0,
      "step": 7614
    },
    {
      "epoch": 0.5857241750634566,
      "grad_norm": 20.125,
      "learning_rate": 2.1214137374048147e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 3316014.0,
      "step": 7615
    },
    {
      "epoch": 0.5858010922236752,
      "grad_norm": 5.21875,
      "learning_rate": 2.1212983616644874e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 3316590.0,
      "step": 7616
    },
    {
      "epoch": 0.5858780093838936,
      "grad_norm": 8.75,
      "learning_rate": 2.1211829859241597e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3317174.0,
      "step": 7617
    },
    {
      "epoch": 0.585954926544112,
      "grad_norm": 11.8125,
      "learning_rate": 2.121067610183832e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9969040155410767,
      "num_tokens": 3317498.0,
      "step": 7618
    },
    {
      "epoch": 0.5860318437043305,
      "grad_norm": 4.25,
      "learning_rate": 2.1209522344435046e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3318040.0,
      "step": 7619
    },
    {
      "epoch": 0.5861087608645489,
      "grad_norm": 32.25,
      "learning_rate": 2.120836858703177e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3318372.0,
      "step": 7620
    },
    {
      "epoch": 0.5861856780247673,
      "grad_norm": 41.5,
      "learning_rate": 2.120721482962849e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 3318716.0,
      "step": 7621
    },
    {
      "epoch": 0.5862625951849858,
      "grad_norm": 18.25,
      "learning_rate": 2.1206061072225214e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 3319288.0,
      "step": 7622
    },
    {
      "epoch": 0.5863395123452042,
      "grad_norm": 6.1875,
      "learning_rate": 2.1204907314821936e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3319868.0,
      "step": 7623
    },
    {
      "epoch": 0.5864164295054226,
      "grad_norm": 15.9375,
      "learning_rate": 2.1203753557418663e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3320432.0,
      "step": 7624
    },
    {
      "epoch": 0.5864933466656411,
      "grad_norm": 19.75,
      "learning_rate": 2.1202599800015386e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3321006.0,
      "step": 7625
    },
    {
      "epoch": 0.5865702638258595,
      "grad_norm": 75.0,
      "learning_rate": 2.1201446042612104e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3321286.0,
      "step": 7626
    },
    {
      "epoch": 0.586647180986078,
      "grad_norm": 19.125,
      "learning_rate": 2.120029228520883e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 3321634.0,
      "step": 7627
    },
    {
      "epoch": 0.5867240981462964,
      "grad_norm": 6.6875,
      "learning_rate": 2.1199138527805553e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 3321962.0,
      "step": 7628
    },
    {
      "epoch": 0.5868010153065149,
      "grad_norm": 17.125,
      "learning_rate": 2.1197984770402276e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 3322292.0,
      "step": 7629
    },
    {
      "epoch": 0.5868779324667334,
      "grad_norm": 36.5,
      "learning_rate": 2.1196831012999003e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 3323102.0,
      "step": 7630
    },
    {
      "epoch": 0.5869548496269518,
      "grad_norm": 117.0,
      "learning_rate": 2.119567725559572e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 3323438.0,
      "step": 7631
    },
    {
      "epoch": 0.5870317667871702,
      "grad_norm": 11.875,
      "learning_rate": 2.119452349819245e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3323986.0,
      "step": 7632
    },
    {
      "epoch": 0.5871086839473887,
      "grad_norm": 16.25,
      "learning_rate": 2.119336974078917e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3324532.0,
      "step": 7633
    },
    {
      "epoch": 0.5871856011076071,
      "grad_norm": 36.0,
      "learning_rate": 2.1192215983385893e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3325128.0,
      "step": 7634
    },
    {
      "epoch": 0.5872625182678256,
      "grad_norm": 18.0,
      "learning_rate": 2.119106222598262e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3325444.0,
      "step": 7635
    },
    {
      "epoch": 0.587339435428044,
      "grad_norm": 17.25,
      "learning_rate": 2.1189908468579343e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3325788.0,
      "step": 7636
    },
    {
      "epoch": 0.5874163525882624,
      "grad_norm": 28.625,
      "learning_rate": 2.118875471117606e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3326152.0,
      "step": 7637
    },
    {
      "epoch": 0.5874932697484809,
      "grad_norm": 26.625,
      "learning_rate": 2.1187600953772788e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3326696.0,
      "step": 7638
    },
    {
      "epoch": 0.5875701869086993,
      "grad_norm": 8.625,
      "learning_rate": 2.118644719636951e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3327020.0,
      "step": 7639
    },
    {
      "epoch": 0.5876471040689177,
      "grad_norm": 15.8125,
      "learning_rate": 2.1185293438966233e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3327604.0,
      "step": 7640
    },
    {
      "epoch": 0.5877240212291363,
      "grad_norm": 10.4375,
      "learning_rate": 2.118413968156296e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3328188.0,
      "step": 7641
    },
    {
      "epoch": 0.5878009383893547,
      "grad_norm": 73.0,
      "learning_rate": 2.118298592415968e-06,
      "loss": 0.047,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 3328446.0,
      "step": 7642
    },
    {
      "epoch": 0.5878778555495731,
      "grad_norm": 8.9375,
      "learning_rate": 2.1181832166756405e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3329008.0,
      "step": 7643
    },
    {
      "epoch": 0.5879547727097916,
      "grad_norm": 23.125,
      "learning_rate": 2.1180678409353127e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9971671104431152,
      "num_tokens": 3329362.0,
      "step": 7644
    },
    {
      "epoch": 0.58803168987001,
      "grad_norm": 7.78125,
      "learning_rate": 2.117952465194985e-06,
      "loss": 0.0062,
      "mean_token_accuracy": 0.9961722493171692,
      "num_tokens": 3330410.0,
      "step": 7645
    },
    {
      "epoch": 0.5881086070302285,
      "grad_norm": 89.0,
      "learning_rate": 2.1178370894546577e-06,
      "loss": 0.0524,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3330690.0,
      "step": 7646
    },
    {
      "epoch": 0.5881855241904469,
      "grad_norm": 31.75,
      "learning_rate": 2.11772171371433e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3331002.0,
      "step": 7647
    },
    {
      "epoch": 0.5882624413506653,
      "grad_norm": 49.5,
      "learning_rate": 2.1176063379740018e-06,
      "loss": 0.0439,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3331258.0,
      "step": 7648
    },
    {
      "epoch": 0.5883393585108838,
      "grad_norm": 9.375,
      "learning_rate": 2.1174909622336745e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 3331592.0,
      "step": 7649
    },
    {
      "epoch": 0.5884162756711022,
      "grad_norm": 79.5,
      "learning_rate": 2.1173755864933467e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3331902.0,
      "step": 7650
    },
    {
      "epoch": 0.5884931928313206,
      "grad_norm": 6.40625,
      "learning_rate": 2.117260210753019e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 3332422.0,
      "step": 7651
    },
    {
      "epoch": 0.5885701099915391,
      "grad_norm": 13.1875,
      "learning_rate": 2.1171448350126917e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3332702.0,
      "step": 7652
    },
    {
      "epoch": 0.5886470271517575,
      "grad_norm": 7.40625,
      "learning_rate": 2.1170294592723635e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3332986.0,
      "step": 7653
    },
    {
      "epoch": 0.588723944311976,
      "grad_norm": 12.3125,
      "learning_rate": 2.116914083532036e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9975308775901794,
      "num_tokens": 3333798.0,
      "step": 7654
    },
    {
      "epoch": 0.5888008614721945,
      "grad_norm": 5.84375,
      "learning_rate": 2.1167987077917084e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 3334340.0,
      "step": 7655
    },
    {
      "epoch": 0.5888777786324129,
      "grad_norm": 9.875,
      "learning_rate": 2.1166833320513807e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 3334670.0,
      "step": 7656
    },
    {
      "epoch": 0.5889546957926314,
      "grad_norm": 5.1875,
      "learning_rate": 2.116567956311053e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3335210.0,
      "step": 7657
    },
    {
      "epoch": 0.5890316129528498,
      "grad_norm": 9.1875,
      "learning_rate": 2.1164525805707256e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 3335810.0,
      "step": 7658
    },
    {
      "epoch": 0.5891085301130682,
      "grad_norm": 12.875,
      "learning_rate": 2.1163372048303975e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3336134.0,
      "step": 7659
    },
    {
      "epoch": 0.5891854472732867,
      "grad_norm": 7.59375,
      "learning_rate": 2.11622182909007e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3336440.0,
      "step": 7660
    },
    {
      "epoch": 0.5892623644335051,
      "grad_norm": 16.125,
      "learning_rate": 2.1161064533497424e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 3336798.0,
      "step": 7661
    },
    {
      "epoch": 0.5893392815937235,
      "grad_norm": 11.0625,
      "learning_rate": 2.1159910776094147e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3337142.0,
      "step": 7662
    },
    {
      "epoch": 0.589416198753942,
      "grad_norm": 10.9375,
      "learning_rate": 2.1158757018690873e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 3337430.0,
      "step": 7663
    },
    {
      "epoch": 0.5894931159141604,
      "grad_norm": 36.0,
      "learning_rate": 2.115760326128759e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 3337764.0,
      "step": 7664
    },
    {
      "epoch": 0.5895700330743789,
      "grad_norm": 10.6875,
      "learning_rate": 2.1156449503884314e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3338034.0,
      "step": 7665
    },
    {
      "epoch": 0.5896469502345973,
      "grad_norm": 5.625,
      "learning_rate": 2.115529574648104e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 3338596.0,
      "step": 7666
    },
    {
      "epoch": 0.5897238673948157,
      "grad_norm": 31.75,
      "learning_rate": 2.1154141989077764e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 3338934.0,
      "step": 7667
    },
    {
      "epoch": 0.5898007845550343,
      "grad_norm": 17.5,
      "learning_rate": 2.1152988231674486e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 3339506.0,
      "step": 7668
    },
    {
      "epoch": 0.5898777017152527,
      "grad_norm": 5.40625,
      "learning_rate": 2.115183447427121e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3340064.0,
      "step": 7669
    },
    {
      "epoch": 0.5899546188754711,
      "grad_norm": 27.875,
      "learning_rate": 2.115068071686793e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 3340424.0,
      "step": 7670
    },
    {
      "epoch": 0.5900315360356896,
      "grad_norm": 6.125,
      "learning_rate": 2.114952695946466e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 3341210.0,
      "step": 7671
    },
    {
      "epoch": 0.590108453195908,
      "grad_norm": 12.625,
      "learning_rate": 2.114837320206138e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3341774.0,
      "step": 7672
    },
    {
      "epoch": 0.5901853703561265,
      "grad_norm": 43.5,
      "learning_rate": 2.1147219444658103e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3342086.0,
      "step": 7673
    },
    {
      "epoch": 0.5902622875163449,
      "grad_norm": 3.796875,
      "learning_rate": 2.114606568725483e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9961758852005005,
      "num_tokens": 3342610.0,
      "step": 7674
    },
    {
      "epoch": 0.5903392046765633,
      "grad_norm": 3.21875,
      "learning_rate": 2.114491192985155e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 3343184.0,
      "step": 7675
    },
    {
      "epoch": 0.5904161218367818,
      "grad_norm": 26.5,
      "learning_rate": 2.114375817244827e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3343496.0,
      "step": 7676
    },
    {
      "epoch": 0.5904930389970002,
      "grad_norm": 78.0,
      "learning_rate": 2.1142604415045e-06,
      "loss": 0.0397,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 3343774.0,
      "step": 7677
    },
    {
      "epoch": 0.5905699561572186,
      "grad_norm": 9.125,
      "learning_rate": 2.114145065764172e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3344316.0,
      "step": 7678
    },
    {
      "epoch": 0.5906468733174371,
      "grad_norm": 22.875,
      "learning_rate": 2.1140296900238443e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3344620.0,
      "step": 7679
    },
    {
      "epoch": 0.5907237904776556,
      "grad_norm": 14.875,
      "learning_rate": 2.1139143142835166e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 3344972.0,
      "step": 7680
    },
    {
      "epoch": 0.590800707637874,
      "grad_norm": 15.6875,
      "learning_rate": 2.113798938543189e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 3345792.0,
      "step": 7681
    },
    {
      "epoch": 0.5908776247980925,
      "grad_norm": 4.25,
      "learning_rate": 2.1136835628028615e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 3346102.0,
      "step": 7682
    },
    {
      "epoch": 0.5909545419583109,
      "grad_norm": 7.84375,
      "learning_rate": 2.1135681870625338e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3346686.0,
      "step": 7683
    },
    {
      "epoch": 0.5910314591185294,
      "grad_norm": 4.34375,
      "learning_rate": 2.113452811322206e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3347232.0,
      "step": 7684
    },
    {
      "epoch": 0.5911083762787478,
      "grad_norm": 9.3125,
      "learning_rate": 2.1133374355818787e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 3347522.0,
      "step": 7685
    },
    {
      "epoch": 0.5911852934389662,
      "grad_norm": 18.125,
      "learning_rate": 2.1132220598415506e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 3347882.0,
      "step": 7686
    },
    {
      "epoch": 0.5912622105991847,
      "grad_norm": 10.125,
      "learning_rate": 2.113106684101223e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9855907559394836,
      "num_tokens": 3348230.0,
      "step": 7687
    },
    {
      "epoch": 0.5913391277594031,
      "grad_norm": 7.5625,
      "learning_rate": 2.1129913083608955e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3348746.0,
      "step": 7688
    },
    {
      "epoch": 0.5914160449196215,
      "grad_norm": 20.25,
      "learning_rate": 2.1128759326205677e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 3349278.0,
      "step": 7689
    },
    {
      "epoch": 0.59149296207984,
      "grad_norm": 10.875,
      "learning_rate": 2.11276055688024e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3349624.0,
      "step": 7690
    },
    {
      "epoch": 0.5915698792400584,
      "grad_norm": 37.25,
      "learning_rate": 2.1126451811399123e-06,
      "loss": 0.0425,
      "mean_token_accuracy": 0.97826087474823,
      "num_tokens": 3349902.0,
      "step": 7691
    },
    {
      "epoch": 0.591646796400277,
      "grad_norm": 21.625,
      "learning_rate": 2.1125298053995845e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3350224.0,
      "step": 7692
    },
    {
      "epoch": 0.5917237135604954,
      "grad_norm": 9.5625,
      "learning_rate": 2.112414429659257e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 3350596.0,
      "step": 7693
    },
    {
      "epoch": 0.5918006307207138,
      "grad_norm": 28.625,
      "learning_rate": 2.1122990539189295e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3350920.0,
      "step": 7694
    },
    {
      "epoch": 0.5918775478809323,
      "grad_norm": 5.03125,
      "learning_rate": 2.1121836781786017e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 3351500.0,
      "step": 7695
    },
    {
      "epoch": 0.5919544650411507,
      "grad_norm": 10.4375,
      "learning_rate": 2.112068302438274e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3351864.0,
      "step": 7696
    },
    {
      "epoch": 0.5920313822013691,
      "grad_norm": 8.1875,
      "learning_rate": 2.1119529266979462e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 3352212.0,
      "step": 7697
    },
    {
      "epoch": 0.5921082993615876,
      "grad_norm": 33.75,
      "learning_rate": 2.1118375509576185e-06,
      "loss": 0.0431,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3352490.0,
      "step": 7698
    },
    {
      "epoch": 0.592185216521806,
      "grad_norm": 11.625,
      "learning_rate": 2.111722175217291e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3353024.0,
      "step": 7699
    },
    {
      "epoch": 0.5922621336820244,
      "grad_norm": 82.0,
      "learning_rate": 2.1116067994769634e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3353352.0,
      "step": 7700
    },
    {
      "epoch": 0.5923390508422429,
      "grad_norm": 16.875,
      "learning_rate": 2.1114914237366357e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3353680.0,
      "step": 7701
    },
    {
      "epoch": 0.5924159680024613,
      "grad_norm": 11.375,
      "learning_rate": 2.111376047996308e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 3353990.0,
      "step": 7702
    },
    {
      "epoch": 0.5924928851626798,
      "grad_norm": 13.5625,
      "learning_rate": 2.11126067225598e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 3354316.0,
      "step": 7703
    },
    {
      "epoch": 0.5925698023228982,
      "grad_norm": 17.375,
      "learning_rate": 2.111145296515653e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 3354916.0,
      "step": 7704
    },
    {
      "epoch": 0.5926467194831166,
      "grad_norm": 5.8125,
      "learning_rate": 2.111029920775325e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3355456.0,
      "step": 7705
    },
    {
      "epoch": 0.5927236366433352,
      "grad_norm": 10.375,
      "learning_rate": 2.1109145450349974e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 3355780.0,
      "step": 7706
    },
    {
      "epoch": 0.5928005538035536,
      "grad_norm": 10.5625,
      "learning_rate": 2.1107991692946697e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9969230890274048,
      "num_tokens": 3356106.0,
      "step": 7707
    },
    {
      "epoch": 0.592877470963772,
      "grad_norm": 37.0,
      "learning_rate": 2.110683793554342e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3356418.0,
      "step": 7708
    },
    {
      "epoch": 0.5929543881239905,
      "grad_norm": 6.9375,
      "learning_rate": 2.110568417814014e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 3356998.0,
      "step": 7709
    },
    {
      "epoch": 0.5930313052842089,
      "grad_norm": 6.9375,
      "learning_rate": 2.110453042073687e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 3357600.0,
      "step": 7710
    },
    {
      "epoch": 0.5931082224444274,
      "grad_norm": 11.25,
      "learning_rate": 2.110337666333359e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9968847632408142,
      "num_tokens": 3357922.0,
      "step": 7711
    },
    {
      "epoch": 0.5931851396046458,
      "grad_norm": 15.25,
      "learning_rate": 2.1102222905930314e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9909256100654602,
      "num_tokens": 3358474.0,
      "step": 7712
    },
    {
      "epoch": 0.5932620567648642,
      "grad_norm": 61.0,
      "learning_rate": 2.1101069148527036e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3358780.0,
      "step": 7713
    },
    {
      "epoch": 0.5933389739250827,
      "grad_norm": 5.53125,
      "learning_rate": 2.109991539112376e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3359322.0,
      "step": 7714
    },
    {
      "epoch": 0.5934158910853011,
      "grad_norm": 3.359375,
      "learning_rate": 2.109876163372048e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3359902.0,
      "step": 7715
    },
    {
      "epoch": 0.5934928082455195,
      "grad_norm": 32.25,
      "learning_rate": 2.109760787631721e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.978723406791687,
      "num_tokens": 3360186.0,
      "step": 7716
    },
    {
      "epoch": 0.593569725405738,
      "grad_norm": 5.875,
      "learning_rate": 2.109645411891393e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3360734.0,
      "step": 7717
    },
    {
      "epoch": 0.5936466425659565,
      "grad_norm": 28.25,
      "learning_rate": 2.1095300361510654e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3361014.0,
      "step": 7718
    },
    {
      "epoch": 0.5937235597261749,
      "grad_norm": 39.75,
      "learning_rate": 2.1094146604107376e-06,
      "loss": 0.0454,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3361294.0,
      "step": 7719
    },
    {
      "epoch": 0.5938004768863934,
      "grad_norm": 24.625,
      "learning_rate": 2.10929928467041e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3361878.0,
      "step": 7720
    },
    {
      "epoch": 0.5938773940466118,
      "grad_norm": 28.25,
      "learning_rate": 2.1091839089300825e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 3362702.0,
      "step": 7721
    },
    {
      "epoch": 0.5939543112068303,
      "grad_norm": 7.34375,
      "learning_rate": 2.109068533189755e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3363014.0,
      "step": 7722
    },
    {
      "epoch": 0.5940312283670487,
      "grad_norm": 7.8125,
      "learning_rate": 2.108953157449427e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3363326.0,
      "step": 7723
    },
    {
      "epoch": 0.5941081455272671,
      "grad_norm": 13.375,
      "learning_rate": 2.1088377817090993e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3363648.0,
      "step": 7724
    },
    {
      "epoch": 0.5941850626874856,
      "grad_norm": 6.5,
      "learning_rate": 2.1087224059687716e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3364194.0,
      "step": 7725
    },
    {
      "epoch": 0.594261979847704,
      "grad_norm": 16.75,
      "learning_rate": 2.108607030228444e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3364534.0,
      "step": 7726
    },
    {
      "epoch": 0.5943388970079224,
      "grad_norm": 32.75,
      "learning_rate": 2.1084916544881165e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9915493130683899,
      "num_tokens": 3364890.0,
      "step": 7727
    },
    {
      "epoch": 0.5944158141681409,
      "grad_norm": 53.0,
      "learning_rate": 2.1083762787477888e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 3365258.0,
      "step": 7728
    },
    {
      "epoch": 0.5944927313283593,
      "grad_norm": 32.5,
      "learning_rate": 2.108260903007461e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 3366318.0,
      "step": 7729
    },
    {
      "epoch": 0.5945696484885779,
      "grad_norm": 14.875,
      "learning_rate": 2.1081455272671333e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3366604.0,
      "step": 7730
    },
    {
      "epoch": 0.5946465656487963,
      "grad_norm": 19.375,
      "learning_rate": 2.1080301515268056e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3367168.0,
      "step": 7731
    },
    {
      "epoch": 0.5947234828090147,
      "grad_norm": 13.8125,
      "learning_rate": 2.1079147757864782e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3367498.0,
      "step": 7732
    },
    {
      "epoch": 0.5948003999692332,
      "grad_norm": 9.75,
      "learning_rate": 2.1077994000461505e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3367848.0,
      "step": 7733
    },
    {
      "epoch": 0.5948773171294516,
      "grad_norm": 13.9375,
      "learning_rate": 2.1076840243058223e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 3368366.0,
      "step": 7734
    },
    {
      "epoch": 0.59495423428967,
      "grad_norm": 17.5,
      "learning_rate": 2.107568648565495e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 3368728.0,
      "step": 7735
    },
    {
      "epoch": 0.5950311514498885,
      "grad_norm": 19.25,
      "learning_rate": 2.1074532728251673e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3369062.0,
      "step": 7736
    },
    {
      "epoch": 0.5951080686101069,
      "grad_norm": 48.0,
      "learning_rate": 2.1073378970848395e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3369370.0,
      "step": 7737
    },
    {
      "epoch": 0.5951849857703254,
      "grad_norm": 11.875,
      "learning_rate": 2.107222521344512e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3369714.0,
      "step": 7738
    },
    {
      "epoch": 0.5952619029305438,
      "grad_norm": 3.9375,
      "learning_rate": 2.1071071456041845e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3370496.0,
      "step": 7739
    },
    {
      "epoch": 0.5953388200907622,
      "grad_norm": 13.5,
      "learning_rate": 2.1069917698638567e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 3370852.0,
      "step": 7740
    },
    {
      "epoch": 0.5954157372509807,
      "grad_norm": 10.1875,
      "learning_rate": 2.106876394123529e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3371426.0,
      "step": 7741
    },
    {
      "epoch": 0.5954926544111991,
      "grad_norm": 6.84375,
      "learning_rate": 2.1067610183832012e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 3372234.0,
      "step": 7742
    },
    {
      "epoch": 0.5955695715714175,
      "grad_norm": 19.0,
      "learning_rate": 2.106645642642874e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 3372796.0,
      "step": 7743
    },
    {
      "epoch": 0.5956464887316361,
      "grad_norm": 2.734375,
      "learning_rate": 2.106530266902546e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3373338.0,
      "step": 7744
    },
    {
      "epoch": 0.5957234058918545,
      "grad_norm": 11.4375,
      "learning_rate": 2.106414891162218e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 3373612.0,
      "step": 7745
    },
    {
      "epoch": 0.5958003230520729,
      "grad_norm": 12.8125,
      "learning_rate": 2.1062995154218907e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 3373934.0,
      "step": 7746
    },
    {
      "epoch": 0.5958772402122914,
      "grad_norm": 13.6875,
      "learning_rate": 2.106184139681563e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3374256.0,
      "step": 7747
    },
    {
      "epoch": 0.5959541573725098,
      "grad_norm": 52.75,
      "learning_rate": 2.106068763941235e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 3374558.0,
      "step": 7748
    },
    {
      "epoch": 0.5960310745327283,
      "grad_norm": 4.5625,
      "learning_rate": 2.105953388200908e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3375128.0,
      "step": 7749
    },
    {
      "epoch": 0.5961079916929467,
      "grad_norm": 7.71875,
      "learning_rate": 2.10583801246058e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 3375700.0,
      "step": 7750
    },
    {
      "epoch": 0.5961849088531651,
      "grad_norm": 21.25,
      "learning_rate": 2.1057226367202524e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9969230890274048,
      "num_tokens": 3376026.0,
      "step": 7751
    },
    {
      "epoch": 0.5962618260133836,
      "grad_norm": 66.5,
      "learning_rate": 2.1056072609799247e-06,
      "loss": 0.0414,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3376310.0,
      "step": 7752
    },
    {
      "epoch": 0.596338743173602,
      "grad_norm": 4.59375,
      "learning_rate": 2.105491885239597e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3376858.0,
      "step": 7753
    },
    {
      "epoch": 0.5964156603338204,
      "grad_norm": 68.5,
      "learning_rate": 2.1053765094992696e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3377168.0,
      "step": 7754
    },
    {
      "epoch": 0.596492577494039,
      "grad_norm": 71.0,
      "learning_rate": 2.105261133758942e-06,
      "loss": 0.0423,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 3377470.0,
      "step": 7755
    },
    {
      "epoch": 0.5965694946542573,
      "grad_norm": 11.0625,
      "learning_rate": 2.1051457580186137e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3377766.0,
      "step": 7756
    },
    {
      "epoch": 0.5966464118144759,
      "grad_norm": 22.0,
      "learning_rate": 2.1050303822782864e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3378088.0,
      "step": 7757
    },
    {
      "epoch": 0.5967233289746943,
      "grad_norm": 27.5,
      "learning_rate": 2.1049150065379586e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3378400.0,
      "step": 7758
    },
    {
      "epoch": 0.5968002461349127,
      "grad_norm": 5.0,
      "learning_rate": 2.104799630797631e-06,
      "loss": 0.0055,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 3379450.0,
      "step": 7759
    },
    {
      "epoch": 0.5968771632951312,
      "grad_norm": 13.3125,
      "learning_rate": 2.1046842550573036e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3379710.0,
      "step": 7760
    },
    {
      "epoch": 0.5969540804553496,
      "grad_norm": 4.1875,
      "learning_rate": 2.104568879316976e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3380274.0,
      "step": 7761
    },
    {
      "epoch": 0.597030997615568,
      "grad_norm": 45.0,
      "learning_rate": 2.104453503576648e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3380556.0,
      "step": 7762
    },
    {
      "epoch": 0.5971079147757865,
      "grad_norm": 18.0,
      "learning_rate": 2.1043381278363204e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 3381072.0,
      "step": 7763
    },
    {
      "epoch": 0.5971848319360049,
      "grad_norm": 31.875,
      "learning_rate": 2.1042227520959926e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3381354.0,
      "step": 7764
    },
    {
      "epoch": 0.5972617490962233,
      "grad_norm": 30.0,
      "learning_rate": 2.104107376355665e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3381684.0,
      "step": 7765
    },
    {
      "epoch": 0.5973386662564418,
      "grad_norm": 45.0,
      "learning_rate": 2.1039920006153376e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3382028.0,
      "step": 7766
    },
    {
      "epoch": 0.5974155834166602,
      "grad_norm": 43.0,
      "learning_rate": 2.1038766248750094e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 3382332.0,
      "step": 7767
    },
    {
      "epoch": 0.5974925005768787,
      "grad_norm": 11.75,
      "learning_rate": 2.103761249134682e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3382616.0,
      "step": 7768
    },
    {
      "epoch": 0.5975694177370972,
      "grad_norm": 17.375,
      "learning_rate": 2.1036458733943543e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9966996908187866,
      "num_tokens": 3382920.0,
      "step": 7769
    },
    {
      "epoch": 0.5976463348973156,
      "grad_norm": 13.5,
      "learning_rate": 2.1035304976540266e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 3383208.0,
      "step": 7770
    },
    {
      "epoch": 0.5977232520575341,
      "grad_norm": 6.59375,
      "learning_rate": 2.1034151219136993e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3383772.0,
      "step": 7771
    },
    {
      "epoch": 0.5978001692177525,
      "grad_norm": 18.25,
      "learning_rate": 2.103299746173371e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3384054.0,
      "step": 7772
    },
    {
      "epoch": 0.5978770863779709,
      "grad_norm": 24.75,
      "learning_rate": 2.1031843704330434e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3384370.0,
      "step": 7773
    },
    {
      "epoch": 0.5979540035381894,
      "grad_norm": 32.75,
      "learning_rate": 2.103068994692716e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3384674.0,
      "step": 7774
    },
    {
      "epoch": 0.5980309206984078,
      "grad_norm": 48.0,
      "learning_rate": 2.1029536189523883e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3384984.0,
      "step": 7775
    },
    {
      "epoch": 0.5981078378586263,
      "grad_norm": 29.625,
      "learning_rate": 2.1028382432120606e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3385742.0,
      "step": 7776
    },
    {
      "epoch": 0.5981847550188447,
      "grad_norm": 39.75,
      "learning_rate": 2.1027228674717332e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3386040.0,
      "step": 7777
    },
    {
      "epoch": 0.5982616721790631,
      "grad_norm": 16.125,
      "learning_rate": 2.102607491731405e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 3386356.0,
      "step": 7778
    },
    {
      "epoch": 0.5983385893392816,
      "grad_norm": 55.5,
      "learning_rate": 2.1024921159910778e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3386666.0,
      "step": 7779
    },
    {
      "epoch": 0.5984155064995,
      "grad_norm": 34.75,
      "learning_rate": 2.10237674025075e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3387240.0,
      "step": 7780
    },
    {
      "epoch": 0.5984924236597184,
      "grad_norm": 8.3125,
      "learning_rate": 2.1022613645104223e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3387522.0,
      "step": 7781
    },
    {
      "epoch": 0.598569340819937,
      "grad_norm": 20.75,
      "learning_rate": 2.102145988770095e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3387826.0,
      "step": 7782
    },
    {
      "epoch": 0.5986462579801554,
      "grad_norm": 21.875,
      "learning_rate": 2.1020306130297668e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3388382.0,
      "step": 7783
    },
    {
      "epoch": 0.5987231751403738,
      "grad_norm": 40.5,
      "learning_rate": 2.101915237289439e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 3388658.0,
      "step": 7784
    },
    {
      "epoch": 0.5988000923005923,
      "grad_norm": 24.375,
      "learning_rate": 2.1017998615491117e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 3389474.0,
      "step": 7785
    },
    {
      "epoch": 0.5988770094608107,
      "grad_norm": 62.25,
      "learning_rate": 2.101684485808784e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 3389812.0,
      "step": 7786
    },
    {
      "epoch": 0.5989539266210292,
      "grad_norm": 15.3125,
      "learning_rate": 2.1015691100684562e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 3390126.0,
      "step": 7787
    },
    {
      "epoch": 0.5990308437812476,
      "grad_norm": 20.25,
      "learning_rate": 2.101453734328129e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9982847571372986,
      "num_tokens": 3390710.0,
      "step": 7788
    },
    {
      "epoch": 0.599107760941466,
      "grad_norm": 36.25,
      "learning_rate": 2.1013383585878008e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 3391014.0,
      "step": 7789
    },
    {
      "epoch": 0.5991846781016845,
      "grad_norm": 11.6875,
      "learning_rate": 2.1012229828474734e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 3391376.0,
      "step": 7790
    },
    {
      "epoch": 0.5992615952619029,
      "grad_norm": 6.65625,
      "learning_rate": 2.1011076071071457e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3391726.0,
      "step": 7791
    },
    {
      "epoch": 0.5993385124221213,
      "grad_norm": 32.75,
      "learning_rate": 2.100992231366818e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3392060.0,
      "step": 7792
    },
    {
      "epoch": 0.5994154295823398,
      "grad_norm": 38.25,
      "learning_rate": 2.1008768556264906e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9764309525489807,
      "num_tokens": 3392360.0,
      "step": 7793
    },
    {
      "epoch": 0.5994923467425582,
      "grad_norm": 11.125,
      "learning_rate": 2.1007614798861625e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3392630.0,
      "step": 7794
    },
    {
      "epoch": 0.5995692639027768,
      "grad_norm": 14.0625,
      "learning_rate": 2.1006461041458347e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3392924.0,
      "step": 7795
    },
    {
      "epoch": 0.5996461810629952,
      "grad_norm": 12.5,
      "learning_rate": 2.1005307284055074e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3393250.0,
      "step": 7796
    },
    {
      "epoch": 0.5997230982232136,
      "grad_norm": 5.125,
      "learning_rate": 2.1004153526651797e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3393798.0,
      "step": 7797
    },
    {
      "epoch": 0.5998000153834321,
      "grad_norm": 22.5,
      "learning_rate": 2.100299976924852e-06,
      "loss": 0.0371,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 3394074.0,
      "step": 7798
    },
    {
      "epoch": 0.5998769325436505,
      "grad_norm": 8.5,
      "learning_rate": 2.100184601184524e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3394402.0,
      "step": 7799
    },
    {
      "epoch": 0.5999538497038689,
      "grad_norm": 27.625,
      "learning_rate": 2.1000692254441964e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 3394742.0,
      "step": 7800
    },
    {
      "epoch": 0.6000307668640874,
      "grad_norm": 8.5,
      "learning_rate": 2.099953849703869e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3395316.0,
      "step": 7801
    },
    {
      "epoch": 0.6001076840243058,
      "grad_norm": 23.75,
      "learning_rate": 2.0998384739635414e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3395632.0,
      "step": 7802
    },
    {
      "epoch": 0.6001846011845242,
      "grad_norm": 8.0,
      "learning_rate": 2.0997230982232136e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9982847571372986,
      "num_tokens": 3396216.0,
      "step": 7803
    },
    {
      "epoch": 0.6002615183447427,
      "grad_norm": 19.0,
      "learning_rate": 2.0996077224828863e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 3396774.0,
      "step": 7804
    },
    {
      "epoch": 0.6003384355049611,
      "grad_norm": 48.75,
      "learning_rate": 2.099492346742558e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3397084.0,
      "step": 7805
    },
    {
      "epoch": 0.6004153526651796,
      "grad_norm": 13.4375,
      "learning_rate": 2.0993769710022304e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 3397908.0,
      "step": 7806
    },
    {
      "epoch": 0.600492269825398,
      "grad_norm": 41.0,
      "learning_rate": 2.099261595261903e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 3398190.0,
      "step": 7807
    },
    {
      "epoch": 0.6005691869856165,
      "grad_norm": 58.75,
      "learning_rate": 2.0991462195215754e-06,
      "loss": 0.0433,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 3398514.0,
      "step": 7808
    },
    {
      "epoch": 0.600646104145835,
      "grad_norm": 14.0,
      "learning_rate": 2.0990308437812476e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3398822.0,
      "step": 7809
    },
    {
      "epoch": 0.6007230213060534,
      "grad_norm": 23.625,
      "learning_rate": 2.09891546804092e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3399174.0,
      "step": 7810
    },
    {
      "epoch": 0.6007999384662718,
      "grad_norm": 46.75,
      "learning_rate": 2.098800092300592e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3399458.0,
      "step": 7811
    },
    {
      "epoch": 0.6008768556264903,
      "grad_norm": 12.625,
      "learning_rate": 2.098684716560265e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3399780.0,
      "step": 7812
    },
    {
      "epoch": 0.6009537727867087,
      "grad_norm": 19.0,
      "learning_rate": 2.098569340819937e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 3400012.0,
      "step": 7813
    },
    {
      "epoch": 0.6010306899469272,
      "grad_norm": 36.0,
      "learning_rate": 2.0984539650796093e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 3400848.0,
      "step": 7814
    },
    {
      "epoch": 0.6011076071071456,
      "grad_norm": 16.0,
      "learning_rate": 2.098338589339282e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 3401166.0,
      "step": 7815
    },
    {
      "epoch": 0.601184524267364,
      "grad_norm": 27.375,
      "learning_rate": 2.098223213598954e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3401484.0,
      "step": 7816
    },
    {
      "epoch": 0.6012614414275825,
      "grad_norm": 31.375,
      "learning_rate": 2.098107837858626e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 3402282.0,
      "step": 7817
    },
    {
      "epoch": 0.6013383585878009,
      "grad_norm": 47.75,
      "learning_rate": 2.0979924621182988e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3402560.0,
      "step": 7818
    },
    {
      "epoch": 0.6014152757480193,
      "grad_norm": 21.625,
      "learning_rate": 2.097877086377971e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 3402820.0,
      "step": 7819
    },
    {
      "epoch": 0.6014921929082379,
      "grad_norm": 21.875,
      "learning_rate": 2.0977617106376433e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3403098.0,
      "step": 7820
    },
    {
      "epoch": 0.6015691100684563,
      "grad_norm": 42.75,
      "learning_rate": 2.0976463348973156e-06,
      "loss": 0.0419,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3403398.0,
      "step": 7821
    },
    {
      "epoch": 0.6016460272286747,
      "grad_norm": 9.625,
      "learning_rate": 2.097530959156988e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3403684.0,
      "step": 7822
    },
    {
      "epoch": 0.6017229443888932,
      "grad_norm": 20.375,
      "learning_rate": 2.0974155834166605e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3404264.0,
      "step": 7823
    },
    {
      "epoch": 0.6017998615491116,
      "grad_norm": 18.125,
      "learning_rate": 2.0973002076763328e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3404808.0,
      "step": 7824
    },
    {
      "epoch": 0.6018767787093301,
      "grad_norm": 20.0,
      "learning_rate": 2.097184831936005e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3405140.0,
      "step": 7825
    },
    {
      "epoch": 0.6019536958695485,
      "grad_norm": 19.125,
      "learning_rate": 2.0970694561956773e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3405686.0,
      "step": 7826
    },
    {
      "epoch": 0.6020306130297669,
      "grad_norm": 28.875,
      "learning_rate": 2.0969540804553495e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9855907559394836,
      "num_tokens": 3406034.0,
      "step": 7827
    },
    {
      "epoch": 0.6021075301899854,
      "grad_norm": 23.25,
      "learning_rate": 2.096838704715022e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 3406860.0,
      "step": 7828
    },
    {
      "epoch": 0.6021844473502038,
      "grad_norm": 17.125,
      "learning_rate": 2.0967233289746945e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 3407380.0,
      "step": 7829
    },
    {
      "epoch": 0.6022613645104222,
      "grad_norm": 24.5,
      "learning_rate": 2.0966079532343667e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 3407940.0,
      "step": 7830
    },
    {
      "epoch": 0.6023382816706407,
      "grad_norm": 18.375,
      "learning_rate": 2.096492577494039e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3408246.0,
      "step": 7831
    },
    {
      "epoch": 0.6024151988308591,
      "grad_norm": 11.0,
      "learning_rate": 2.0963772017537112e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3408792.0,
      "step": 7832
    },
    {
      "epoch": 0.6024921159910777,
      "grad_norm": 16.125,
      "learning_rate": 2.0962618260133835e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 3409106.0,
      "step": 7833
    },
    {
      "epoch": 0.6025690331512961,
      "grad_norm": 21.25,
      "learning_rate": 2.0961464502730558e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3409390.0,
      "step": 7834
    },
    {
      "epoch": 0.6026459503115145,
      "grad_norm": 18.25,
      "learning_rate": 2.0960310745327284e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 3409724.0,
      "step": 7835
    },
    {
      "epoch": 0.602722867471733,
      "grad_norm": 9.8125,
      "learning_rate": 2.0959156987924007e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3410266.0,
      "step": 7836
    },
    {
      "epoch": 0.6027997846319514,
      "grad_norm": 17.5,
      "learning_rate": 2.095800323052073e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3410524.0,
      "step": 7837
    },
    {
      "epoch": 0.6028767017921698,
      "grad_norm": 5.78125,
      "learning_rate": 2.0956849473117452e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3411304.0,
      "step": 7838
    },
    {
      "epoch": 0.6029536189523883,
      "grad_norm": 24.75,
      "learning_rate": 2.0955695715714175e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3411850.0,
      "step": 7839
    },
    {
      "epoch": 0.6030305361126067,
      "grad_norm": 73.0,
      "learning_rate": 2.09545419583109e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3412178.0,
      "step": 7840
    },
    {
      "epoch": 0.6031074532728252,
      "grad_norm": 12.75,
      "learning_rate": 2.0953388200907624e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 3412726.0,
      "step": 7841
    },
    {
      "epoch": 0.6031843704330436,
      "grad_norm": 14.9375,
      "learning_rate": 2.0952234443504347e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3413010.0,
      "step": 7842
    },
    {
      "epoch": 0.603261287593262,
      "grad_norm": 36.5,
      "learning_rate": 2.095108068610107e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3413268.0,
      "step": 7843
    },
    {
      "epoch": 0.6033382047534805,
      "grad_norm": 10.5,
      "learning_rate": 2.094992692869779e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 3413558.0,
      "step": 7844
    },
    {
      "epoch": 0.603415121913699,
      "grad_norm": 24.875,
      "learning_rate": 2.0948773171294514e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3414138.0,
      "step": 7845
    },
    {
      "epoch": 0.6034920390739174,
      "grad_norm": 22.75,
      "learning_rate": 2.094761941389124e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3414688.0,
      "step": 7846
    },
    {
      "epoch": 0.6035689562341359,
      "grad_norm": 20.75,
      "learning_rate": 2.0946465656487964e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3415274.0,
      "step": 7847
    },
    {
      "epoch": 0.6036458733943543,
      "grad_norm": 18.75,
      "learning_rate": 2.0945311899084686e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3415584.0,
      "step": 7848
    },
    {
      "epoch": 0.6037227905545727,
      "grad_norm": 7.96875,
      "learning_rate": 2.094415814168141e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 3415922.0,
      "step": 7849
    },
    {
      "epoch": 0.6037997077147912,
      "grad_norm": 23.75,
      "learning_rate": 2.094300438427813e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 3416182.0,
      "step": 7850
    },
    {
      "epoch": 0.6038766248750096,
      "grad_norm": 10.875,
      "learning_rate": 2.094185062687486e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3416738.0,
      "step": 7851
    },
    {
      "epoch": 0.6039535420352281,
      "grad_norm": 5.65625,
      "learning_rate": 2.094069686947158e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3417284.0,
      "step": 7852
    },
    {
      "epoch": 0.6040304591954465,
      "grad_norm": 2.984375,
      "learning_rate": 2.0939543112068304e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3417840.0,
      "step": 7853
    },
    {
      "epoch": 0.6041073763556649,
      "grad_norm": 6.65625,
      "learning_rate": 2.0938389354665026e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9974293112754822,
      "num_tokens": 3418620.0,
      "step": 7854
    },
    {
      "epoch": 0.6041842935158834,
      "grad_norm": 35.25,
      "learning_rate": 2.093723559726175e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3418930.0,
      "step": 7855
    },
    {
      "epoch": 0.6042612106761018,
      "grad_norm": 26.625,
      "learning_rate": 2.093608183985847e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9965635538101196,
      "num_tokens": 3419222.0,
      "step": 7856
    },
    {
      "epoch": 0.6043381278363202,
      "grad_norm": 8.3125,
      "learning_rate": 2.09349280824552e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3419544.0,
      "step": 7857
    },
    {
      "epoch": 0.6044150449965388,
      "grad_norm": 40.25,
      "learning_rate": 2.093377432505192e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 3419852.0,
      "step": 7858
    },
    {
      "epoch": 0.6044919621567572,
      "grad_norm": 11.6875,
      "learning_rate": 2.0932620567648643e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3420408.0,
      "step": 7859
    },
    {
      "epoch": 0.6045688793169757,
      "grad_norm": 2.609375,
      "learning_rate": 2.0931466810245366e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 3420956.0,
      "step": 7860
    },
    {
      "epoch": 0.6046457964771941,
      "grad_norm": 19.125,
      "learning_rate": 2.093031305284209e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 3421548.0,
      "step": 7861
    },
    {
      "epoch": 0.6047227136374125,
      "grad_norm": 4.6875,
      "learning_rate": 2.0929159295438815e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3422112.0,
      "step": 7862
    },
    {
      "epoch": 0.604799630797631,
      "grad_norm": 10.8125,
      "learning_rate": 2.0928005538035538e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3422658.0,
      "step": 7863
    },
    {
      "epoch": 0.6048765479578494,
      "grad_norm": 33.25,
      "learning_rate": 2.092685178063226e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3422914.0,
      "step": 7864
    },
    {
      "epoch": 0.6049534651180678,
      "grad_norm": 12.375,
      "learning_rate": 2.0925698023228983e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3423488.0,
      "step": 7865
    },
    {
      "epoch": 0.6050303822782863,
      "grad_norm": 41.0,
      "learning_rate": 2.0924544265825706e-06,
      "loss": 0.043,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 3423816.0,
      "step": 7866
    },
    {
      "epoch": 0.6051072994385047,
      "grad_norm": 33.5,
      "learning_rate": 2.092339050842243e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 3424122.0,
      "step": 7867
    },
    {
      "epoch": 0.6051842165987231,
      "grad_norm": 15.1875,
      "learning_rate": 2.0922236751019155e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9981982111930847,
      "num_tokens": 3424678.0,
      "step": 7868
    },
    {
      "epoch": 0.6052611337589416,
      "grad_norm": 40.0,
      "learning_rate": 2.0921082993615878e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3424998.0,
      "step": 7869
    },
    {
      "epoch": 0.60533805091916,
      "grad_norm": 18.125,
      "learning_rate": 2.09199292362126e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3425334.0,
      "step": 7870
    },
    {
      "epoch": 0.6054149680793786,
      "grad_norm": 64.5,
      "learning_rate": 2.0918775478809323e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3425612.0,
      "step": 7871
    },
    {
      "epoch": 0.605491885239597,
      "grad_norm": 25.625,
      "learning_rate": 2.0917621721406045e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 3425944.0,
      "step": 7872
    },
    {
      "epoch": 0.6055688023998154,
      "grad_norm": 24.25,
      "learning_rate": 2.0916467964002772e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 3426532.0,
      "step": 7873
    },
    {
      "epoch": 0.6056457195600339,
      "grad_norm": 68.0,
      "learning_rate": 2.0915314206599495e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 3427128.0,
      "step": 7874
    },
    {
      "epoch": 0.6057226367202523,
      "grad_norm": 47.5,
      "learning_rate": 2.0914160449196213e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3427448.0,
      "step": 7875
    },
    {
      "epoch": 0.6057995538804707,
      "grad_norm": 38.25,
      "learning_rate": 2.091300669179294e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3427748.0,
      "step": 7876
    },
    {
      "epoch": 0.6058764710406892,
      "grad_norm": 45.0,
      "learning_rate": 2.0911852934389662e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3428078.0,
      "step": 7877
    },
    {
      "epoch": 0.6059533882009076,
      "grad_norm": 15.8125,
      "learning_rate": 2.0910699176986385e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3428622.0,
      "step": 7878
    },
    {
      "epoch": 0.6060303053611261,
      "grad_norm": 14.6875,
      "learning_rate": 2.090954541958311e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3429196.0,
      "step": 7879
    },
    {
      "epoch": 0.6061072225213445,
      "grad_norm": 2.578125,
      "learning_rate": 2.0908391662179834e-06,
      "loss": 0.0035,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3430240.0,
      "step": 7880
    },
    {
      "epoch": 0.6061841396815629,
      "grad_norm": 31.0,
      "learning_rate": 2.0907237904776557e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3430524.0,
      "step": 7881
    },
    {
      "epoch": 0.6062610568417814,
      "grad_norm": 62.75,
      "learning_rate": 2.090608414737328e-06,
      "loss": 0.0431,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3430804.0,
      "step": 7882
    },
    {
      "epoch": 0.6063379740019998,
      "grad_norm": 2.1875,
      "learning_rate": 2.0904930389970002e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.9974293112754822,
      "num_tokens": 3431584.0,
      "step": 7883
    },
    {
      "epoch": 0.6064148911622183,
      "grad_norm": 14.6875,
      "learning_rate": 2.0903776632566725e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3431870.0,
      "step": 7884
    },
    {
      "epoch": 0.6064918083224368,
      "grad_norm": 10.0,
      "learning_rate": 2.090262287516345e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3432192.0,
      "step": 7885
    },
    {
      "epoch": 0.6065687254826552,
      "grad_norm": 8.8125,
      "learning_rate": 2.090146911776017e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3432772.0,
      "step": 7886
    },
    {
      "epoch": 0.6066456426428736,
      "grad_norm": 20.75,
      "learning_rate": 2.0900315360356897e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 3433290.0,
      "step": 7887
    },
    {
      "epoch": 0.6067225598030921,
      "grad_norm": 23.625,
      "learning_rate": 2.089916160295362e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 3433616.0,
      "step": 7888
    },
    {
      "epoch": 0.6067994769633105,
      "grad_norm": 12.3125,
      "learning_rate": 2.089800784555034e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3434180.0,
      "step": 7889
    },
    {
      "epoch": 0.606876394123529,
      "grad_norm": 66.0,
      "learning_rate": 2.089685408814707e-06,
      "loss": 0.0399,
      "mean_token_accuracy": 0.982758641242981,
      "num_tokens": 3434416.0,
      "step": 7890
    },
    {
      "epoch": 0.6069533112837474,
      "grad_norm": 7.9375,
      "learning_rate": 2.089570033074379e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 3434756.0,
      "step": 7891
    },
    {
      "epoch": 0.6070302284439658,
      "grad_norm": 13.625,
      "learning_rate": 2.089454657334051e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 3435096.0,
      "step": 7892
    },
    {
      "epoch": 0.6071071456041843,
      "grad_norm": 44.5,
      "learning_rate": 2.0893392815937236e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 3435400.0,
      "step": 7893
    },
    {
      "epoch": 0.6071840627644027,
      "grad_norm": 99.0,
      "learning_rate": 2.089223905853396e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3435706.0,
      "step": 7894
    },
    {
      "epoch": 0.6072609799246211,
      "grad_norm": 25.125,
      "learning_rate": 2.089108530113068e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 3436520.0,
      "step": 7895
    },
    {
      "epoch": 0.6073378970848397,
      "grad_norm": 59.75,
      "learning_rate": 2.088993154372741e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3436832.0,
      "step": 7896
    },
    {
      "epoch": 0.607414814245058,
      "grad_norm": 38.75,
      "learning_rate": 2.0888777786324127e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3437154.0,
      "step": 7897
    },
    {
      "epoch": 0.6074917314052766,
      "grad_norm": 67.5,
      "learning_rate": 2.0887624028920854e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3437410.0,
      "step": 7898
    },
    {
      "epoch": 0.607568648565495,
      "grad_norm": 27.625,
      "learning_rate": 2.0886470271517576e-06,
      "loss": 0.0065,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 3438454.0,
      "step": 7899
    },
    {
      "epoch": 0.6076455657257134,
      "grad_norm": 35.5,
      "learning_rate": 2.08853165141143e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3439018.0,
      "step": 7900
    },
    {
      "epoch": 0.6077224828859319,
      "grad_norm": 29.5,
      "learning_rate": 2.0884162756711026e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3439344.0,
      "step": 7901
    },
    {
      "epoch": 0.6077994000461503,
      "grad_norm": 29.5,
      "learning_rate": 2.0883008999307744e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3439604.0,
      "step": 7902
    },
    {
      "epoch": 0.6078763172063687,
      "grad_norm": 18.125,
      "learning_rate": 2.0881855241904467e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 3440140.0,
      "step": 7903
    },
    {
      "epoch": 0.6079532343665872,
      "grad_norm": 61.75,
      "learning_rate": 2.0880701484501193e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.996874988079071,
      "num_tokens": 3440464.0,
      "step": 7904
    },
    {
      "epoch": 0.6080301515268056,
      "grad_norm": 38.25,
      "learning_rate": 2.0879547727097916e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 3440826.0,
      "step": 7905
    },
    {
      "epoch": 0.608107068687024,
      "grad_norm": 7.5,
      "learning_rate": 2.087839396969464e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3441372.0,
      "step": 7906
    },
    {
      "epoch": 0.6081839858472425,
      "grad_norm": 2.453125,
      "learning_rate": 2.0877240212291365e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3441906.0,
      "step": 7907
    },
    {
      "epoch": 0.6082609030074609,
      "grad_norm": 12.5625,
      "learning_rate": 2.0876086454888084e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9973545074462891,
      "num_tokens": 3442664.0,
      "step": 7908
    },
    {
      "epoch": 0.6083378201676795,
      "grad_norm": 18.5,
      "learning_rate": 2.087493269748481e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 3443444.0,
      "step": 7909
    },
    {
      "epoch": 0.6084147373278979,
      "grad_norm": 3.265625,
      "learning_rate": 2.0873778940081533e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3444000.0,
      "step": 7910
    },
    {
      "epoch": 0.6084916544881163,
      "grad_norm": 10.625,
      "learning_rate": 2.0872625182678256e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3444286.0,
      "step": 7911
    },
    {
      "epoch": 0.6085685716483348,
      "grad_norm": 29.0,
      "learning_rate": 2.0871471425274982e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3444632.0,
      "step": 7912
    },
    {
      "epoch": 0.6086454888085532,
      "grad_norm": 26.625,
      "learning_rate": 2.08703176678717e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 3444950.0,
      "step": 7913
    },
    {
      "epoch": 0.6087224059687716,
      "grad_norm": 13.5625,
      "learning_rate": 2.0869163910468423e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3445472.0,
      "step": 7914
    },
    {
      "epoch": 0.6087993231289901,
      "grad_norm": 21.5,
      "learning_rate": 2.086801015306515e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3446016.0,
      "step": 7915
    },
    {
      "epoch": 0.6088762402892085,
      "grad_norm": 9.25,
      "learning_rate": 2.0866856395661873e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3446352.0,
      "step": 7916
    },
    {
      "epoch": 0.608953157449427,
      "grad_norm": 55.75,
      "learning_rate": 2.0865702638258595e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 3446684.0,
      "step": 7917
    },
    {
      "epoch": 0.6090300746096454,
      "grad_norm": 12.125,
      "learning_rate": 2.0864548880855322e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3446990.0,
      "step": 7918
    },
    {
      "epoch": 0.6091069917698638,
      "grad_norm": 22.625,
      "learning_rate": 2.086339512345204e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 3447306.0,
      "step": 7919
    },
    {
      "epoch": 0.6091839089300823,
      "grad_norm": 22.5,
      "learning_rate": 2.0862241366048767e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9781022071838379,
      "num_tokens": 3447582.0,
      "step": 7920
    },
    {
      "epoch": 0.6092608260903007,
      "grad_norm": 12.25,
      "learning_rate": 2.086108760864549e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3447916.0,
      "step": 7921
    },
    {
      "epoch": 0.6093377432505191,
      "grad_norm": 14.25,
      "learning_rate": 2.0859933851242212e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9825327396392822,
      "num_tokens": 3448146.0,
      "step": 7922
    },
    {
      "epoch": 0.6094146604107377,
      "grad_norm": 6.5,
      "learning_rate": 2.085878009383894e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 3448666.0,
      "step": 7923
    },
    {
      "epoch": 0.6094915775709561,
      "grad_norm": 14.125,
      "learning_rate": 2.0857626336435658e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3448988.0,
      "step": 7924
    },
    {
      "epoch": 0.6095684947311745,
      "grad_norm": 7.5,
      "learning_rate": 2.085647257903238e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 3449796.0,
      "step": 7925
    },
    {
      "epoch": 0.609645411891393,
      "grad_norm": 22.0,
      "learning_rate": 2.0855318821629107e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 3450134.0,
      "step": 7926
    },
    {
      "epoch": 0.6097223290516114,
      "grad_norm": 7.4375,
      "learning_rate": 2.085416506422583e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3450690.0,
      "step": 7927
    },
    {
      "epoch": 0.6097992462118299,
      "grad_norm": 11.1875,
      "learning_rate": 2.0853011306822552e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3451248.0,
      "step": 7928
    },
    {
      "epoch": 0.6098761633720483,
      "grad_norm": 7.375,
      "learning_rate": 2.085185754941928e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3451560.0,
      "step": 7929
    },
    {
      "epoch": 0.6099530805322667,
      "grad_norm": 22.5,
      "learning_rate": 2.0850703792015997e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3451838.0,
      "step": 7930
    },
    {
      "epoch": 0.6100299976924852,
      "grad_norm": 23.5,
      "learning_rate": 2.0849550034612724e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3452402.0,
      "step": 7931
    },
    {
      "epoch": 0.6101069148527036,
      "grad_norm": 21.5,
      "learning_rate": 2.0848396277209447e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3452680.0,
      "step": 7932
    },
    {
      "epoch": 0.610183832012922,
      "grad_norm": 37.5,
      "learning_rate": 2.084724251980617e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 3453014.0,
      "step": 7933
    },
    {
      "epoch": 0.6102607491731405,
      "grad_norm": 7.5625,
      "learning_rate": 2.084608876240289e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3453298.0,
      "step": 7934
    },
    {
      "epoch": 0.610337666333359,
      "grad_norm": 6.5,
      "learning_rate": 2.0844935004999615e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3453620.0,
      "step": 7935
    },
    {
      "epoch": 0.6104145834935775,
      "grad_norm": 16.0,
      "learning_rate": 2.0843781247596337e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 3454402.0,
      "step": 7936
    },
    {
      "epoch": 0.6104915006537959,
      "grad_norm": 6.1875,
      "learning_rate": 2.0842627490193064e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9968454241752625,
      "num_tokens": 3454720.0,
      "step": 7937
    },
    {
      "epoch": 0.6105684178140143,
      "grad_norm": 19.125,
      "learning_rate": 2.0841473732789787e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3455070.0,
      "step": 7938
    },
    {
      "epoch": 0.6106453349742328,
      "grad_norm": 15.875,
      "learning_rate": 2.084031997538651e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 3455404.0,
      "step": 7939
    },
    {
      "epoch": 0.6107222521344512,
      "grad_norm": 9.9375,
      "learning_rate": 2.083916621798323e-06,
      "loss": 0.0049,
      "mean_token_accuracy": 0.9971618056297302,
      "num_tokens": 3456464.0,
      "step": 7940
    },
    {
      "epoch": 0.6107991692946696,
      "grad_norm": 4.59375,
      "learning_rate": 2.0838012460579954e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3457012.0,
      "step": 7941
    },
    {
      "epoch": 0.6108760864548881,
      "grad_norm": 28.875,
      "learning_rate": 2.0836858703176677e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3457318.0,
      "step": 7942
    },
    {
      "epoch": 0.6109530036151065,
      "grad_norm": 19.875,
      "learning_rate": 2.0835704945773404e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3457636.0,
      "step": 7943
    },
    {
      "epoch": 0.6110299207753249,
      "grad_norm": 18.875,
      "learning_rate": 2.0834551188370126e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 3457988.0,
      "step": 7944
    },
    {
      "epoch": 0.6111068379355434,
      "grad_norm": 4.59375,
      "learning_rate": 2.083339743096685e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 3458540.0,
      "step": 7945
    },
    {
      "epoch": 0.6111837550957618,
      "grad_norm": 4.5,
      "learning_rate": 2.083224367356357e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 3459116.0,
      "step": 7946
    },
    {
      "epoch": 0.6112606722559804,
      "grad_norm": 24.375,
      "learning_rate": 2.0831089916160294e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3459422.0,
      "step": 7947
    },
    {
      "epoch": 0.6113375894161988,
      "grad_norm": 120.0,
      "learning_rate": 2.082993615875702e-06,
      "loss": 0.0445,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3459722.0,
      "step": 7948
    },
    {
      "epoch": 0.6114145065764172,
      "grad_norm": 14.125,
      "learning_rate": 2.0828782401353743e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 3460478.0,
      "step": 7949
    },
    {
      "epoch": 0.6114914237366357,
      "grad_norm": 18.75,
      "learning_rate": 2.0827628643950466e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 3460778.0,
      "step": 7950
    },
    {
      "epoch": 0.6115683408968541,
      "grad_norm": 16.75,
      "learning_rate": 2.082647488654719e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3461102.0,
      "step": 7951
    },
    {
      "epoch": 0.6116452580570725,
      "grad_norm": 16.875,
      "learning_rate": 2.082532112914391e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3461358.0,
      "step": 7952
    },
    {
      "epoch": 0.611722175217291,
      "grad_norm": 12.9375,
      "learning_rate": 2.0824167371740634e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9971346855163574,
      "num_tokens": 3462408.0,
      "step": 7953
    },
    {
      "epoch": 0.6117990923775094,
      "grad_norm": 9.5,
      "learning_rate": 2.082301361433736e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3462982.0,
      "step": 7954
    },
    {
      "epoch": 0.6118760095377279,
      "grad_norm": 25.625,
      "learning_rate": 2.0821859856934083e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3463556.0,
      "step": 7955
    },
    {
      "epoch": 0.6119529266979463,
      "grad_norm": 66.5,
      "learning_rate": 2.0820706099530806e-06,
      "loss": 0.0482,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3463864.0,
      "step": 7956
    },
    {
      "epoch": 0.6120298438581647,
      "grad_norm": 71.5,
      "learning_rate": 2.081955234212753e-06,
      "loss": 0.0564,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 3464168.0,
      "step": 7957
    },
    {
      "epoch": 0.6121067610183832,
      "grad_norm": 12.0,
      "learning_rate": 2.081839858472425e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3464514.0,
      "step": 7958
    },
    {
      "epoch": 0.6121836781786016,
      "grad_norm": 15.5,
      "learning_rate": 2.0817244827320978e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9948892593383789,
      "num_tokens": 3465102.0,
      "step": 7959
    },
    {
      "epoch": 0.61226059533882,
      "grad_norm": 36.25,
      "learning_rate": 2.08160910699177e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3465358.0,
      "step": 7960
    },
    {
      "epoch": 0.6123375124990386,
      "grad_norm": 24.875,
      "learning_rate": 2.0814937312514423e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3465658.0,
      "step": 7961
    },
    {
      "epoch": 0.612414429659257,
      "grad_norm": 15.9375,
      "learning_rate": 2.0813783555111145e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 3466194.0,
      "step": 7962
    },
    {
      "epoch": 0.6124913468194755,
      "grad_norm": 18.875,
      "learning_rate": 2.081262979770787e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3466774.0,
      "step": 7963
    },
    {
      "epoch": 0.6125682639796939,
      "grad_norm": 4.0,
      "learning_rate": 2.081147604030459e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3467342.0,
      "step": 7964
    },
    {
      "epoch": 0.6126451811399123,
      "grad_norm": 10.8125,
      "learning_rate": 2.0810322282901317e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 3467668.0,
      "step": 7965
    },
    {
      "epoch": 0.6127220983001308,
      "grad_norm": 3.03125,
      "learning_rate": 2.080916852549804e-06,
      "loss": 0.0069,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 3468476.0,
      "step": 7966
    },
    {
      "epoch": 0.6127990154603492,
      "grad_norm": 33.25,
      "learning_rate": 2.0808014768094763e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3468786.0,
      "step": 7967
    },
    {
      "epoch": 0.6128759326205676,
      "grad_norm": 81.0,
      "learning_rate": 2.0806861010691485e-06,
      "loss": 0.0553,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 3469088.0,
      "step": 7968
    },
    {
      "epoch": 0.6129528497807861,
      "grad_norm": 25.375,
      "learning_rate": 2.0805707253288208e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3469404.0,
      "step": 7969
    },
    {
      "epoch": 0.6130297669410045,
      "grad_norm": 38.25,
      "learning_rate": 2.0804553495884935e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3469716.0,
      "step": 7970
    },
    {
      "epoch": 0.6131066841012229,
      "grad_norm": 34.25,
      "learning_rate": 2.0803399738481657e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3469974.0,
      "step": 7971
    },
    {
      "epoch": 0.6131836012614414,
      "grad_norm": 37.5,
      "learning_rate": 2.080224598107838e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9928057789802551,
      "num_tokens": 3470254.0,
      "step": 7972
    },
    {
      "epoch": 0.6132605184216599,
      "grad_norm": 19.625,
      "learning_rate": 2.0801092223675102e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 3470590.0,
      "step": 7973
    },
    {
      "epoch": 0.6133374355818784,
      "grad_norm": 9.1875,
      "learning_rate": 2.0799938466271825e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3470896.0,
      "step": 7974
    },
    {
      "epoch": 0.6134143527420968,
      "grad_norm": 9.8125,
      "learning_rate": 2.0798784708868547e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 3471684.0,
      "step": 7975
    },
    {
      "epoch": 0.6134912699023152,
      "grad_norm": 9.5,
      "learning_rate": 2.0797630951465274e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 3471994.0,
      "step": 7976
    },
    {
      "epoch": 0.6135681870625337,
      "grad_norm": 10.8125,
      "learning_rate": 2.0796477194061997e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3472574.0,
      "step": 7977
    },
    {
      "epoch": 0.6136451042227521,
      "grad_norm": 14.25,
      "learning_rate": 2.079532343665872e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 3472890.0,
      "step": 7978
    },
    {
      "epoch": 0.6137220213829705,
      "grad_norm": 24.875,
      "learning_rate": 2.079416967925544e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 3473230.0,
      "step": 7979
    },
    {
      "epoch": 0.613798938543189,
      "grad_norm": 25.0,
      "learning_rate": 2.0793015921852165e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 3473580.0,
      "step": 7980
    },
    {
      "epoch": 0.6138758557034074,
      "grad_norm": 18.625,
      "learning_rate": 2.079186216444889e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 3474132.0,
      "step": 7981
    },
    {
      "epoch": 0.6139527728636259,
      "grad_norm": 8.1875,
      "learning_rate": 2.0790708407045614e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3474696.0,
      "step": 7982
    },
    {
      "epoch": 0.6140296900238443,
      "grad_norm": 4.21875,
      "learning_rate": 2.0789554649642337e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3475238.0,
      "step": 7983
    },
    {
      "epoch": 0.6141066071840627,
      "grad_norm": 65.0,
      "learning_rate": 2.078840089223906e-06,
      "loss": 0.0452,
      "mean_token_accuracy": 0.976190447807312,
      "num_tokens": 3475578.0,
      "step": 7984
    },
    {
      "epoch": 0.6141835243442813,
      "grad_norm": 13.125,
      "learning_rate": 2.078724713483578e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 3476094.0,
      "step": 7985
    },
    {
      "epoch": 0.6142604415044997,
      "grad_norm": 29.5,
      "learning_rate": 2.0786093377432504e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3476388.0,
      "step": 7986
    },
    {
      "epoch": 0.6143373586647181,
      "grad_norm": 33.0,
      "learning_rate": 2.078493962002923e-06,
      "loss": 0.05,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3476716.0,
      "step": 7987
    },
    {
      "epoch": 0.6144142758249366,
      "grad_norm": 20.375,
      "learning_rate": 2.0783785862625954e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3477236.0,
      "step": 7988
    },
    {
      "epoch": 0.614491192985155,
      "grad_norm": 26.125,
      "learning_rate": 2.0782632105222676e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3477534.0,
      "step": 7989
    },
    {
      "epoch": 0.6145681101453734,
      "grad_norm": 3.21875,
      "learning_rate": 2.07814783478194e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 3478104.0,
      "step": 7990
    },
    {
      "epoch": 0.6146450273055919,
      "grad_norm": 5.5,
      "learning_rate": 2.078032459041612e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 3478900.0,
      "step": 7991
    },
    {
      "epoch": 0.6147219444658103,
      "grad_norm": 26.0,
      "learning_rate": 2.0779170833012844e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 3479212.0,
      "step": 7992
    },
    {
      "epoch": 0.6147988616260288,
      "grad_norm": 6.90625,
      "learning_rate": 2.077801707560957e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3479532.0,
      "step": 7993
    },
    {
      "epoch": 0.6148757787862472,
      "grad_norm": 22.75,
      "learning_rate": 2.0776863318206293e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3479786.0,
      "step": 7994
    },
    {
      "epoch": 0.6149526959464656,
      "grad_norm": 20.125,
      "learning_rate": 2.0775709560803016e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3480376.0,
      "step": 7995
    },
    {
      "epoch": 0.6150296131066841,
      "grad_norm": 13.6875,
      "learning_rate": 2.077455580339974e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3480672.0,
      "step": 7996
    },
    {
      "epoch": 0.6151065302669025,
      "grad_norm": 12.375,
      "learning_rate": 2.077340204599646e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 3481244.0,
      "step": 7997
    },
    {
      "epoch": 0.6151834474271209,
      "grad_norm": 19.625,
      "learning_rate": 2.077224828859319e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3481792.0,
      "step": 7998
    },
    {
      "epoch": 0.6152603645873395,
      "grad_norm": 15.8125,
      "learning_rate": 2.077109453118991e-06,
      "loss": 0.0074,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 3482572.0,
      "step": 7999
    },
    {
      "epoch": 0.6153372817475579,
      "grad_norm": 6.8125,
      "learning_rate": 2.076994077378663e-06,
      "loss": 0.0188,
      "step": 8000
    },
    {
      "epoch": 0.6153372817475579,
      "eval_loss": 0.0191638320684433,
      "eval_mean_token_accuracy": 0.9923640036582947,
      "eval_num_tokens": 3482908.0,
      "eval_runtime": 21.0705,
      "eval_samples_per_second": 9.492,
      "eval_steps_per_second": 1.186,
      "step": 8000
    },
    {
      "epoch": 0.6154141989077764,
      "grad_norm": 27.625,
      "learning_rate": 2.0768787016383356e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9892784357070923,
      "num_tokens": 3483230.0,
      "step": 8001
    },
    {
      "epoch": 0.6154911160679948,
      "grad_norm": 32.0,
      "learning_rate": 2.076763325898008e-06,
      "loss": 0.0452,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3483536.0,
      "step": 8002
    },
    {
      "epoch": 0.6155680332282132,
      "grad_norm": 31.375,
      "learning_rate": 2.07664795015768e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3483794.0,
      "step": 8003
    },
    {
      "epoch": 0.6156449503884317,
      "grad_norm": 12.3125,
      "learning_rate": 2.0765325744173528e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3484338.0,
      "step": 8004
    },
    {
      "epoch": 0.6157218675486501,
      "grad_norm": 6.21875,
      "learning_rate": 2.0764171986770246e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9982847571372986,
      "num_tokens": 3484922.0,
      "step": 8005
    },
    {
      "epoch": 0.6157987847088685,
      "grad_norm": 35.5,
      "learning_rate": 2.0763018229366973e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3485206.0,
      "step": 8006
    },
    {
      "epoch": 0.615875701869087,
      "grad_norm": 24.875,
      "learning_rate": 2.0761864471963695e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3485516.0,
      "step": 8007
    },
    {
      "epoch": 0.6159526190293054,
      "grad_norm": 26.875,
      "learning_rate": 2.076071071456042e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3485850.0,
      "step": 8008
    },
    {
      "epoch": 0.6160295361895238,
      "grad_norm": 11.9375,
      "learning_rate": 2.0759556957157145e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3486374.0,
      "step": 8009
    },
    {
      "epoch": 0.6161064533497423,
      "grad_norm": 16.875,
      "learning_rate": 2.0758403199753867e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3486704.0,
      "step": 8010
    },
    {
      "epoch": 0.6161833705099607,
      "grad_norm": 17.75,
      "learning_rate": 2.0757249442350586e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3487272.0,
      "step": 8011
    },
    {
      "epoch": 0.6162602876701793,
      "grad_norm": 40.75,
      "learning_rate": 2.0756095684947313e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3487600.0,
      "step": 8012
    },
    {
      "epoch": 0.6163372048303977,
      "grad_norm": 34.0,
      "learning_rate": 2.0754941927544035e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3488134.0,
      "step": 8013
    },
    {
      "epoch": 0.6164141219906161,
      "grad_norm": 24.375,
      "learning_rate": 2.0753788170140758e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3488690.0,
      "step": 8014
    },
    {
      "epoch": 0.6164910391508346,
      "grad_norm": 18.25,
      "learning_rate": 2.0752634412737485e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 3489712.0,
      "step": 8015
    },
    {
      "epoch": 0.616567956311053,
      "grad_norm": 15.375,
      "learning_rate": 2.0751480655334203e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3489994.0,
      "step": 8016
    },
    {
      "epoch": 0.6166448734712714,
      "grad_norm": 30.625,
      "learning_rate": 2.075032689793093e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 3490274.0,
      "step": 8017
    },
    {
      "epoch": 0.6167217906314899,
      "grad_norm": 33.25,
      "learning_rate": 2.0749173140527652e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 3491070.0,
      "step": 8018
    },
    {
      "epoch": 0.6167987077917083,
      "grad_norm": 19.75,
      "learning_rate": 2.0748019383124375e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3491382.0,
      "step": 8019
    },
    {
      "epoch": 0.6168756249519268,
      "grad_norm": 10.875,
      "learning_rate": 2.07468656257211e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9969040155410767,
      "num_tokens": 3491706.0,
      "step": 8020
    },
    {
      "epoch": 0.6169525421121452,
      "grad_norm": 14.125,
      "learning_rate": 2.0745711868317824e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3492000.0,
      "step": 8021
    },
    {
      "epoch": 0.6170294592723636,
      "grad_norm": 6.5,
      "learning_rate": 2.0744558110914543e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3492332.0,
      "step": 8022
    },
    {
      "epoch": 0.6171063764325821,
      "grad_norm": 21.875,
      "learning_rate": 2.074340435351127e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3492626.0,
      "step": 8023
    },
    {
      "epoch": 0.6171832935928006,
      "grad_norm": 64.5,
      "learning_rate": 2.074225059610799e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3492928.0,
      "step": 8024
    },
    {
      "epoch": 0.617260210753019,
      "grad_norm": 30.125,
      "learning_rate": 2.0741096838704715e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3493212.0,
      "step": 8025
    },
    {
      "epoch": 0.6173371279132375,
      "grad_norm": 31.25,
      "learning_rate": 2.073994308130144e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 3493490.0,
      "step": 8026
    },
    {
      "epoch": 0.6174140450734559,
      "grad_norm": 26.75,
      "learning_rate": 2.073878932389816e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 3493776.0,
      "step": 8027
    },
    {
      "epoch": 0.6174909622336743,
      "grad_norm": 8.25,
      "learning_rate": 2.0737635566494887e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 3494590.0,
      "step": 8028
    },
    {
      "epoch": 0.6175678793938928,
      "grad_norm": 18.25,
      "learning_rate": 2.073648180909161e-06,
      "loss": 0.0074,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 3495634.0,
      "step": 8029
    },
    {
      "epoch": 0.6176447965541112,
      "grad_norm": 26.375,
      "learning_rate": 2.073532805168833e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 3496418.0,
      "step": 8030
    },
    {
      "epoch": 0.6177217137143297,
      "grad_norm": 13.4375,
      "learning_rate": 2.073417429428506e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9971346855163574,
      "num_tokens": 3496768.0,
      "step": 8031
    },
    {
      "epoch": 0.6177986308745481,
      "grad_norm": 21.5,
      "learning_rate": 2.073302053688178e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3497114.0,
      "step": 8032
    },
    {
      "epoch": 0.6178755480347665,
      "grad_norm": 11.6875,
      "learning_rate": 2.07318667794785e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3497660.0,
      "step": 8033
    },
    {
      "epoch": 0.617952465194985,
      "grad_norm": 8.5,
      "learning_rate": 2.0730713022075226e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3498206.0,
      "step": 8034
    },
    {
      "epoch": 0.6180293823552034,
      "grad_norm": 20.0,
      "learning_rate": 2.072955926467195e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 3498494.0,
      "step": 8035
    },
    {
      "epoch": 0.6181062995154218,
      "grad_norm": 9.8125,
      "learning_rate": 2.072840550726867e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3499038.0,
      "step": 8036
    },
    {
      "epoch": 0.6181832166756404,
      "grad_norm": 48.5,
      "learning_rate": 2.07272517498654e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3499348.0,
      "step": 8037
    },
    {
      "epoch": 0.6182601338358588,
      "grad_norm": 11.25,
      "learning_rate": 2.0726097992462117e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3499892.0,
      "step": 8038
    },
    {
      "epoch": 0.6183370509960773,
      "grad_norm": 5.125,
      "learning_rate": 2.0724944235058843e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3500440.0,
      "step": 8039
    },
    {
      "epoch": 0.6184139681562957,
      "grad_norm": 18.0,
      "learning_rate": 2.0723790477655566e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 3500726.0,
      "step": 8040
    },
    {
      "epoch": 0.6184908853165141,
      "grad_norm": 36.25,
      "learning_rate": 2.072263672025229e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 3501032.0,
      "step": 8041
    },
    {
      "epoch": 0.6185678024767326,
      "grad_norm": 30.125,
      "learning_rate": 2.072148296284901e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3501344.0,
      "step": 8042
    },
    {
      "epoch": 0.618644719636951,
      "grad_norm": 25.0,
      "learning_rate": 2.0720329205445734e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3501626.0,
      "step": 8043
    },
    {
      "epoch": 0.6187216367971694,
      "grad_norm": 10.875,
      "learning_rate": 2.0719175448042456e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3502180.0,
      "step": 8044
    },
    {
      "epoch": 0.6187985539573879,
      "grad_norm": 11.6875,
      "learning_rate": 2.0718021690639183e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3502492.0,
      "step": 8045
    },
    {
      "epoch": 0.6188754711176063,
      "grad_norm": 13.3125,
      "learning_rate": 2.0716867933235906e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3502776.0,
      "step": 8046
    },
    {
      "epoch": 0.6189523882778247,
      "grad_norm": 9.0,
      "learning_rate": 2.071571417583263e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3503088.0,
      "step": 8047
    },
    {
      "epoch": 0.6190293054380432,
      "grad_norm": 20.375,
      "learning_rate": 2.0714560418429355e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 3503368.0,
      "step": 8048
    },
    {
      "epoch": 0.6191062225982616,
      "grad_norm": 14.25,
      "learning_rate": 2.0713406661026073e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3503890.0,
      "step": 8049
    },
    {
      "epoch": 0.6191831397584802,
      "grad_norm": 36.25,
      "learning_rate": 2.0712252903622796e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3504452.0,
      "step": 8050
    },
    {
      "epoch": 0.6192600569186986,
      "grad_norm": 31.125,
      "learning_rate": 2.0711099146219523e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3504770.0,
      "step": 8051
    },
    {
      "epoch": 0.619336974078917,
      "grad_norm": 62.5,
      "learning_rate": 2.0709945388816245e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 3505108.0,
      "step": 8052
    },
    {
      "epoch": 0.6194138912391355,
      "grad_norm": 43.5,
      "learning_rate": 2.070879163141297e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3505408.0,
      "step": 8053
    },
    {
      "epoch": 0.6194908083993539,
      "grad_norm": 17.5,
      "learning_rate": 2.070763787400969e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3505948.0,
      "step": 8054
    },
    {
      "epoch": 0.6195677255595723,
      "grad_norm": 15.6875,
      "learning_rate": 2.0706484116606413e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3506268.0,
      "step": 8055
    },
    {
      "epoch": 0.6196446427197908,
      "grad_norm": 13.1875,
      "learning_rate": 2.070533035920314e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3506614.0,
      "step": 8056
    },
    {
      "epoch": 0.6197215598800092,
      "grad_norm": 14.3125,
      "learning_rate": 2.0704176601799863e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9971510171890259,
      "num_tokens": 3506966.0,
      "step": 8057
    },
    {
      "epoch": 0.6197984770402277,
      "grad_norm": 10.3125,
      "learning_rate": 2.0703022844396585e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3507314.0,
      "step": 8058
    },
    {
      "epoch": 0.6198753942004461,
      "grad_norm": 25.5,
      "learning_rate": 2.070186908699331e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3507570.0,
      "step": 8059
    },
    {
      "epoch": 0.6199523113606645,
      "grad_norm": 13.8125,
      "learning_rate": 2.070071532959003e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 3507874.0,
      "step": 8060
    },
    {
      "epoch": 0.620029228520883,
      "grad_norm": 17.5,
      "learning_rate": 2.0699561572186753e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 3508170.0,
      "step": 8061
    },
    {
      "epoch": 0.6201061456811015,
      "grad_norm": 10.6875,
      "learning_rate": 2.069840781478348e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 3508754.0,
      "step": 8062
    },
    {
      "epoch": 0.6201830628413199,
      "grad_norm": 10.375,
      "learning_rate": 2.0697254057380202e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 3509026.0,
      "step": 8063
    },
    {
      "epoch": 0.6202599800015384,
      "grad_norm": 3.796875,
      "learning_rate": 2.0696100299976925e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 3509812.0,
      "step": 8064
    },
    {
      "epoch": 0.6203368971617568,
      "grad_norm": 9.125,
      "learning_rate": 2.0694946542573647e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3510362.0,
      "step": 8065
    },
    {
      "epoch": 0.6204138143219753,
      "grad_norm": 9.375,
      "learning_rate": 2.069379278517037e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 3511166.0,
      "step": 8066
    },
    {
      "epoch": 0.6204907314821937,
      "grad_norm": 22.625,
      "learning_rate": 2.0692639027767097e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3511448.0,
      "step": 8067
    },
    {
      "epoch": 0.6205676486424121,
      "grad_norm": 18.75,
      "learning_rate": 2.069148527036382e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9930070042610168,
      "num_tokens": 3511736.0,
      "step": 8068
    },
    {
      "epoch": 0.6206445658026306,
      "grad_norm": 11.25,
      "learning_rate": 2.069033151296054e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3512080.0,
      "step": 8069
    },
    {
      "epoch": 0.620721482962849,
      "grad_norm": 10.25,
      "learning_rate": 2.068917775555727e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3512654.0,
      "step": 8070
    },
    {
      "epoch": 0.6207984001230674,
      "grad_norm": 43.0,
      "learning_rate": 2.0688023998153987e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 3512982.0,
      "step": 8071
    },
    {
      "epoch": 0.6208753172832859,
      "grad_norm": 24.75,
      "learning_rate": 2.068687024075071e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 3513314.0,
      "step": 8072
    },
    {
      "epoch": 0.6209522344435043,
      "grad_norm": 16.25,
      "learning_rate": 2.0685716483347437e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3513632.0,
      "step": 8073
    },
    {
      "epoch": 0.6210291516037227,
      "grad_norm": 12.0625,
      "learning_rate": 2.068456272594416e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 3513950.0,
      "step": 8074
    },
    {
      "epoch": 0.6211060687639413,
      "grad_norm": 6.96875,
      "learning_rate": 2.068340896854088e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3514260.0,
      "step": 8075
    },
    {
      "epoch": 0.6211829859241597,
      "grad_norm": 16.0,
      "learning_rate": 2.0682255211137604e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3514568.0,
      "step": 8076
    },
    {
      "epoch": 0.6212599030843782,
      "grad_norm": 20.0,
      "learning_rate": 2.0681101453734327e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 3515376.0,
      "step": 8077
    },
    {
      "epoch": 0.6213368202445966,
      "grad_norm": 12.3125,
      "learning_rate": 2.0679947696331054e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3515922.0,
      "step": 8078
    },
    {
      "epoch": 0.621413737404815,
      "grad_norm": 8.8125,
      "learning_rate": 2.0678793938927776e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3516234.0,
      "step": 8079
    },
    {
      "epoch": 0.6214906545650335,
      "grad_norm": 9.3125,
      "learning_rate": 2.06776401815245e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3516570.0,
      "step": 8080
    },
    {
      "epoch": 0.6215675717252519,
      "grad_norm": 5.6875,
      "learning_rate": 2.067648642412122e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 3517356.0,
      "step": 8081
    },
    {
      "epoch": 0.6216444888854703,
      "grad_norm": 5.65625,
      "learning_rate": 2.0675332666717944e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3517930.0,
      "step": 8082
    },
    {
      "epoch": 0.6217214060456888,
      "grad_norm": 9.9375,
      "learning_rate": 2.0674178909314667e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 3518740.0,
      "step": 8083
    },
    {
      "epoch": 0.6217983232059072,
      "grad_norm": 12.6875,
      "learning_rate": 2.0673025151911393e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 3519316.0,
      "step": 8084
    },
    {
      "epoch": 0.6218752403661257,
      "grad_norm": 36.0,
      "learning_rate": 2.0671871394508116e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3519632.0,
      "step": 8085
    },
    {
      "epoch": 0.6219521575263441,
      "grad_norm": 34.75,
      "learning_rate": 2.067071763710484e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3519886.0,
      "step": 8086
    },
    {
      "epoch": 0.6220290746865625,
      "grad_norm": 21.625,
      "learning_rate": 2.066956387970156e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3520204.0,
      "step": 8087
    },
    {
      "epoch": 0.6221059918467811,
      "grad_norm": 19.75,
      "learning_rate": 2.0668410122298284e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3520522.0,
      "step": 8088
    },
    {
      "epoch": 0.6221829090069995,
      "grad_norm": 18.25,
      "learning_rate": 2.066725636489501e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 3520840.0,
      "step": 8089
    },
    {
      "epoch": 0.6222598261672179,
      "grad_norm": 4.46875,
      "learning_rate": 2.0666102607491733e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 3521636.0,
      "step": 8090
    },
    {
      "epoch": 0.6223367433274364,
      "grad_norm": 5.84375,
      "learning_rate": 2.0664948850088456e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9974554777145386,
      "num_tokens": 3522424.0,
      "step": 8091
    },
    {
      "epoch": 0.6224136604876548,
      "grad_norm": 8.0625,
      "learning_rate": 2.066379509268518e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3522746.0,
      "step": 8092
    },
    {
      "epoch": 0.6224905776478732,
      "grad_norm": 12.25,
      "learning_rate": 2.06626413352819e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 3523324.0,
      "step": 8093
    },
    {
      "epoch": 0.6225674948080917,
      "grad_norm": 6.15625,
      "learning_rate": 2.0661487577878623e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3523870.0,
      "step": 8094
    },
    {
      "epoch": 0.6226444119683101,
      "grad_norm": 4.65625,
      "learning_rate": 2.066033382047535e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 3524650.0,
      "step": 8095
    },
    {
      "epoch": 0.6227213291285286,
      "grad_norm": 37.5,
      "learning_rate": 2.0659180063072073e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3524910.0,
      "step": 8096
    },
    {
      "epoch": 0.622798246288747,
      "grad_norm": 10.8125,
      "learning_rate": 2.0658026305668795e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3525208.0,
      "step": 8097
    },
    {
      "epoch": 0.6228751634489654,
      "grad_norm": 17.5,
      "learning_rate": 2.065687254826552e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3525520.0,
      "step": 8098
    },
    {
      "epoch": 0.622952080609184,
      "grad_norm": 9.5625,
      "learning_rate": 2.065571879086224e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 3525854.0,
      "step": 8099
    },
    {
      "epoch": 0.6230289977694023,
      "grad_norm": 4.03125,
      "learning_rate": 2.0654565033458963e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3526418.0,
      "step": 8100
    },
    {
      "epoch": 0.6231059149296208,
      "grad_norm": 28.875,
      "learning_rate": 2.065341127605569e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9930070042610168,
      "num_tokens": 3526706.0,
      "step": 8101
    },
    {
      "epoch": 0.6231828320898393,
      "grad_norm": 21.375,
      "learning_rate": 2.0652257518652413e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3527022.0,
      "step": 8102
    },
    {
      "epoch": 0.6232597492500577,
      "grad_norm": 16.25,
      "learning_rate": 2.0651103761249135e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 3527330.0,
      "step": 8103
    },
    {
      "epoch": 0.6233366664102762,
      "grad_norm": 33.0,
      "learning_rate": 2.0649950003845858e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 3527862.0,
      "step": 8104
    },
    {
      "epoch": 0.6234135835704946,
      "grad_norm": 22.375,
      "learning_rate": 2.064879624644258e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 3528642.0,
      "step": 8105
    },
    {
      "epoch": 0.623490500730713,
      "grad_norm": 30.625,
      "learning_rate": 2.0647642489039307e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3528972.0,
      "step": 8106
    },
    {
      "epoch": 0.6235674178909315,
      "grad_norm": 63.25,
      "learning_rate": 2.064648873163603e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3529276.0,
      "step": 8107
    },
    {
      "epoch": 0.6236443350511499,
      "grad_norm": 82.0,
      "learning_rate": 2.0645334974232752e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 3529648.0,
      "step": 8108
    },
    {
      "epoch": 0.6237212522113683,
      "grad_norm": 34.0,
      "learning_rate": 2.0644181216829475e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3529970.0,
      "step": 8109
    },
    {
      "epoch": 0.6237981693715868,
      "grad_norm": 11.0625,
      "learning_rate": 2.0643027459426197e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 3530310.0,
      "step": 8110
    },
    {
      "epoch": 0.6238750865318052,
      "grad_norm": 25.0,
      "learning_rate": 2.064187370202292e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3530856.0,
      "step": 8111
    },
    {
      "epoch": 0.6239520036920236,
      "grad_norm": 12.9375,
      "learning_rate": 2.0640719944619647e-06,
      "loss": 0.0069,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 3531636.0,
      "step": 8112
    },
    {
      "epoch": 0.6240289208522422,
      "grad_norm": 18.5,
      "learning_rate": 2.063956618721637e-06,
      "loss": 0.0041,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3532920.0,
      "step": 8113
    },
    {
      "epoch": 0.6241058380124606,
      "grad_norm": 51.0,
      "learning_rate": 2.063841242981309e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3533236.0,
      "step": 8114
    },
    {
      "epoch": 0.6241827551726791,
      "grad_norm": 34.5,
      "learning_rate": 2.0637258672409815e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3533570.0,
      "step": 8115
    },
    {
      "epoch": 0.6242596723328975,
      "grad_norm": 10.1875,
      "learning_rate": 2.0636104915006537e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3533894.0,
      "step": 8116
    },
    {
      "epoch": 0.6243365894931159,
      "grad_norm": 25.625,
      "learning_rate": 2.0634951157603264e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3534234.0,
      "step": 8117
    },
    {
      "epoch": 0.6244135066533344,
      "grad_norm": 48.25,
      "learning_rate": 2.0633797400199987e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3534532.0,
      "step": 8118
    },
    {
      "epoch": 0.6244904238135528,
      "grad_norm": 9.3125,
      "learning_rate": 2.0632643642796705e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3535106.0,
      "step": 8119
    },
    {
      "epoch": 0.6245673409737712,
      "grad_norm": 60.25,
      "learning_rate": 2.063148988539343e-06,
      "loss": 0.0478,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 3535438.0,
      "step": 8120
    },
    {
      "epoch": 0.6246442581339897,
      "grad_norm": 9.1875,
      "learning_rate": 2.0630336127990154e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3535990.0,
      "step": 8121
    },
    {
      "epoch": 0.6247211752942081,
      "grad_norm": 56.75,
      "learning_rate": 2.0629182370586877e-06,
      "loss": 0.0435,
      "mean_token_accuracy": 0.9788732528686523,
      "num_tokens": 3536276.0,
      "step": 8122
    },
    {
      "epoch": 0.6247980924544266,
      "grad_norm": 34.25,
      "learning_rate": 2.0628028613183604e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3536606.0,
      "step": 8123
    },
    {
      "epoch": 0.624875009614645,
      "grad_norm": 17.25,
      "learning_rate": 2.0626874855780326e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3536924.0,
      "step": 8124
    },
    {
      "epoch": 0.6249519267748634,
      "grad_norm": 33.75,
      "learning_rate": 2.062572109837705e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3537486.0,
      "step": 8125
    },
    {
      "epoch": 0.625028843935082,
      "grad_norm": 87.0,
      "learning_rate": 2.062456734097377e-06,
      "loss": 0.0371,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 3537788.0,
      "step": 8126
    },
    {
      "epoch": 0.6251057610953004,
      "grad_norm": 13.0625,
      "learning_rate": 2.0623413583570494e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3538104.0,
      "step": 8127
    },
    {
      "epoch": 0.6251826782555188,
      "grad_norm": 26.75,
      "learning_rate": 2.062225982616722e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3538700.0,
      "step": 8128
    },
    {
      "epoch": 0.6252595954157373,
      "grad_norm": 13.25,
      "learning_rate": 2.0621106068763943e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3539270.0,
      "step": 8129
    },
    {
      "epoch": 0.6253365125759557,
      "grad_norm": 28.75,
      "learning_rate": 2.061995231136066e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9971751570701599,
      "num_tokens": 3539626.0,
      "step": 8130
    },
    {
      "epoch": 0.6254134297361741,
      "grad_norm": 22.875,
      "learning_rate": 2.061879855395739e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3539966.0,
      "step": 8131
    },
    {
      "epoch": 0.6254903468963926,
      "grad_norm": 26.125,
      "learning_rate": 2.061764479655411e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 3540524.0,
      "step": 8132
    },
    {
      "epoch": 0.625567264056611,
      "grad_norm": 20.5,
      "learning_rate": 2.0616491039150834e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3540854.0,
      "step": 8133
    },
    {
      "epoch": 0.6256441812168295,
      "grad_norm": 23.75,
      "learning_rate": 2.061533728174756e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3541134.0,
      "step": 8134
    },
    {
      "epoch": 0.6257210983770479,
      "grad_norm": 7.40625,
      "learning_rate": 2.0614183524344283e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9961758852005005,
      "num_tokens": 3541658.0,
      "step": 8135
    },
    {
      "epoch": 0.6257980155372663,
      "grad_norm": 14.8125,
      "learning_rate": 2.0613029766941006e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 3541946.0,
      "step": 8136
    },
    {
      "epoch": 0.6258749326974848,
      "grad_norm": 12.125,
      "learning_rate": 2.061187600953773e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3542496.0,
      "step": 8137
    },
    {
      "epoch": 0.6259518498577032,
      "grad_norm": 19.375,
      "learning_rate": 2.061072225213445e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3543050.0,
      "step": 8138
    },
    {
      "epoch": 0.6260287670179217,
      "grad_norm": 7.0625,
      "learning_rate": 2.0609568494731178e-06,
      "loss": 0.0065,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 3543876.0,
      "step": 8139
    },
    {
      "epoch": 0.6261056841781402,
      "grad_norm": 47.25,
      "learning_rate": 2.06084147373279e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3544154.0,
      "step": 8140
    },
    {
      "epoch": 0.6261826013383586,
      "grad_norm": 20.625,
      "learning_rate": 2.060726097992462e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3544452.0,
      "step": 8141
    },
    {
      "epoch": 0.6262595184985771,
      "grad_norm": 24.625,
      "learning_rate": 2.0606107222521345e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9971014261245728,
      "num_tokens": 3544798.0,
      "step": 8142
    },
    {
      "epoch": 0.6263364356587955,
      "grad_norm": 9.9375,
      "learning_rate": 2.060495346511807e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3545058.0,
      "step": 8143
    },
    {
      "epoch": 0.6264133528190139,
      "grad_norm": 7.09375,
      "learning_rate": 2.060379970771479e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 3545382.0,
      "step": 8144
    },
    {
      "epoch": 0.6264902699792324,
      "grad_norm": 5.78125,
      "learning_rate": 2.0602645950311517e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9981784820556641,
      "num_tokens": 3545932.0,
      "step": 8145
    },
    {
      "epoch": 0.6265671871394508,
      "grad_norm": 17.625,
      "learning_rate": 2.0601492192908236e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 3546242.0,
      "step": 8146
    },
    {
      "epoch": 0.6266441042996692,
      "grad_norm": 8.75,
      "learning_rate": 2.0600338435504963e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 3546810.0,
      "step": 8147
    },
    {
      "epoch": 0.6267210214598877,
      "grad_norm": 7.5625,
      "learning_rate": 2.0599184678101685e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3547396.0,
      "step": 8148
    },
    {
      "epoch": 0.6267979386201061,
      "grad_norm": 9.6875,
      "learning_rate": 2.0598030920698408e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3547954.0,
      "step": 8149
    },
    {
      "epoch": 0.6268748557803245,
      "grad_norm": 44.25,
      "learning_rate": 2.059687716329513e-06,
      "loss": 0.051,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 3548256.0,
      "step": 8150
    },
    {
      "epoch": 0.626951772940543,
      "grad_norm": 38.75,
      "learning_rate": 2.0595723405891857e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3548576.0,
      "step": 8151
    },
    {
      "epoch": 0.6270286901007615,
      "grad_norm": 28.125,
      "learning_rate": 2.0594569648488576e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3548856.0,
      "step": 8152
    },
    {
      "epoch": 0.62710560726098,
      "grad_norm": 21.125,
      "learning_rate": 2.0593415891085302e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3549178.0,
      "step": 8153
    },
    {
      "epoch": 0.6271825244211984,
      "grad_norm": 18.875,
      "learning_rate": 2.0592262133682025e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 3549694.0,
      "step": 8154
    },
    {
      "epoch": 0.6272594415814168,
      "grad_norm": 9.0,
      "learning_rate": 2.0591108376278748e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3550264.0,
      "step": 8155
    },
    {
      "epoch": 0.6273363587416353,
      "grad_norm": 14.75,
      "learning_rate": 2.0589954618875474e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3550598.0,
      "step": 8156
    },
    {
      "epoch": 0.6274132759018537,
      "grad_norm": 11.375,
      "learning_rate": 2.0588800861472193e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3551152.0,
      "step": 8157
    },
    {
      "epoch": 0.6274901930620721,
      "grad_norm": 37.5,
      "learning_rate": 2.0587647104068915e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3551434.0,
      "step": 8158
    },
    {
      "epoch": 0.6275671102222906,
      "grad_norm": 22.125,
      "learning_rate": 2.058649334666564e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 3552258.0,
      "step": 8159
    },
    {
      "epoch": 0.627644027382509,
      "grad_norm": 39.5,
      "learning_rate": 2.0585339589262365e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3552842.0,
      "step": 8160
    },
    {
      "epoch": 0.6277209445427275,
      "grad_norm": 4.84375,
      "learning_rate": 2.0584185831859087e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 3553386.0,
      "step": 8161
    },
    {
      "epoch": 0.6277978617029459,
      "grad_norm": 20.125,
      "learning_rate": 2.0583032074455814e-06,
      "loss": 0.0069,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 3554190.0,
      "step": 8162
    },
    {
      "epoch": 0.6278747788631643,
      "grad_norm": 5.71875,
      "learning_rate": 2.0581878317052532e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3554742.0,
      "step": 8163
    },
    {
      "epoch": 0.6279516960233829,
      "grad_norm": 13.375,
      "learning_rate": 2.058072455964926e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3555306.0,
      "step": 8164
    },
    {
      "epoch": 0.6280286131836013,
      "grad_norm": 18.75,
      "learning_rate": 2.057957080224598e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 3555606.0,
      "step": 8165
    },
    {
      "epoch": 0.6281055303438197,
      "grad_norm": 13.1875,
      "learning_rate": 2.0578417044842704e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3555932.0,
      "step": 8166
    },
    {
      "epoch": 0.6281824475040382,
      "grad_norm": 27.625,
      "learning_rate": 2.057726328743943e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3556188.0,
      "step": 8167
    },
    {
      "epoch": 0.6282593646642566,
      "grad_norm": 17.5,
      "learning_rate": 2.057610953003615e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3556474.0,
      "step": 8168
    },
    {
      "epoch": 0.6283362818244751,
      "grad_norm": 6.6875,
      "learning_rate": 2.0574955772632872e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.998266875743866,
      "num_tokens": 3557052.0,
      "step": 8169
    },
    {
      "epoch": 0.6284131989846935,
      "grad_norm": 13.0625,
      "learning_rate": 2.05738020152296e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3557638.0,
      "step": 8170
    },
    {
      "epoch": 0.6284901161449119,
      "grad_norm": 8.3125,
      "learning_rate": 2.057264825782632e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3558224.0,
      "step": 8171
    },
    {
      "epoch": 0.6285670333051304,
      "grad_norm": 23.625,
      "learning_rate": 2.0571494500423044e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 3558558.0,
      "step": 8172
    },
    {
      "epoch": 0.6286439504653488,
      "grad_norm": 5.09375,
      "learning_rate": 2.057034074301977e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.994575023651123,
      "num_tokens": 3559112.0,
      "step": 8173
    },
    {
      "epoch": 0.6287208676255672,
      "grad_norm": 10.25,
      "learning_rate": 2.056918698561649e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3559436.0,
      "step": 8174
    },
    {
      "epoch": 0.6287977847857857,
      "grad_norm": 5.8125,
      "learning_rate": 2.0568033228213216e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3560016.0,
      "step": 8175
    },
    {
      "epoch": 0.6288747019460041,
      "grad_norm": 10.25,
      "learning_rate": 2.056687947080994e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3560362.0,
      "step": 8176
    },
    {
      "epoch": 0.6289516191062225,
      "grad_norm": 8.1875,
      "learning_rate": 2.056572571340666e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 3560932.0,
      "step": 8177
    },
    {
      "epoch": 0.6290285362664411,
      "grad_norm": 23.125,
      "learning_rate": 2.056457195600339e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 3561280.0,
      "step": 8178
    },
    {
      "epoch": 0.6291054534266595,
      "grad_norm": 15.4375,
      "learning_rate": 2.0563418198600106e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 3561796.0,
      "step": 8179
    },
    {
      "epoch": 0.629182370586878,
      "grad_norm": 2.796875,
      "learning_rate": 2.056226444119683e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3562380.0,
      "step": 8180
    },
    {
      "epoch": 0.6292592877470964,
      "grad_norm": 9.375,
      "learning_rate": 2.0561110683793556e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9987244606018066,
      "num_tokens": 3563166.0,
      "step": 8181
    },
    {
      "epoch": 0.6293362049073148,
      "grad_norm": 19.5,
      "learning_rate": 2.055995692639028e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3563506.0,
      "step": 8182
    },
    {
      "epoch": 0.6294131220675333,
      "grad_norm": 32.0,
      "learning_rate": 2.0558803168987e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3563824.0,
      "step": 8183
    },
    {
      "epoch": 0.6294900392277517,
      "grad_norm": 3.765625,
      "learning_rate": 2.0557649411583724e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 3564398.0,
      "step": 8184
    },
    {
      "epoch": 0.6295669563879701,
      "grad_norm": 85.0,
      "learning_rate": 2.0556495654180446e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3564656.0,
      "step": 8185
    },
    {
      "epoch": 0.6296438735481886,
      "grad_norm": 44.75,
      "learning_rate": 2.0555341896777173e-06,
      "loss": 0.04,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3564938.0,
      "step": 8186
    },
    {
      "epoch": 0.629720790708407,
      "grad_norm": 10.875,
      "learning_rate": 2.0554188139373896e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3565266.0,
      "step": 8187
    },
    {
      "epoch": 0.6297977078686255,
      "grad_norm": 26.875,
      "learning_rate": 2.055303438197062e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3565564.0,
      "step": 8188
    },
    {
      "epoch": 0.629874625028844,
      "grad_norm": 3.875,
      "learning_rate": 2.0551880624567345e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 3566344.0,
      "step": 8189
    },
    {
      "epoch": 0.6299515421890624,
      "grad_norm": 49.0,
      "learning_rate": 2.0550726867164063e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3566940.0,
      "step": 8190
    },
    {
      "epoch": 0.6300284593492809,
      "grad_norm": 12.9375,
      "learning_rate": 2.0549573109760786e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3567486.0,
      "step": 8191
    },
    {
      "epoch": 0.6301053765094993,
      "grad_norm": 31.875,
      "learning_rate": 2.0548419352357513e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 3567822.0,
      "step": 8192
    },
    {
      "epoch": 0.6301822936697177,
      "grad_norm": 19.375,
      "learning_rate": 2.0547265594954235e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 3568112.0,
      "step": 8193
    },
    {
      "epoch": 0.6302592108299362,
      "grad_norm": 34.25,
      "learning_rate": 2.0546111837550958e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9975308775901794,
      "num_tokens": 3568924.0,
      "step": 8194
    },
    {
      "epoch": 0.6303361279901546,
      "grad_norm": 36.75,
      "learning_rate": 2.054495808014768e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9975308775901794,
      "num_tokens": 3569736.0,
      "step": 8195
    },
    {
      "epoch": 0.630413045150373,
      "grad_norm": 101.5,
      "learning_rate": 2.0543804322744403e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3570046.0,
      "step": 8196
    },
    {
      "epoch": 0.6304899623105915,
      "grad_norm": 20.375,
      "learning_rate": 2.054265056534113e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3570396.0,
      "step": 8197
    },
    {
      "epoch": 0.6305668794708099,
      "grad_norm": 37.25,
      "learning_rate": 2.0541496807937852e-06,
      "loss": 0.0375,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3570656.0,
      "step": 8198
    },
    {
      "epoch": 0.6306437966310284,
      "grad_norm": 26.25,
      "learning_rate": 2.0540343050534575e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3571220.0,
      "step": 8199
    },
    {
      "epoch": 0.6307207137912468,
      "grad_norm": 64.5,
      "learning_rate": 2.05391892931313e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9941520690917969,
      "num_tokens": 3571564.0,
      "step": 8200
    },
    {
      "epoch": 0.6307976309514652,
      "grad_norm": 15.875,
      "learning_rate": 2.053803553572802e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 3571914.0,
      "step": 8201
    },
    {
      "epoch": 0.6308745481116838,
      "grad_norm": 4.28125,
      "learning_rate": 2.0536881778324743e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3572470.0,
      "step": 8202
    },
    {
      "epoch": 0.6309514652719022,
      "grad_norm": 10.125,
      "learning_rate": 2.053572802092147e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9987244606018066,
      "num_tokens": 3573256.0,
      "step": 8203
    },
    {
      "epoch": 0.6310283824321206,
      "grad_norm": 47.5,
      "learning_rate": 2.053457426351819e-06,
      "loss": 0.0486,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3573562.0,
      "step": 8204
    },
    {
      "epoch": 0.6311052995923391,
      "grad_norm": 10.125,
      "learning_rate": 2.0533420506114915e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 3574134.0,
      "step": 8205
    },
    {
      "epoch": 0.6311822167525575,
      "grad_norm": 31.75,
      "learning_rate": 2.0532266748711637e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3574668.0,
      "step": 8206
    },
    {
      "epoch": 0.631259133912776,
      "grad_norm": 13.875,
      "learning_rate": 2.053111299130836e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3575214.0,
      "step": 8207
    },
    {
      "epoch": 0.6313360510729944,
      "grad_norm": 21.75,
      "learning_rate": 2.0529959233905087e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3575508.0,
      "step": 8208
    },
    {
      "epoch": 0.6314129682332128,
      "grad_norm": 16.75,
      "learning_rate": 2.052880547650181e-06,
      "loss": 0.0408,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 3575760.0,
      "step": 8209
    },
    {
      "epoch": 0.6314898853934313,
      "grad_norm": 5.78125,
      "learning_rate": 2.052765171909853e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 3576540.0,
      "step": 8210
    },
    {
      "epoch": 0.6315668025536497,
      "grad_norm": 7.46875,
      "learning_rate": 2.0526497961695254e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 3576906.0,
      "step": 8211
    },
    {
      "epoch": 0.6316437197138681,
      "grad_norm": 7.09375,
      "learning_rate": 2.0525344204291977e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3577480.0,
      "step": 8212
    },
    {
      "epoch": 0.6317206368740866,
      "grad_norm": 4.34375,
      "learning_rate": 2.05241904468887e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9974358677864075,
      "num_tokens": 3578262.0,
      "step": 8213
    },
    {
      "epoch": 0.631797554034305,
      "grad_norm": 8.75,
      "learning_rate": 2.0523036689485426e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3578588.0,
      "step": 8214
    },
    {
      "epoch": 0.6318744711945234,
      "grad_norm": 64.5,
      "learning_rate": 2.052188293208215e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3578894.0,
      "step": 8215
    },
    {
      "epoch": 0.631951388354742,
      "grad_norm": 32.5,
      "learning_rate": 2.052072917467887e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3579174.0,
      "step": 8216
    },
    {
      "epoch": 0.6320283055149604,
      "grad_norm": 22.375,
      "learning_rate": 2.0519575417275594e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3579490.0,
      "step": 8217
    },
    {
      "epoch": 0.6321052226751789,
      "grad_norm": 54.25,
      "learning_rate": 2.0518421659872317e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 3579762.0,
      "step": 8218
    },
    {
      "epoch": 0.6321821398353973,
      "grad_norm": 78.0,
      "learning_rate": 2.051726790246904e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 3580124.0,
      "step": 8219
    },
    {
      "epoch": 0.6322590569956157,
      "grad_norm": 19.875,
      "learning_rate": 2.0516114145065766e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 3580640.0,
      "step": 8220
    },
    {
      "epoch": 0.6323359741558342,
      "grad_norm": 54.0,
      "learning_rate": 2.051496038766249e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3580946.0,
      "step": 8221
    },
    {
      "epoch": 0.6324128913160526,
      "grad_norm": 9.5625,
      "learning_rate": 2.051380663025921e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3581492.0,
      "step": 8222
    },
    {
      "epoch": 0.632489808476271,
      "grad_norm": 48.5,
      "learning_rate": 2.0512652872855934e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 3581808.0,
      "step": 8223
    },
    {
      "epoch": 0.6325667256364895,
      "grad_norm": 42.75,
      "learning_rate": 2.0511499115452656e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 3582404.0,
      "step": 8224
    },
    {
      "epoch": 0.6326436427967079,
      "grad_norm": 28.75,
      "learning_rate": 2.0510345358049383e-06,
      "loss": 0.045,
      "mean_token_accuracy": 0.978787899017334,
      "num_tokens": 3582736.0,
      "step": 8225
    },
    {
      "epoch": 0.6327205599569264,
      "grad_norm": 54.0,
      "learning_rate": 2.0509191600646106e-06,
      "loss": 0.0425,
      "mean_token_accuracy": 0.9821428656578064,
      "num_tokens": 3582964.0,
      "step": 8226
    },
    {
      "epoch": 0.6327974771171448,
      "grad_norm": 4.71875,
      "learning_rate": 2.050803784324283e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9975429773330688,
      "num_tokens": 3583780.0,
      "step": 8227
    },
    {
      "epoch": 0.6328743942773633,
      "grad_norm": 40.75,
      "learning_rate": 2.050688408583955e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3584090.0,
      "step": 8228
    },
    {
      "epoch": 0.6329513114375818,
      "grad_norm": 29.25,
      "learning_rate": 2.0505730328436274e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3584400.0,
      "step": 8229
    },
    {
      "epoch": 0.6330282285978002,
      "grad_norm": 35.25,
      "learning_rate": 2.0504576571032996e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3585182.0,
      "step": 8230
    },
    {
      "epoch": 0.6331051457580186,
      "grad_norm": 21.625,
      "learning_rate": 2.0503422813629723e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 3585978.0,
      "step": 8231
    },
    {
      "epoch": 0.6331820629182371,
      "grad_norm": 28.875,
      "learning_rate": 2.0502269056226446e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 3586534.0,
      "step": 8232
    },
    {
      "epoch": 0.6332589800784555,
      "grad_norm": 38.0,
      "learning_rate": 2.050111529882317e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3586846.0,
      "step": 8233
    },
    {
      "epoch": 0.6333358972386739,
      "grad_norm": 32.0,
      "learning_rate": 2.049996154141989e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 3587438.0,
      "step": 8234
    },
    {
      "epoch": 0.6334128143988924,
      "grad_norm": 53.0,
      "learning_rate": 2.0498807784016613e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3587778.0,
      "step": 8235
    },
    {
      "epoch": 0.6334897315591108,
      "grad_norm": 42.0,
      "learning_rate": 2.049765402661334e-06,
      "loss": 0.0451,
      "mean_token_accuracy": 0.9787985682487488,
      "num_tokens": 3588062.0,
      "step": 8236
    },
    {
      "epoch": 0.6335666487193293,
      "grad_norm": 24.75,
      "learning_rate": 2.0496500269210063e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 3588396.0,
      "step": 8237
    },
    {
      "epoch": 0.6336435658795477,
      "grad_norm": 6.8125,
      "learning_rate": 2.0495346511806785e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3588942.0,
      "step": 8238
    },
    {
      "epoch": 0.6337204830397661,
      "grad_norm": 10.8125,
      "learning_rate": 2.0494192754403508e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 3589518.0,
      "step": 8239
    },
    {
      "epoch": 0.6337974001999847,
      "grad_norm": 6.625,
      "learning_rate": 2.049303899700023e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 3590038.0,
      "step": 8240
    },
    {
      "epoch": 0.633874317360203,
      "grad_norm": 60.0,
      "learning_rate": 2.0491885239596953e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3590370.0,
      "step": 8241
    },
    {
      "epoch": 0.6339512345204215,
      "grad_norm": 11.125,
      "learning_rate": 2.049073148219368e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.998171865940094,
      "num_tokens": 3590918.0,
      "step": 8242
    },
    {
      "epoch": 0.63402815168064,
      "grad_norm": 5.1875,
      "learning_rate": 2.0489577724790402e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3591240.0,
      "step": 8243
    },
    {
      "epoch": 0.6341050688408584,
      "grad_norm": 14.0625,
      "learning_rate": 2.0488423967387125e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 3592036.0,
      "step": 8244
    },
    {
      "epoch": 0.6341819860010769,
      "grad_norm": 17.5,
      "learning_rate": 2.0487270209983848e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3592570.0,
      "step": 8245
    },
    {
      "epoch": 0.6342589031612953,
      "grad_norm": 52.75,
      "learning_rate": 2.048611645258057e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 3592872.0,
      "step": 8246
    },
    {
      "epoch": 0.6343358203215137,
      "grad_norm": 12.25,
      "learning_rate": 2.0484962695177297e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9948365092277527,
      "num_tokens": 3593454.0,
      "step": 8247
    },
    {
      "epoch": 0.6344127374817322,
      "grad_norm": 12.9375,
      "learning_rate": 2.048380893777402e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 3593788.0,
      "step": 8248
    },
    {
      "epoch": 0.6344896546419506,
      "grad_norm": 44.5,
      "learning_rate": 2.048265518037074e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3594128.0,
      "step": 8249
    },
    {
      "epoch": 0.634566571802169,
      "grad_norm": 24.75,
      "learning_rate": 2.0481501422967465e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 3594494.0,
      "step": 8250
    },
    {
      "epoch": 0.6346434889623875,
      "grad_norm": 33.5,
      "learning_rate": 2.0480347665564187e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3594774.0,
      "step": 8251
    },
    {
      "epoch": 0.6347204061226059,
      "grad_norm": 28.875,
      "learning_rate": 2.047919390816091e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 3595048.0,
      "step": 8252
    },
    {
      "epoch": 0.6347973232828243,
      "grad_norm": 21.75,
      "learning_rate": 2.0478040150757637e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3595592.0,
      "step": 8253
    },
    {
      "epoch": 0.6348742404430429,
      "grad_norm": 6.90625,
      "learning_rate": 2.047688639335436e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3595874.0,
      "step": 8254
    },
    {
      "epoch": 0.6349511576032613,
      "grad_norm": 27.375,
      "learning_rate": 2.047573263595108e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9948892593383789,
      "num_tokens": 3596462.0,
      "step": 8255
    },
    {
      "epoch": 0.6350280747634798,
      "grad_norm": 13.6875,
      "learning_rate": 2.0474578878547804e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 3596776.0,
      "step": 8256
    },
    {
      "epoch": 0.6351049919236982,
      "grad_norm": 57.75,
      "learning_rate": 2.0473425121144527e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3597084.0,
      "step": 8257
    },
    {
      "epoch": 0.6351819090839166,
      "grad_norm": 26.5,
      "learning_rate": 2.0472271363741254e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 3597398.0,
      "step": 8258
    },
    {
      "epoch": 0.6352588262441351,
      "grad_norm": 13.0,
      "learning_rate": 2.0471117606337976e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 3597970.0,
      "step": 8259
    },
    {
      "epoch": 0.6353357434043535,
      "grad_norm": 31.0,
      "learning_rate": 2.0469963848934695e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 3598318.0,
      "step": 8260
    },
    {
      "epoch": 0.6354126605645719,
      "grad_norm": 23.0,
      "learning_rate": 2.046881009153142e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3598630.0,
      "step": 8261
    },
    {
      "epoch": 0.6354895777247904,
      "grad_norm": 18.25,
      "learning_rate": 2.0467656334128144e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 3598980.0,
      "step": 8262
    },
    {
      "epoch": 0.6355664948850088,
      "grad_norm": 25.0,
      "learning_rate": 2.0466502576724867e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 3599302.0,
      "step": 8263
    },
    {
      "epoch": 0.6356434120452273,
      "grad_norm": 20.25,
      "learning_rate": 2.0465348819321594e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 3599612.0,
      "step": 8264
    },
    {
      "epoch": 0.6357203292054457,
      "grad_norm": 30.0,
      "learning_rate": 2.0464195061918316e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 3600192.0,
      "step": 8265
    },
    {
      "epoch": 0.6357972463656641,
      "grad_norm": 8.4375,
      "learning_rate": 2.046304130451504e-06,
      "loss": 0.0053,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 3601236.0,
      "step": 8266
    },
    {
      "epoch": 0.6358741635258827,
      "grad_norm": 33.25,
      "learning_rate": 2.046188754711176e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3601496.0,
      "step": 8267
    },
    {
      "epoch": 0.6359510806861011,
      "grad_norm": 39.25,
      "learning_rate": 2.0460733789708484e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3601808.0,
      "step": 8268
    },
    {
      "epoch": 0.6360279978463195,
      "grad_norm": 18.0,
      "learning_rate": 2.0459580032305206e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3602376.0,
      "step": 8269
    },
    {
      "epoch": 0.636104915006538,
      "grad_norm": 6.59375,
      "learning_rate": 2.0458426274901933e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3602744.0,
      "step": 8270
    },
    {
      "epoch": 0.6361818321667564,
      "grad_norm": 19.0,
      "learning_rate": 2.045727251749865e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 3603078.0,
      "step": 8271
    },
    {
      "epoch": 0.6362587493269748,
      "grad_norm": 22.625,
      "learning_rate": 2.045611876009538e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3603390.0,
      "step": 8272
    },
    {
      "epoch": 0.6363356664871933,
      "grad_norm": 7.40625,
      "learning_rate": 2.04549650026921e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 3603906.0,
      "step": 8273
    },
    {
      "epoch": 0.6364125836474117,
      "grad_norm": 23.875,
      "learning_rate": 2.0453811245288824e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3604270.0,
      "step": 8274
    },
    {
      "epoch": 0.6364895008076302,
      "grad_norm": 9.5,
      "learning_rate": 2.045265748788555e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3604820.0,
      "step": 8275
    },
    {
      "epoch": 0.6365664179678486,
      "grad_norm": 60.75,
      "learning_rate": 2.0451503730482273e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3605122.0,
      "step": 8276
    },
    {
      "epoch": 0.636643335128067,
      "grad_norm": 28.875,
      "learning_rate": 2.045034997307899e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3605438.0,
      "step": 8277
    },
    {
      "epoch": 0.6367202522882855,
      "grad_norm": 15.0,
      "learning_rate": 2.044919621567572e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3605788.0,
      "step": 8278
    },
    {
      "epoch": 0.636797169448504,
      "grad_norm": 35.75,
      "learning_rate": 2.044804245827244e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3606098.0,
      "step": 8279
    },
    {
      "epoch": 0.6368740866087224,
      "grad_norm": 27.25,
      "learning_rate": 2.0446888700869163e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3606656.0,
      "step": 8280
    },
    {
      "epoch": 0.6369510037689409,
      "grad_norm": 17.875,
      "learning_rate": 2.044573494346589e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.994413435459137,
      "num_tokens": 3607016.0,
      "step": 8281
    },
    {
      "epoch": 0.6370279209291593,
      "grad_norm": 10.0,
      "learning_rate": 2.044458118606261e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3607556.0,
      "step": 8282
    },
    {
      "epoch": 0.6371048380893778,
      "grad_norm": 33.0,
      "learning_rate": 2.0443427428659335e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3607862.0,
      "step": 8283
    },
    {
      "epoch": 0.6371817552495962,
      "grad_norm": 58.25,
      "learning_rate": 2.0442273671256058e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 3608200.0,
      "step": 8284
    },
    {
      "epoch": 0.6372586724098146,
      "grad_norm": 43.0,
      "learning_rate": 2.044111991385278e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3608762.0,
      "step": 8285
    },
    {
      "epoch": 0.6373355895700331,
      "grad_norm": 35.5,
      "learning_rate": 2.0439966156449507e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 3609582.0,
      "step": 8286
    },
    {
      "epoch": 0.6374125067302515,
      "grad_norm": 28.5,
      "learning_rate": 2.0438812399046226e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3610152.0,
      "step": 8287
    },
    {
      "epoch": 0.6374894238904699,
      "grad_norm": 9.9375,
      "learning_rate": 2.043765864164295e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3610448.0,
      "step": 8288
    },
    {
      "epoch": 0.6375663410506884,
      "grad_norm": 29.25,
      "learning_rate": 2.0436504884239675e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 3610724.0,
      "step": 8289
    },
    {
      "epoch": 0.6376432582109068,
      "grad_norm": 72.0,
      "learning_rate": 2.0435351126836398e-06,
      "loss": 0.0432,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3611030.0,
      "step": 8290
    },
    {
      "epoch": 0.6377201753711254,
      "grad_norm": 38.5,
      "learning_rate": 2.043419736943312e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 3611818.0,
      "step": 8291
    },
    {
      "epoch": 0.6377970925313438,
      "grad_norm": 30.75,
      "learning_rate": 2.0433043612029847e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3612158.0,
      "step": 8292
    },
    {
      "epoch": 0.6378740096915622,
      "grad_norm": 7.71875,
      "learning_rate": 2.0431889854626565e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 3612736.0,
      "step": 8293
    },
    {
      "epoch": 0.6379509268517807,
      "grad_norm": 15.8125,
      "learning_rate": 2.0430736097223292e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 3613288.0,
      "step": 8294
    },
    {
      "epoch": 0.6380278440119991,
      "grad_norm": 64.5,
      "learning_rate": 2.0429582339820015e-06,
      "loss": 0.0406,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 3613594.0,
      "step": 8295
    },
    {
      "epoch": 0.6381047611722175,
      "grad_norm": 5.71875,
      "learning_rate": 2.0428428582416737e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3614168.0,
      "step": 8296
    },
    {
      "epoch": 0.638181678332436,
      "grad_norm": 29.375,
      "learning_rate": 2.0427274825013464e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3614426.0,
      "step": 8297
    },
    {
      "epoch": 0.6382585954926544,
      "grad_norm": 13.4375,
      "learning_rate": 2.0426121067610182e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 3614766.0,
      "step": 8298
    },
    {
      "epoch": 0.6383355126528728,
      "grad_norm": 30.625,
      "learning_rate": 2.0424967310206905e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3615308.0,
      "step": 8299
    },
    {
      "epoch": 0.6384124298130913,
      "grad_norm": 84.0,
      "learning_rate": 2.042381355280363e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 3615910.0,
      "step": 8300
    },
    {
      "epoch": 0.6384893469733097,
      "grad_norm": 16.875,
      "learning_rate": 2.0422659795400354e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 3616462.0,
      "step": 8301
    },
    {
      "epoch": 0.6385662641335282,
      "grad_norm": 39.75,
      "learning_rate": 2.0421506037997077e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3617036.0,
      "step": 8302
    },
    {
      "epoch": 0.6386431812937466,
      "grad_norm": 16.75,
      "learning_rate": 2.0420352280593804e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3617570.0,
      "step": 8303
    },
    {
      "epoch": 0.638720098453965,
      "grad_norm": 31.875,
      "learning_rate": 2.0419198523190522e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 3617910.0,
      "step": 8304
    },
    {
      "epoch": 0.6387970156141836,
      "grad_norm": 13.4375,
      "learning_rate": 2.041804476578725e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3618454.0,
      "step": 8305
    },
    {
      "epoch": 0.638873932774402,
      "grad_norm": 35.75,
      "learning_rate": 2.041689100838397e-06,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 3618708.0,
      "step": 8306
    },
    {
      "epoch": 0.6389508499346204,
      "grad_norm": 3.953125,
      "learning_rate": 2.0415737250980694e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3619272.0,
      "step": 8307
    },
    {
      "epoch": 0.6390277670948389,
      "grad_norm": 88.5,
      "learning_rate": 2.041458349357742e-06,
      "loss": 0.0567,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 3619576.0,
      "step": 8308
    },
    {
      "epoch": 0.6391046842550573,
      "grad_norm": 9.0,
      "learning_rate": 2.041342973617414e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3620100.0,
      "step": 8309
    },
    {
      "epoch": 0.6391816014152758,
      "grad_norm": 4.46875,
      "learning_rate": 2.041227597877086e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3620686.0,
      "step": 8310
    },
    {
      "epoch": 0.6392585185754942,
      "grad_norm": 53.0,
      "learning_rate": 2.041112222136759e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3620992.0,
      "step": 8311
    },
    {
      "epoch": 0.6393354357357126,
      "grad_norm": 42.0,
      "learning_rate": 2.040996846396431e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3621274.0,
      "step": 8312
    },
    {
      "epoch": 0.6394123528959311,
      "grad_norm": 19.5,
      "learning_rate": 2.0408814706561034e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3621582.0,
      "step": 8313
    },
    {
      "epoch": 0.6394892700561495,
      "grad_norm": 32.5,
      "learning_rate": 2.040766094915776e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 3621878.0,
      "step": 8314
    },
    {
      "epoch": 0.6395661872163679,
      "grad_norm": 59.0,
      "learning_rate": 2.040650719175448e-06,
      "loss": 0.0448,
      "mean_token_accuracy": 0.9786477088928223,
      "num_tokens": 3622160.0,
      "step": 8315
    },
    {
      "epoch": 0.6396431043765864,
      "grad_norm": 14.75,
      "learning_rate": 2.0405353434351206e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 3622520.0,
      "step": 8316
    },
    {
      "epoch": 0.6397200215368049,
      "grad_norm": 7.4375,
      "learning_rate": 2.040419967694793e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9967741966247559,
      "num_tokens": 3622832.0,
      "step": 8317
    },
    {
      "epoch": 0.6397969386970233,
      "grad_norm": 29.875,
      "learning_rate": 2.040304591954465e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 3623380.0,
      "step": 8318
    },
    {
      "epoch": 0.6398738558572418,
      "grad_norm": 10.9375,
      "learning_rate": 2.0401892162141374e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3623928.0,
      "step": 8319
    },
    {
      "epoch": 0.6399507730174602,
      "grad_norm": 45.25,
      "learning_rate": 2.0400738404738096e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 3624228.0,
      "step": 8320
    },
    {
      "epoch": 0.6400276901776787,
      "grad_norm": 24.625,
      "learning_rate": 2.039958464733482e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 3624806.0,
      "step": 8321
    },
    {
      "epoch": 0.6401046073378971,
      "grad_norm": 8.375,
      "learning_rate": 2.0398430889931546e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9963833689689636,
      "num_tokens": 3625360.0,
      "step": 8322
    },
    {
      "epoch": 0.6401815244981155,
      "grad_norm": 77.5,
      "learning_rate": 2.039727713252827e-06,
      "loss": 0.0353,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3625616.0,
      "step": 8323
    },
    {
      "epoch": 0.640258441658334,
      "grad_norm": 22.5,
      "learning_rate": 2.039612337512499e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9966044425964355,
      "num_tokens": 3626206.0,
      "step": 8324
    },
    {
      "epoch": 0.6403353588185524,
      "grad_norm": 2.78125,
      "learning_rate": 2.0394969617721713e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3626792.0,
      "step": 8325
    },
    {
      "epoch": 0.6404122759787708,
      "grad_norm": 15.4375,
      "learning_rate": 2.0393815860318436e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3627288.0,
      "step": 8326
    },
    {
      "epoch": 0.6404891931389893,
      "grad_norm": 20.625,
      "learning_rate": 2.039266210291516e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3627582.0,
      "step": 8327
    },
    {
      "epoch": 0.6405661102992077,
      "grad_norm": 38.0,
      "learning_rate": 2.0391508345511885e-06,
      "loss": 0.0469,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3627860.0,
      "step": 8328
    },
    {
      "epoch": 0.6406430274594263,
      "grad_norm": 53.75,
      "learning_rate": 2.039035458810861e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3628172.0,
      "step": 8329
    },
    {
      "epoch": 0.6407199446196447,
      "grad_norm": 41.75,
      "learning_rate": 2.038920083070533e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3628514.0,
      "step": 8330
    },
    {
      "epoch": 0.6407968617798631,
      "grad_norm": 28.5,
      "learning_rate": 2.0388047073302053e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 3628852.0,
      "step": 8331
    },
    {
      "epoch": 0.6408737789400816,
      "grad_norm": 17.0,
      "learning_rate": 2.0386893315898776e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 3629190.0,
      "step": 8332
    },
    {
      "epoch": 0.6409506961003,
      "grad_norm": 26.0,
      "learning_rate": 2.0385739558495502e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3629472.0,
      "step": 8333
    },
    {
      "epoch": 0.6410276132605184,
      "grad_norm": 16.375,
      "learning_rate": 2.0384585801092225e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3630022.0,
      "step": 8334
    },
    {
      "epoch": 0.6411045304207369,
      "grad_norm": 5.25,
      "learning_rate": 2.0383432043688948e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3630804.0,
      "step": 8335
    },
    {
      "epoch": 0.6411814475809553,
      "grad_norm": 27.125,
      "learning_rate": 2.038227828628567e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3631376.0,
      "step": 8336
    },
    {
      "epoch": 0.6412583647411737,
      "grad_norm": 14.9375,
      "learning_rate": 2.0381124528882393e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3631960.0,
      "step": 8337
    },
    {
      "epoch": 0.6413352819013922,
      "grad_norm": 18.25,
      "learning_rate": 2.0379970771479115e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3632240.0,
      "step": 8338
    },
    {
      "epoch": 0.6414121990616106,
      "grad_norm": 13.625,
      "learning_rate": 2.0378817014075842e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3632562.0,
      "step": 8339
    },
    {
      "epoch": 0.6414891162218291,
      "grad_norm": 9.1875,
      "learning_rate": 2.0377663256672565e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 3633094.0,
      "step": 8340
    },
    {
      "epoch": 0.6415660333820475,
      "grad_norm": 7.09375,
      "learning_rate": 2.0376509499269287e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 3633898.0,
      "step": 8341
    },
    {
      "epoch": 0.6416429505422659,
      "grad_norm": 6.78125,
      "learning_rate": 2.037535574186601e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3634218.0,
      "step": 8342
    },
    {
      "epoch": 0.6417198677024845,
      "grad_norm": 29.375,
      "learning_rate": 2.0374201984462733e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3634522.0,
      "step": 8343
    },
    {
      "epoch": 0.6417967848627029,
      "grad_norm": 11.125,
      "learning_rate": 2.037304822705946e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3635118.0,
      "step": 8344
    },
    {
      "epoch": 0.6418737020229213,
      "grad_norm": 11.25,
      "learning_rate": 2.037189446965618e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3635668.0,
      "step": 8345
    },
    {
      "epoch": 0.6419506191831398,
      "grad_norm": 8.125,
      "learning_rate": 2.0370740712252904e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3636018.0,
      "step": 8346
    },
    {
      "epoch": 0.6420275363433582,
      "grad_norm": 10.6875,
      "learning_rate": 2.0369586954849627e-06,
      "loss": 0.0041,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3636854.0,
      "step": 8347
    },
    {
      "epoch": 0.6421044535035767,
      "grad_norm": 97.5,
      "learning_rate": 2.036843319744635e-06,
      "loss": 0.049,
      "mean_token_accuracy": 0.9785932898521423,
      "num_tokens": 3637182.0,
      "step": 8348
    },
    {
      "epoch": 0.6421813706637951,
      "grad_norm": 63.75,
      "learning_rate": 2.0367279440043072e-06,
      "loss": 0.0523,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3637464.0,
      "step": 8349
    },
    {
      "epoch": 0.6422582878240135,
      "grad_norm": 7.4375,
      "learning_rate": 2.03661256826398e-06,
      "loss": 0.0074,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 3638262.0,
      "step": 8350
    },
    {
      "epoch": 0.642335204984232,
      "grad_norm": 36.75,
      "learning_rate": 2.036497192523652e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3638592.0,
      "step": 8351
    },
    {
      "epoch": 0.6424121221444504,
      "grad_norm": 36.75,
      "learning_rate": 2.0363818167833244e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3638926.0,
      "step": 8352
    },
    {
      "epoch": 0.6424890393046688,
      "grad_norm": 39.75,
      "learning_rate": 2.0362664410429967e-06,
      "loss": 0.0473,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 3639188.0,
      "step": 8353
    },
    {
      "epoch": 0.6425659564648873,
      "grad_norm": 36.0,
      "learning_rate": 2.036151065302669e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 3639544.0,
      "step": 8354
    },
    {
      "epoch": 0.6426428736251057,
      "grad_norm": 45.25,
      "learning_rate": 2.0360356895623416e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3640084.0,
      "step": 8355
    },
    {
      "epoch": 0.6427197907853242,
      "grad_norm": 69.0,
      "learning_rate": 2.035920313822014e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 3640390.0,
      "step": 8356
    },
    {
      "epoch": 0.6427967079455427,
      "grad_norm": 14.125,
      "learning_rate": 2.035804938081686e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3640674.0,
      "step": 8357
    },
    {
      "epoch": 0.6428736251057611,
      "grad_norm": 5.90625,
      "learning_rate": 2.0356895623413584e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3641244.0,
      "step": 8358
    },
    {
      "epoch": 0.6429505422659796,
      "grad_norm": 13.0625,
      "learning_rate": 2.0355741866010307e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 3641564.0,
      "step": 8359
    },
    {
      "epoch": 0.643027459426198,
      "grad_norm": 10.3125,
      "learning_rate": 2.035458810860703e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3642154.0,
      "step": 8360
    },
    {
      "epoch": 0.6431043765864164,
      "grad_norm": 23.5,
      "learning_rate": 2.0353434351203756e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 3642502.0,
      "step": 8361
    },
    {
      "epoch": 0.6431812937466349,
      "grad_norm": 31.875,
      "learning_rate": 2.035228059380048e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3642810.0,
      "step": 8362
    },
    {
      "epoch": 0.6432582109068533,
      "grad_norm": 14.9375,
      "learning_rate": 2.03511268363972e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 3643100.0,
      "step": 8363
    },
    {
      "epoch": 0.6433351280670717,
      "grad_norm": 4.09375,
      "learning_rate": 2.0349973078993924e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3643670.0,
      "step": 8364
    },
    {
      "epoch": 0.6434120452272902,
      "grad_norm": 7.0,
      "learning_rate": 2.0348819321590646e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3644256.0,
      "step": 8365
    },
    {
      "epoch": 0.6434889623875086,
      "grad_norm": 6.59375,
      "learning_rate": 2.0347665564187373e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 3644836.0,
      "step": 8366
    },
    {
      "epoch": 0.6435658795477271,
      "grad_norm": 6.46875,
      "learning_rate": 2.0346511806784096e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 3645124.0,
      "step": 8367
    },
    {
      "epoch": 0.6436427967079456,
      "grad_norm": 15.125,
      "learning_rate": 2.034535804938082e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3645408.0,
      "step": 8368
    },
    {
      "epoch": 0.643719713868164,
      "grad_norm": 5.40625,
      "learning_rate": 2.034420429197754e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3645930.0,
      "step": 8369
    },
    {
      "epoch": 0.6437966310283825,
      "grad_norm": 7.5,
      "learning_rate": 2.0343050534574263e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3646228.0,
      "step": 8370
    },
    {
      "epoch": 0.6438735481886009,
      "grad_norm": 10.5625,
      "learning_rate": 2.0341896777170986e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3646548.0,
      "step": 8371
    },
    {
      "epoch": 0.6439504653488193,
      "grad_norm": 10.75,
      "learning_rate": 2.0340743019767713e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3646894.0,
      "step": 8372
    },
    {
      "epoch": 0.6440273825090378,
      "grad_norm": 8.375,
      "learning_rate": 2.0339589262364435e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3647216.0,
      "step": 8373
    },
    {
      "epoch": 0.6441042996692562,
      "grad_norm": 11.5,
      "learning_rate": 2.033843550496116e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3647534.0,
      "step": 8374
    },
    {
      "epoch": 0.6441812168294746,
      "grad_norm": 5.65625,
      "learning_rate": 2.033728174755788e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3648068.0,
      "step": 8375
    },
    {
      "epoch": 0.6442581339896931,
      "grad_norm": 14.9375,
      "learning_rate": 2.0336127990154603e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3648396.0,
      "step": 8376
    },
    {
      "epoch": 0.6443350511499115,
      "grad_norm": 8.625,
      "learning_rate": 2.0334974232751326e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3648712.0,
      "step": 8377
    },
    {
      "epoch": 0.64441196831013,
      "grad_norm": 20.25,
      "learning_rate": 2.0333820475348052e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3649080.0,
      "step": 8378
    },
    {
      "epoch": 0.6444888854703484,
      "grad_norm": 8.0625,
      "learning_rate": 2.0332666717944775e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 3649372.0,
      "step": 8379
    },
    {
      "epoch": 0.6445658026305668,
      "grad_norm": 12.1875,
      "learning_rate": 2.0331512960541498e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3649956.0,
      "step": 8380
    },
    {
      "epoch": 0.6446427197907854,
      "grad_norm": 11.625,
      "learning_rate": 2.033035920313822e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 3650764.0,
      "step": 8381
    },
    {
      "epoch": 0.6447196369510038,
      "grad_norm": 26.625,
      "learning_rate": 2.0329205445734943e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3651086.0,
      "step": 8382
    },
    {
      "epoch": 0.6447965541112222,
      "grad_norm": 29.875,
      "learning_rate": 2.032805168833167e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3651402.0,
      "step": 8383
    },
    {
      "epoch": 0.6448734712714407,
      "grad_norm": 3.578125,
      "learning_rate": 2.0326897930928392e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3651972.0,
      "step": 8384
    },
    {
      "epoch": 0.6449503884316591,
      "grad_norm": 33.75,
      "learning_rate": 2.032574417352511e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9858356714248657,
      "num_tokens": 3652326.0,
      "step": 8385
    },
    {
      "epoch": 0.6450273055918776,
      "grad_norm": 8.5,
      "learning_rate": 2.0324590416121837e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 3652910.0,
      "step": 8386
    },
    {
      "epoch": 0.645104222752096,
      "grad_norm": 5.59375,
      "learning_rate": 2.032343665871856e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 3653486.0,
      "step": 8387
    },
    {
      "epoch": 0.6451811399123144,
      "grad_norm": 44.5,
      "learning_rate": 2.0322282901315283e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3653810.0,
      "step": 8388
    },
    {
      "epoch": 0.6452580570725329,
      "grad_norm": 29.875,
      "learning_rate": 2.032112914391201e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3654066.0,
      "step": 8389
    },
    {
      "epoch": 0.6453349742327513,
      "grad_norm": 4.84375,
      "learning_rate": 2.0319975386508728e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3654584.0,
      "step": 8390
    },
    {
      "epoch": 0.6454118913929697,
      "grad_norm": 8.8125,
      "learning_rate": 2.0318821629105455e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.994413435459137,
      "num_tokens": 3654944.0,
      "step": 8391
    },
    {
      "epoch": 0.6454888085531882,
      "grad_norm": 25.375,
      "learning_rate": 2.0317667871702177e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 3655284.0,
      "step": 8392
    },
    {
      "epoch": 0.6455657257134066,
      "grad_norm": 43.75,
      "learning_rate": 2.03165141142989e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3655594.0,
      "step": 8393
    },
    {
      "epoch": 0.6456426428736252,
      "grad_norm": 26.875,
      "learning_rate": 2.0315360356895626e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3655900.0,
      "step": 8394
    },
    {
      "epoch": 0.6457195600338436,
      "grad_norm": 8.8125,
      "learning_rate": 2.031420659949235e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3656228.0,
      "step": 8395
    },
    {
      "epoch": 0.645796477194062,
      "grad_norm": 3.828125,
      "learning_rate": 2.0313052842089067e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3656796.0,
      "step": 8396
    },
    {
      "epoch": 0.6458733943542805,
      "grad_norm": 7.9375,
      "learning_rate": 2.0311899084685794e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3657382.0,
      "step": 8397
    },
    {
      "epoch": 0.6459503115144989,
      "grad_norm": 6.8125,
      "learning_rate": 2.0310745327282517e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 3657700.0,
      "step": 8398
    },
    {
      "epoch": 0.6460272286747173,
      "grad_norm": 4.21875,
      "learning_rate": 2.030959156987924e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3658270.0,
      "step": 8399
    },
    {
      "epoch": 0.6461041458349358,
      "grad_norm": 8.0,
      "learning_rate": 2.0308437812475966e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 3658846.0,
      "step": 8400
    },
    {
      "epoch": 0.6461810629951542,
      "grad_norm": 11.25,
      "learning_rate": 2.0307284055072685e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 3659190.0,
      "step": 8401
    },
    {
      "epoch": 0.6462579801553726,
      "grad_norm": 4.125,
      "learning_rate": 2.030613029766941e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3659708.0,
      "step": 8402
    },
    {
      "epoch": 0.6463348973155911,
      "grad_norm": 4.09375,
      "learning_rate": 2.0304976540266134e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3660248.0,
      "step": 8403
    },
    {
      "epoch": 0.6464118144758095,
      "grad_norm": 3.171875,
      "learning_rate": 2.0303822782862857e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3660804.0,
      "step": 8404
    },
    {
      "epoch": 0.646488731636028,
      "grad_norm": 3.4375,
      "learning_rate": 2.0302669025459583e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3661388.0,
      "step": 8405
    },
    {
      "epoch": 0.6465656487962465,
      "grad_norm": 2.96875,
      "learning_rate": 2.0301515268056306e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3661936.0,
      "step": 8406
    },
    {
      "epoch": 0.6466425659564649,
      "grad_norm": 12.5,
      "learning_rate": 2.0300361510653024e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9839743375778198,
      "num_tokens": 3662250.0,
      "step": 8407
    },
    {
      "epoch": 0.6467194831166834,
      "grad_norm": 3.171875,
      "learning_rate": 2.029920775324975e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3663008.0,
      "step": 8408
    },
    {
      "epoch": 0.6467964002769018,
      "grad_norm": 7.3125,
      "learning_rate": 2.0298053995846474e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3663548.0,
      "step": 8409
    },
    {
      "epoch": 0.6468733174371202,
      "grad_norm": 3.625,
      "learning_rate": 2.0296900238443196e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 3664124.0,
      "step": 8410
    },
    {
      "epoch": 0.6469502345973387,
      "grad_norm": 6.5625,
      "learning_rate": 2.0295746481039923e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3664642.0,
      "step": 8411
    },
    {
      "epoch": 0.6470271517575571,
      "grad_norm": 6.65625,
      "learning_rate": 2.029459272363664e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3665210.0,
      "step": 8412
    },
    {
      "epoch": 0.6471040689177756,
      "grad_norm": 22.875,
      "learning_rate": 2.029343896623337e-06,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 3665514.0,
      "step": 8413
    },
    {
      "epoch": 0.647180986077994,
      "grad_norm": 11.6875,
      "learning_rate": 2.029228520883009e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3665860.0,
      "step": 8414
    },
    {
      "epoch": 0.6472579032382124,
      "grad_norm": 8.75,
      "learning_rate": 2.0291131451426813e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3666170.0,
      "step": 8415
    },
    {
      "epoch": 0.6473348203984309,
      "grad_norm": 4.65625,
      "learning_rate": 2.028997769402354e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 3666742.0,
      "step": 8416
    },
    {
      "epoch": 0.6474117375586493,
      "grad_norm": 11.375,
      "learning_rate": 2.0288823936620263e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3667058.0,
      "step": 8417
    },
    {
      "epoch": 0.6474886547188677,
      "grad_norm": 9.4375,
      "learning_rate": 2.028767017921698e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3667352.0,
      "step": 8418
    },
    {
      "epoch": 0.6475655718790863,
      "grad_norm": 4.21875,
      "learning_rate": 2.028651642181371e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3667904.0,
      "step": 8419
    },
    {
      "epoch": 0.6476424890393047,
      "grad_norm": 9.0625,
      "learning_rate": 2.028536266441043e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3668164.0,
      "step": 8420
    },
    {
      "epoch": 0.6477194061995231,
      "grad_norm": 4.53125,
      "learning_rate": 2.0284208907007153e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3668704.0,
      "step": 8421
    },
    {
      "epoch": 0.6477963233597416,
      "grad_norm": 11.75,
      "learning_rate": 2.028305514960388e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3669028.0,
      "step": 8422
    },
    {
      "epoch": 0.64787324051996,
      "grad_norm": 6.03125,
      "learning_rate": 2.02819013922006e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3669592.0,
      "step": 8423
    },
    {
      "epoch": 0.6479501576801785,
      "grad_norm": 9.5625,
      "learning_rate": 2.0280747634797325e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3670126.0,
      "step": 8424
    },
    {
      "epoch": 0.6480270748403969,
      "grad_norm": 5.84375,
      "learning_rate": 2.0279593877394048e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3670666.0,
      "step": 8425
    },
    {
      "epoch": 0.6481039920006153,
      "grad_norm": 4.9375,
      "learning_rate": 2.027844011999077e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 3671236.0,
      "step": 8426
    },
    {
      "epoch": 0.6481809091608338,
      "grad_norm": 20.0,
      "learning_rate": 2.0277286362587493e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 3671550.0,
      "step": 8427
    },
    {
      "epoch": 0.6482578263210522,
      "grad_norm": 20.125,
      "learning_rate": 2.0276132605184215e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9849397540092468,
      "num_tokens": 3671884.0,
      "step": 8428
    },
    {
      "epoch": 0.6483347434812706,
      "grad_norm": 30.5,
      "learning_rate": 2.027497884778094e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3672206.0,
      "step": 8429
    },
    {
      "epoch": 0.6484116606414891,
      "grad_norm": 20.25,
      "learning_rate": 2.0273825090377665e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 3672556.0,
      "step": 8430
    },
    {
      "epoch": 0.6484885778017075,
      "grad_norm": 13.3125,
      "learning_rate": 2.0272671332974387e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3672876.0,
      "step": 8431
    },
    {
      "epoch": 0.6485654949619261,
      "grad_norm": 5.53125,
      "learning_rate": 2.027151757557111e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3673456.0,
      "step": 8432
    },
    {
      "epoch": 0.6486424121221445,
      "grad_norm": 14.3125,
      "learning_rate": 2.0270363818167837e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3674000.0,
      "step": 8433
    },
    {
      "epoch": 0.6487193292823629,
      "grad_norm": 49.75,
      "learning_rate": 2.0269210060764555e-06,
      "loss": 0.0522,
      "mean_token_accuracy": 0.9788732528686523,
      "num_tokens": 3674286.0,
      "step": 8434
    },
    {
      "epoch": 0.6487962464425814,
      "grad_norm": 10.5625,
      "learning_rate": 2.0268056303361278e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3674842.0,
      "step": 8435
    },
    {
      "epoch": 0.6488731636027998,
      "grad_norm": 30.125,
      "learning_rate": 2.0266902545958005e-06,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 3675122.0,
      "step": 8436
    },
    {
      "epoch": 0.6489500807630182,
      "grad_norm": 10.1875,
      "learning_rate": 2.0265748788554727e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3675680.0,
      "step": 8437
    },
    {
      "epoch": 0.6490269979232367,
      "grad_norm": 12.0625,
      "learning_rate": 2.026459503115145e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3676264.0,
      "step": 8438
    },
    {
      "epoch": 0.6491039150834551,
      "grad_norm": 6.0,
      "learning_rate": 2.0263441273748172e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3676782.0,
      "step": 8439
    },
    {
      "epoch": 0.6491808322436735,
      "grad_norm": 7.5625,
      "learning_rate": 2.0262287516344895e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3677068.0,
      "step": 8440
    },
    {
      "epoch": 0.649257749403892,
      "grad_norm": 8.3125,
      "learning_rate": 2.026113375894162e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3677590.0,
      "step": 8441
    },
    {
      "epoch": 0.6493346665641104,
      "grad_norm": 4.28125,
      "learning_rate": 2.0259980001538344e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3678110.0,
      "step": 8442
    },
    {
      "epoch": 0.649411583724329,
      "grad_norm": 7.21875,
      "learning_rate": 2.0258826244135067e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3678438.0,
      "step": 8443
    },
    {
      "epoch": 0.6494885008845473,
      "grad_norm": 7.625,
      "learning_rate": 2.0257672486731794e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9981308579444885,
      "num_tokens": 3678974.0,
      "step": 8444
    },
    {
      "epoch": 0.6495654180447658,
      "grad_norm": 8.9375,
      "learning_rate": 2.025651872932851e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 3679260.0,
      "step": 8445
    },
    {
      "epoch": 0.6496423352049843,
      "grad_norm": 22.75,
      "learning_rate": 2.0255364971925235e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 3679608.0,
      "step": 8446
    },
    {
      "epoch": 0.6497192523652027,
      "grad_norm": 13.1875,
      "learning_rate": 2.025421121452196e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3679938.0,
      "step": 8447
    },
    {
      "epoch": 0.6497961695254211,
      "grad_norm": 10.3125,
      "learning_rate": 2.0253057457118684e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 3680488.0,
      "step": 8448
    },
    {
      "epoch": 0.6498730866856396,
      "grad_norm": 5.4375,
      "learning_rate": 2.0251903699715407e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3680832.0,
      "step": 8449
    },
    {
      "epoch": 0.649950003845858,
      "grad_norm": 13.6875,
      "learning_rate": 2.025074994231213e-06,
      "loss": 0.0063,
      "mean_token_accuracy": 0.9970559477806091,
      "num_tokens": 3681854.0,
      "step": 8450
    },
    {
      "epoch": 0.6500269210060765,
      "grad_norm": 21.125,
      "learning_rate": 2.024959618490885e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 3682198.0,
      "step": 8451
    },
    {
      "epoch": 0.6501038381662949,
      "grad_norm": 8.375,
      "learning_rate": 2.024844242750558e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3682506.0,
      "step": 8452
    },
    {
      "epoch": 0.6501807553265133,
      "grad_norm": 11.6875,
      "learning_rate": 2.02472886701023e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.98525071144104,
      "num_tokens": 3682846.0,
      "step": 8453
    },
    {
      "epoch": 0.6502576724867318,
      "grad_norm": 6.40625,
      "learning_rate": 2.0246134912699024e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3683392.0,
      "step": 8454
    },
    {
      "epoch": 0.6503345896469502,
      "grad_norm": 1.9453125,
      "learning_rate": 2.0244981155295746e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 3683950.0,
      "step": 8455
    },
    {
      "epoch": 0.6504115068071686,
      "grad_norm": 5.34375,
      "learning_rate": 2.024382739789247e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 3684754.0,
      "step": 8456
    },
    {
      "epoch": 0.6504884239673872,
      "grad_norm": 15.1875,
      "learning_rate": 2.024267364048919e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 3685110.0,
      "step": 8457
    },
    {
      "epoch": 0.6505653411276056,
      "grad_norm": 10.0,
      "learning_rate": 2.024151988308592e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3685438.0,
      "step": 8458
    },
    {
      "epoch": 0.650642258287824,
      "grad_norm": 26.375,
      "learning_rate": 2.024036612568264e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3685758.0,
      "step": 8459
    },
    {
      "epoch": 0.6507191754480425,
      "grad_norm": 27.375,
      "learning_rate": 2.0239212368279363e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3686052.0,
      "step": 8460
    },
    {
      "epoch": 0.6507960926082609,
      "grad_norm": 2.21875,
      "learning_rate": 2.0238058610876086e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3686616.0,
      "step": 8461
    },
    {
      "epoch": 0.6508730097684794,
      "grad_norm": 18.75,
      "learning_rate": 2.023690485347281e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 3686922.0,
      "step": 8462
    },
    {
      "epoch": 0.6509499269286978,
      "grad_norm": 7.75,
      "learning_rate": 2.0235751096069535e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 3687278.0,
      "step": 8463
    },
    {
      "epoch": 0.6510268440889162,
      "grad_norm": 4.1875,
      "learning_rate": 2.023459733866626e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3687840.0,
      "step": 8464
    },
    {
      "epoch": 0.6511037612491347,
      "grad_norm": 33.25,
      "learning_rate": 2.023344358126298e-06,
      "loss": 0.0393,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 3688154.0,
      "step": 8465
    },
    {
      "epoch": 0.6511806784093531,
      "grad_norm": 29.5,
      "learning_rate": 2.0232289823859703e-06,
      "loss": 0.048,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3688410.0,
      "step": 8466
    },
    {
      "epoch": 0.6512575955695715,
      "grad_norm": 6.15625,
      "learning_rate": 2.0231136066456426e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3688974.0,
      "step": 8467
    },
    {
      "epoch": 0.65133451272979,
      "grad_norm": 48.5,
      "learning_rate": 2.022998230905315e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3689270.0,
      "step": 8468
    },
    {
      "epoch": 0.6514114298900084,
      "grad_norm": 14.6875,
      "learning_rate": 2.0228828551649875e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 3689626.0,
      "step": 8469
    },
    {
      "epoch": 0.651488347050227,
      "grad_norm": 8.75,
      "learning_rate": 2.0227674794246598e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 3689972.0,
      "step": 8470
    },
    {
      "epoch": 0.6515652642104454,
      "grad_norm": 16.375,
      "learning_rate": 2.022652103684332e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3690294.0,
      "step": 8471
    },
    {
      "epoch": 0.6516421813706638,
      "grad_norm": 23.875,
      "learning_rate": 2.0225367279440043e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3690874.0,
      "step": 8472
    },
    {
      "epoch": 0.6517190985308823,
      "grad_norm": 10.6875,
      "learning_rate": 2.0224213522036765e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3691196.0,
      "step": 8473
    },
    {
      "epoch": 0.6517960156911007,
      "grad_norm": 12.5625,
      "learning_rate": 2.0223059764633492e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3691518.0,
      "step": 8474
    },
    {
      "epoch": 0.6518729328513191,
      "grad_norm": 16.125,
      "learning_rate": 2.0221906007230215e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3692072.0,
      "step": 8475
    },
    {
      "epoch": 0.6519498500115376,
      "grad_norm": 14.1875,
      "learning_rate": 2.0220752249826937e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 3692388.0,
      "step": 8476
    },
    {
      "epoch": 0.652026767171756,
      "grad_norm": 5.40625,
      "learning_rate": 2.021959849242366e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 3692920.0,
      "step": 8477
    },
    {
      "epoch": 0.6521036843319744,
      "grad_norm": 11.75,
      "learning_rate": 2.0218444735020383e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3693468.0,
      "step": 8478
    },
    {
      "epoch": 0.6521806014921929,
      "grad_norm": 21.25,
      "learning_rate": 2.0217290977617105e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3693764.0,
      "step": 8479
    },
    {
      "epoch": 0.6522575186524113,
      "grad_norm": 13.8125,
      "learning_rate": 2.021613722021383e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3694308.0,
      "step": 8480
    },
    {
      "epoch": 0.6523344358126298,
      "grad_norm": 5.21875,
      "learning_rate": 2.0214983462810555e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 3694654.0,
      "step": 8481
    },
    {
      "epoch": 0.6524113529728482,
      "grad_norm": 64.5,
      "learning_rate": 2.0213829705407277e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3694914.0,
      "step": 8482
    },
    {
      "epoch": 0.6524882701330667,
      "grad_norm": 34.5,
      "learning_rate": 2.0212675948004e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3695198.0,
      "step": 8483
    },
    {
      "epoch": 0.6525651872932852,
      "grad_norm": 11.3125,
      "learning_rate": 2.0211522190600722e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3695766.0,
      "step": 8484
    },
    {
      "epoch": 0.6526421044535036,
      "grad_norm": 15.625,
      "learning_rate": 2.0210368433197445e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3696314.0,
      "step": 8485
    },
    {
      "epoch": 0.652719021613722,
      "grad_norm": 22.875,
      "learning_rate": 2.020921467579417e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3696644.0,
      "step": 8486
    },
    {
      "epoch": 0.6527959387739405,
      "grad_norm": 10.625,
      "learning_rate": 2.0208060918390894e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.993127167224884,
      "num_tokens": 3696936.0,
      "step": 8487
    },
    {
      "epoch": 0.6528728559341589,
      "grad_norm": 12.875,
      "learning_rate": 2.0206907160987617e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3697484.0,
      "step": 8488
    },
    {
      "epoch": 0.6529497730943774,
      "grad_norm": 11.5625,
      "learning_rate": 2.020575340358434e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9889298677444458,
      "num_tokens": 3697756.0,
      "step": 8489
    },
    {
      "epoch": 0.6530266902545958,
      "grad_norm": 7.75,
      "learning_rate": 2.020459964618106e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 3698320.0,
      "step": 8490
    },
    {
      "epoch": 0.6531036074148142,
      "grad_norm": 5.15625,
      "learning_rate": 2.020344588877779e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3698904.0,
      "step": 8491
    },
    {
      "epoch": 0.6531805245750327,
      "grad_norm": 18.625,
      "learning_rate": 2.020229213137451e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 3699210.0,
      "step": 8492
    },
    {
      "epoch": 0.6532574417352511,
      "grad_norm": 14.6875,
      "learning_rate": 2.020113837397123e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 3699474.0,
      "step": 8493
    },
    {
      "epoch": 0.6533343588954695,
      "grad_norm": 6.09375,
      "learning_rate": 2.0199984616567957e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3700044.0,
      "step": 8494
    },
    {
      "epoch": 0.653411276055688,
      "grad_norm": 31.25,
      "learning_rate": 2.019883085916468e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3700328.0,
      "step": 8495
    },
    {
      "epoch": 0.6534881932159065,
      "grad_norm": 3.53125,
      "learning_rate": 2.01976771017614e-06,
      "loss": 0.0068,
      "mean_token_accuracy": 0.9975669384002686,
      "num_tokens": 3701152.0,
      "step": 8496
    },
    {
      "epoch": 0.6535651103761249,
      "grad_norm": 20.125,
      "learning_rate": 2.019652334435813e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9971014261245728,
      "num_tokens": 3701500.0,
      "step": 8497
    },
    {
      "epoch": 0.6536420275363434,
      "grad_norm": 51.0,
      "learning_rate": 2.019536958695485e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 3701792.0,
      "step": 8498
    },
    {
      "epoch": 0.6537189446965618,
      "grad_norm": 39.75,
      "learning_rate": 2.0194215829551574e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 3702106.0,
      "step": 8499
    },
    {
      "epoch": 0.6537958618567803,
      "grad_norm": 20.375,
      "learning_rate": 2.0193062072148296e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 3702406.0,
      "step": 8500
    },
    {
      "epoch": 0.6538727790169987,
      "grad_norm": 3.453125,
      "learning_rate": 2.019190831474502e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 3703214.0,
      "step": 8501
    },
    {
      "epoch": 0.6539496961772171,
      "grad_norm": 18.125,
      "learning_rate": 2.0190754557341746e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3703526.0,
      "step": 8502
    },
    {
      "epoch": 0.6540266133374356,
      "grad_norm": 46.25,
      "learning_rate": 2.018960079993847e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3703858.0,
      "step": 8503
    },
    {
      "epoch": 0.654103530497654,
      "grad_norm": 7.34375,
      "learning_rate": 2.0188447042535187e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3704164.0,
      "step": 8504
    },
    {
      "epoch": 0.6541804476578724,
      "grad_norm": 21.25,
      "learning_rate": 2.0187293285131913e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3704504.0,
      "step": 8505
    },
    {
      "epoch": 0.6542573648180909,
      "grad_norm": 6.34375,
      "learning_rate": 2.0186139527728636e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 3705308.0,
      "step": 8506
    },
    {
      "epoch": 0.6543342819783093,
      "grad_norm": 10.4375,
      "learning_rate": 2.018498577032536e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 3705886.0,
      "step": 8507
    },
    {
      "epoch": 0.6544111991385279,
      "grad_norm": 20.5,
      "learning_rate": 2.0183832012922085e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3706192.0,
      "step": 8508
    },
    {
      "epoch": 0.6544881162987463,
      "grad_norm": 16.625,
      "learning_rate": 2.018267825551881e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 3706496.0,
      "step": 8509
    },
    {
      "epoch": 0.6545650334589647,
      "grad_norm": 12.875,
      "learning_rate": 2.018152449811553e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 3707316.0,
      "step": 8510
    },
    {
      "epoch": 0.6546419506191832,
      "grad_norm": 17.125,
      "learning_rate": 2.0180370740712253e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 3707648.0,
      "step": 8511
    },
    {
      "epoch": 0.6547188677794016,
      "grad_norm": 16.125,
      "learning_rate": 2.0179216983308976e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 3707950.0,
      "step": 8512
    },
    {
      "epoch": 0.65479578493962,
      "grad_norm": 11.625,
      "learning_rate": 2.0178063225905703e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 3708306.0,
      "step": 8513
    },
    {
      "epoch": 0.6548727020998385,
      "grad_norm": 4.21875,
      "learning_rate": 2.0176909468502425e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3709064.0,
      "step": 8514
    },
    {
      "epoch": 0.6549496192600569,
      "grad_norm": 16.75,
      "learning_rate": 2.0175755711099143e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3709648.0,
      "step": 8515
    },
    {
      "epoch": 0.6550265364202754,
      "grad_norm": 11.875,
      "learning_rate": 2.017460195369587e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 3710008.0,
      "step": 8516
    },
    {
      "epoch": 0.6551034535804938,
      "grad_norm": 13.9375,
      "learning_rate": 2.0173448196292593e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3710288.0,
      "step": 8517
    },
    {
      "epoch": 0.6551803707407122,
      "grad_norm": 14.3125,
      "learning_rate": 2.0172294438889315e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3710850.0,
      "step": 8518
    },
    {
      "epoch": 0.6552572879009307,
      "grad_norm": 3.84375,
      "learning_rate": 2.0171140681486042e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3711396.0,
      "step": 8519
    },
    {
      "epoch": 0.6553342050611491,
      "grad_norm": 8.9375,
      "learning_rate": 2.0169986924082765e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3711726.0,
      "step": 8520
    },
    {
      "epoch": 0.6554111222213675,
      "grad_norm": 8.4375,
      "learning_rate": 2.0168833166679487e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3712274.0,
      "step": 8521
    },
    {
      "epoch": 0.6554880393815861,
      "grad_norm": 12.5,
      "learning_rate": 2.016767940927621e-06,
      "loss": 0.0068,
      "mean_token_accuracy": 0.9960745573043823,
      "num_tokens": 3713296.0,
      "step": 8522
    },
    {
      "epoch": 0.6555649565418045,
      "grad_norm": 7.90625,
      "learning_rate": 2.0166525651872933e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 3713568.0,
      "step": 8523
    },
    {
      "epoch": 0.6556418737020229,
      "grad_norm": 10.125,
      "learning_rate": 2.016537189446966e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3713912.0,
      "step": 8524
    },
    {
      "epoch": 0.6557187908622414,
      "grad_norm": 41.0,
      "learning_rate": 2.016421813706638e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3714196.0,
      "step": 8525
    },
    {
      "epoch": 0.6557957080224598,
      "grad_norm": 10.9375,
      "learning_rate": 2.01630643796631e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 3714770.0,
      "step": 8526
    },
    {
      "epoch": 0.6558726251826783,
      "grad_norm": 26.125,
      "learning_rate": 2.0161910622259827e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3715122.0,
      "step": 8527
    },
    {
      "epoch": 0.6559495423428967,
      "grad_norm": 42.75,
      "learning_rate": 2.016075686485655e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 3715442.0,
      "step": 8528
    },
    {
      "epoch": 0.6560264595031151,
      "grad_norm": 7.21875,
      "learning_rate": 2.0159603107453272e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 3716030.0,
      "step": 8529
    },
    {
      "epoch": 0.6561033766633336,
      "grad_norm": 19.125,
      "learning_rate": 2.015844935005e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 3716554.0,
      "step": 8530
    },
    {
      "epoch": 0.656180293823552,
      "grad_norm": 6.875,
      "learning_rate": 2.0157295592646717e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 3717314.0,
      "step": 8531
    },
    {
      "epoch": 0.6562572109837704,
      "grad_norm": 6.4375,
      "learning_rate": 2.0156141835243444e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3717848.0,
      "step": 8532
    },
    {
      "epoch": 0.656334128143989,
      "grad_norm": 14.0625,
      "learning_rate": 2.0154988077840167e-06,
      "loss": 0.0513,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 3718100.0,
      "step": 8533
    },
    {
      "epoch": 0.6564110453042074,
      "grad_norm": 9.9375,
      "learning_rate": 2.015383432043689e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 3718688.0,
      "step": 8534
    },
    {
      "epoch": 0.6564879624644259,
      "grad_norm": 73.5,
      "learning_rate": 2.015268056303361e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3719024.0,
      "step": 8535
    },
    {
      "epoch": 0.6565648796246443,
      "grad_norm": 13.0625,
      "learning_rate": 2.015152680563034e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9961977005004883,
      "num_tokens": 3719288.0,
      "step": 8536
    },
    {
      "epoch": 0.6566417967848627,
      "grad_norm": 15.6875,
      "learning_rate": 2.0150373048227057e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3719614.0,
      "step": 8537
    },
    {
      "epoch": 0.6567187139450812,
      "grad_norm": 7.0625,
      "learning_rate": 2.0149219290823784e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 3720428.0,
      "step": 8538
    },
    {
      "epoch": 0.6567956311052996,
      "grad_norm": 17.125,
      "learning_rate": 2.0148065533420507e-06,
      "loss": 0.005,
      "mean_token_accuracy": 0.9971618056297302,
      "num_tokens": 3721488.0,
      "step": 8539
    },
    {
      "epoch": 0.656872548265518,
      "grad_norm": 21.375,
      "learning_rate": 2.014691177601723e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 3722062.0,
      "step": 8540
    },
    {
      "epoch": 0.6569494654257365,
      "grad_norm": 9.5,
      "learning_rate": 2.0145758018613956e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3722632.0,
      "step": 8541
    },
    {
      "epoch": 0.6570263825859549,
      "grad_norm": 53.75,
      "learning_rate": 2.0144604261210674e-06,
      "loss": 0.051,
      "mean_token_accuracy": 0.980079710483551,
      "num_tokens": 3722884.0,
      "step": 8542
    },
    {
      "epoch": 0.6571032997461733,
      "grad_norm": 69.5,
      "learning_rate": 2.0143450503807397e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3723162.0,
      "step": 8543
    },
    {
      "epoch": 0.6571802169063918,
      "grad_norm": 40.75,
      "learning_rate": 2.0142296746404124e-06,
      "loss": 0.0421,
      "mean_token_accuracy": 0.9825072884559631,
      "num_tokens": 3723506.0,
      "step": 8544
    },
    {
      "epoch": 0.6572571340666102,
      "grad_norm": 3.28125,
      "learning_rate": 2.0141142989000846e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.998123824596405,
      "num_tokens": 3724040.0,
      "step": 8545
    },
    {
      "epoch": 0.6573340512268288,
      "grad_norm": 7.6875,
      "learning_rate": 2.013998923159757e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 3724396.0,
      "step": 8546
    },
    {
      "epoch": 0.6574109683870472,
      "grad_norm": 32.75,
      "learning_rate": 2.0138835474194296e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3724708.0,
      "step": 8547
    },
    {
      "epoch": 0.6574878855472656,
      "grad_norm": 7.75,
      "learning_rate": 2.0137681716791014e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 3725046.0,
      "step": 8548
    },
    {
      "epoch": 0.6575648027074841,
      "grad_norm": 9.75,
      "learning_rate": 2.013652795938774e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 3725632.0,
      "step": 8549
    },
    {
      "epoch": 0.6576417198677025,
      "grad_norm": 54.25,
      "learning_rate": 2.0135374201984463e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 3725960.0,
      "step": 8550
    },
    {
      "epoch": 0.6577186370279209,
      "grad_norm": 39.75,
      "learning_rate": 2.0134220444581186e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9922178983688354,
      "num_tokens": 3726218.0,
      "step": 8551
    },
    {
      "epoch": 0.6577955541881394,
      "grad_norm": 31.625,
      "learning_rate": 2.0133066687177913e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9781022071838379,
      "num_tokens": 3726494.0,
      "step": 8552
    },
    {
      "epoch": 0.6578724713483578,
      "grad_norm": 20.5,
      "learning_rate": 2.013191292977463e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3726830.0,
      "step": 8553
    },
    {
      "epoch": 0.6579493885085763,
      "grad_norm": 10.1875,
      "learning_rate": 2.0130759172371354e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3727392.0,
      "step": 8554
    },
    {
      "epoch": 0.6580263056687947,
      "grad_norm": 35.75,
      "learning_rate": 2.012960541496808e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 3727666.0,
      "step": 8555
    },
    {
      "epoch": 0.6581032228290131,
      "grad_norm": 18.625,
      "learning_rate": 2.0128451657564803e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 3728008.0,
      "step": 8556
    },
    {
      "epoch": 0.6581801399892316,
      "grad_norm": 19.25,
      "learning_rate": 2.0127297900161526e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3728292.0,
      "step": 8557
    },
    {
      "epoch": 0.65825705714945,
      "grad_norm": 4.90625,
      "learning_rate": 2.012614414275825e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3728838.0,
      "step": 8558
    },
    {
      "epoch": 0.6583339743096684,
      "grad_norm": 3.4375,
      "learning_rate": 2.012499038535497e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9961464405059814,
      "num_tokens": 3729358.0,
      "step": 8559
    },
    {
      "epoch": 0.658410891469887,
      "grad_norm": 18.125,
      "learning_rate": 2.0123836627951698e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 3729654.0,
      "step": 8560
    },
    {
      "epoch": 0.6584878086301054,
      "grad_norm": 4.1875,
      "learning_rate": 2.012268287054842e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3730218.0,
      "step": 8561
    },
    {
      "epoch": 0.6585647257903238,
      "grad_norm": 13.75,
      "learning_rate": 2.0121529113145143e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9968253970146179,
      "num_tokens": 3730534.0,
      "step": 8562
    },
    {
      "epoch": 0.6586416429505423,
      "grad_norm": 7.0,
      "learning_rate": 2.012037535574187e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 3730884.0,
      "step": 8563
    },
    {
      "epoch": 0.6587185601107607,
      "grad_norm": 2.984375,
      "learning_rate": 2.011922159833859e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9961464405059814,
      "num_tokens": 3731404.0,
      "step": 8564
    },
    {
      "epoch": 0.6587954772709792,
      "grad_norm": 9.875,
      "learning_rate": 2.011806784093531e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 3731964.0,
      "step": 8565
    },
    {
      "epoch": 0.6588723944311976,
      "grad_norm": 9.1875,
      "learning_rate": 2.0116914083532037e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 3732284.0,
      "step": 8566
    },
    {
      "epoch": 0.658949311591416,
      "grad_norm": 19.625,
      "learning_rate": 2.011576032612876e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9809523820877075,
      "num_tokens": 3732600.0,
      "step": 8567
    },
    {
      "epoch": 0.6590262287516345,
      "grad_norm": 18.375,
      "learning_rate": 2.0114606568725483e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3732882.0,
      "step": 8568
    },
    {
      "epoch": 0.6591031459118529,
      "grad_norm": 7.34375,
      "learning_rate": 2.0113452811322205e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3733180.0,
      "step": 8569
    },
    {
      "epoch": 0.6591800630720713,
      "grad_norm": 3.25,
      "learning_rate": 2.0112299053918928e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 3733698.0,
      "step": 8570
    },
    {
      "epoch": 0.6592569802322898,
      "grad_norm": 7.90625,
      "learning_rate": 2.0111145296515655e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3734010.0,
      "step": 8571
    },
    {
      "epoch": 0.6593338973925083,
      "grad_norm": 10.375,
      "learning_rate": 2.0109991539112377e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3734572.0,
      "step": 8572
    },
    {
      "epoch": 0.6594108145527268,
      "grad_norm": 2.6875,
      "learning_rate": 2.01088377817091e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3735330.0,
      "step": 8573
    },
    {
      "epoch": 0.6594877317129452,
      "grad_norm": 8.9375,
      "learning_rate": 2.0107684024305827e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3735916.0,
      "step": 8574
    },
    {
      "epoch": 0.6595646488731636,
      "grad_norm": 16.0,
      "learning_rate": 2.0106530266902545e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3736226.0,
      "step": 8575
    },
    {
      "epoch": 0.6596415660333821,
      "grad_norm": 6.0,
      "learning_rate": 2.0105376509499268e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 3736742.0,
      "step": 8576
    },
    {
      "epoch": 0.6597184831936005,
      "grad_norm": 5.40625,
      "learning_rate": 2.0104222752095994e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9982046484947205,
      "num_tokens": 3737300.0,
      "step": 8577
    },
    {
      "epoch": 0.6597954003538189,
      "grad_norm": 7.0625,
      "learning_rate": 2.0103068994692717e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3737886.0,
      "step": 8578
    },
    {
      "epoch": 0.6598723175140374,
      "grad_norm": 36.5,
      "learning_rate": 2.010191523728944e-06,
      "loss": 0.0397,
      "mean_token_accuracy": 0.9786477088928223,
      "num_tokens": 3738168.0,
      "step": 8579
    },
    {
      "epoch": 0.6599492346742558,
      "grad_norm": 43.75,
      "learning_rate": 2.010076147988616e-06,
      "loss": 0.0506,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 3738444.0,
      "step": 8580
    },
    {
      "epoch": 0.6600261518344742,
      "grad_norm": 66.5,
      "learning_rate": 2.0099607722482885e-06,
      "loss": 0.0486,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 3738750.0,
      "step": 8581
    },
    {
      "epoch": 0.6601030689946927,
      "grad_norm": 49.5,
      "learning_rate": 2.009845396507961e-06,
      "loss": 0.0485,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3739028.0,
      "step": 8582
    },
    {
      "epoch": 0.6601799861549111,
      "grad_norm": 39.75,
      "learning_rate": 2.0097300207676334e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 3739366.0,
      "step": 8583
    },
    {
      "epoch": 0.6602569033151297,
      "grad_norm": 14.25,
      "learning_rate": 2.0096146450273057e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3739910.0,
      "step": 8584
    },
    {
      "epoch": 0.660333820475348,
      "grad_norm": 4.0,
      "learning_rate": 2.0094992692869783e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 3740486.0,
      "step": 8585
    },
    {
      "epoch": 0.6604107376355665,
      "grad_norm": 12.875,
      "learning_rate": 2.00938389354665e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3740794.0,
      "step": 8586
    },
    {
      "epoch": 0.660487654795785,
      "grad_norm": 15.625,
      "learning_rate": 2.0092685178063224e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3741358.0,
      "step": 8587
    },
    {
      "epoch": 0.6605645719560034,
      "grad_norm": 7.46875,
      "learning_rate": 2.009153142065995e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3742166.0,
      "step": 8588
    },
    {
      "epoch": 0.6606414891162218,
      "grad_norm": 12.3125,
      "learning_rate": 2.0090377663256674e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 3742474.0,
      "step": 8589
    },
    {
      "epoch": 0.6607184062764403,
      "grad_norm": 6.5,
      "learning_rate": 2.0089223905853396e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3743048.0,
      "step": 8590
    },
    {
      "epoch": 0.6607953234366587,
      "grad_norm": 20.375,
      "learning_rate": 2.008807014845012e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 3743394.0,
      "step": 8591
    },
    {
      "epoch": 0.6608722405968772,
      "grad_norm": 7.78125,
      "learning_rate": 2.008691639104684e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3743714.0,
      "step": 8592
    },
    {
      "epoch": 0.6609491577570956,
      "grad_norm": 8.3125,
      "learning_rate": 2.008576263364357e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3744036.0,
      "step": 8593
    },
    {
      "epoch": 0.661026074917314,
      "grad_norm": 12.5,
      "learning_rate": 2.008460887624029e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 3744354.0,
      "step": 8594
    },
    {
      "epoch": 0.6611029920775325,
      "grad_norm": 11.125,
      "learning_rate": 2.0083455118837014e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3744926.0,
      "step": 8595
    },
    {
      "epoch": 0.6611799092377509,
      "grad_norm": 7.875,
      "learning_rate": 2.0082301361433736e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3745222.0,
      "step": 8596
    },
    {
      "epoch": 0.6612568263979693,
      "grad_norm": 4.59375,
      "learning_rate": 2.008114760403046e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3745746.0,
      "step": 8597
    },
    {
      "epoch": 0.6613337435581879,
      "grad_norm": 14.6875,
      "learning_rate": 2.007999384662718e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 3746346.0,
      "step": 8598
    },
    {
      "epoch": 0.6614106607184063,
      "grad_norm": 13.25,
      "learning_rate": 2.007884008922391e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3746626.0,
      "step": 8599
    },
    {
      "epoch": 0.6614875778786247,
      "grad_norm": 29.5,
      "learning_rate": 2.007768633182063e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 3746912.0,
      "step": 8600
    },
    {
      "epoch": 0.6615644950388432,
      "grad_norm": 4.75,
      "learning_rate": 2.0076532574417353e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3747224.0,
      "step": 8601
    },
    {
      "epoch": 0.6616414121990616,
      "grad_norm": 9.125,
      "learning_rate": 2.0075378817014076e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3747478.0,
      "step": 8602
    },
    {
      "epoch": 0.6617183293592801,
      "grad_norm": 7.28125,
      "learning_rate": 2.00742250596108e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3748052.0,
      "step": 8603
    },
    {
      "epoch": 0.6617952465194985,
      "grad_norm": 22.375,
      "learning_rate": 2.007307130220752e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3748384.0,
      "step": 8604
    },
    {
      "epoch": 0.6618721636797169,
      "grad_norm": 11.25,
      "learning_rate": 2.0071917544804248e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3748692.0,
      "step": 8605
    },
    {
      "epoch": 0.6619490808399354,
      "grad_norm": 8.6875,
      "learning_rate": 2.007076378740097e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 3749000.0,
      "step": 8606
    },
    {
      "epoch": 0.6620259980001538,
      "grad_norm": 15.1875,
      "learning_rate": 2.0069610029997693e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3749302.0,
      "step": 8607
    },
    {
      "epoch": 0.6621029151603722,
      "grad_norm": 11.0,
      "learning_rate": 2.0068456272594416e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9943342804908752,
      "num_tokens": 3749656.0,
      "step": 8608
    },
    {
      "epoch": 0.6621798323205907,
      "grad_norm": 25.375,
      "learning_rate": 2.006730251519114e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.996874988079071,
      "num_tokens": 3749980.0,
      "step": 8609
    },
    {
      "epoch": 0.6622567494808091,
      "grad_norm": 3.5,
      "learning_rate": 2.0066148757787865e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 3750788.0,
      "step": 8610
    },
    {
      "epoch": 0.6623336666410277,
      "grad_norm": 13.1875,
      "learning_rate": 2.0064995000384588e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 3751068.0,
      "step": 8611
    },
    {
      "epoch": 0.6624105838012461,
      "grad_norm": 9.3125,
      "learning_rate": 2.006384124298131e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3751658.0,
      "step": 8612
    },
    {
      "epoch": 0.6624875009614645,
      "grad_norm": 7.03125,
      "learning_rate": 2.0062687485578033e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 3752218.0,
      "step": 8613
    },
    {
      "epoch": 0.662564418121683,
      "grad_norm": 18.875,
      "learning_rate": 2.0061533728174755e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9927007555961609,
      "num_tokens": 3752494.0,
      "step": 8614
    },
    {
      "epoch": 0.6626413352819014,
      "grad_norm": 12.0625,
      "learning_rate": 2.0060379970771478e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3752818.0,
      "step": 8615
    },
    {
      "epoch": 0.6627182524421198,
      "grad_norm": 6.53125,
      "learning_rate": 2.0059226213368205e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3753118.0,
      "step": 8616
    },
    {
      "epoch": 0.6627951696023383,
      "grad_norm": 4.5,
      "learning_rate": 2.0058072455964927e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3753668.0,
      "step": 8617
    },
    {
      "epoch": 0.6628720867625567,
      "grad_norm": 12.875,
      "learning_rate": 2.005691869856165e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3753928.0,
      "step": 8618
    },
    {
      "epoch": 0.6629490039227752,
      "grad_norm": 4.6875,
      "learning_rate": 2.0055764941158372e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3754480.0,
      "step": 8619
    },
    {
      "epoch": 0.6630259210829936,
      "grad_norm": 5.53125,
      "learning_rate": 2.0054611183755095e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3755044.0,
      "step": 8620
    },
    {
      "epoch": 0.663102838243212,
      "grad_norm": 4.75,
      "learning_rate": 2.005345742635182e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3755598.0,
      "step": 8621
    },
    {
      "epoch": 0.6631797554034305,
      "grad_norm": 10.75,
      "learning_rate": 2.0052303668948544e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 3755886.0,
      "step": 8622
    },
    {
      "epoch": 0.663256672563649,
      "grad_norm": 39.0,
      "learning_rate": 2.0051149911545267e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 3756166.0,
      "step": 8623
    },
    {
      "epoch": 0.6633335897238674,
      "grad_norm": 61.75,
      "learning_rate": 2.004999615414199e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3756444.0,
      "step": 8624
    },
    {
      "epoch": 0.6634105068840859,
      "grad_norm": 8.6875,
      "learning_rate": 2.004884239673871e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 3756762.0,
      "step": 8625
    },
    {
      "epoch": 0.6634874240443043,
      "grad_norm": 4.03125,
      "learning_rate": 2.0047688639335435e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3757308.0,
      "step": 8626
    },
    {
      "epoch": 0.6635643412045227,
      "grad_norm": 8.6875,
      "learning_rate": 2.004653488193216e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3757630.0,
      "step": 8627
    },
    {
      "epoch": 0.6636412583647412,
      "grad_norm": 18.0,
      "learning_rate": 2.0045381124528884e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3757910.0,
      "step": 8628
    },
    {
      "epoch": 0.6637181755249596,
      "grad_norm": 17.375,
      "learning_rate": 2.0044227367125607e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3758244.0,
      "step": 8629
    },
    {
      "epoch": 0.6637950926851781,
      "grad_norm": 7.9375,
      "learning_rate": 2.004307360972233e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 3759064.0,
      "step": 8630
    },
    {
      "epoch": 0.6638720098453965,
      "grad_norm": 32.0,
      "learning_rate": 2.004191985231905e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3759396.0,
      "step": 8631
    },
    {
      "epoch": 0.6639489270056149,
      "grad_norm": 4.75,
      "learning_rate": 2.004076609491578e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3759970.0,
      "step": 8632
    },
    {
      "epoch": 0.6640258441658334,
      "grad_norm": 8.875,
      "learning_rate": 2.00396123375125e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 3760520.0,
      "step": 8633
    },
    {
      "epoch": 0.6641027613260518,
      "grad_norm": 5.5,
      "learning_rate": 2.003845858010922e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 3761072.0,
      "step": 8634
    },
    {
      "epoch": 0.6641796784862702,
      "grad_norm": 8.25,
      "learning_rate": 2.0037304822705946e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3761646.0,
      "step": 8635
    },
    {
      "epoch": 0.6642565956464888,
      "grad_norm": 3.8125,
      "learning_rate": 2.003615106530267e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 3762190.0,
      "step": 8636
    },
    {
      "epoch": 0.6643335128067072,
      "grad_norm": 23.5,
      "learning_rate": 2.003499730789939e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3762530.0,
      "step": 8637
    },
    {
      "epoch": 0.6644104299669257,
      "grad_norm": 42.25,
      "learning_rate": 2.003384355049612e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3762844.0,
      "step": 8638
    },
    {
      "epoch": 0.6644873471271441,
      "grad_norm": 23.0,
      "learning_rate": 2.003268979309284e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3763190.0,
      "step": 8639
    },
    {
      "epoch": 0.6645642642873625,
      "grad_norm": 4.34375,
      "learning_rate": 2.0031536035689564e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9981850981712341,
      "num_tokens": 3763742.0,
      "step": 8640
    },
    {
      "epoch": 0.664641181447581,
      "grad_norm": 5.1875,
      "learning_rate": 2.0030382278286286e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 3764262.0,
      "step": 8641
    },
    {
      "epoch": 0.6647180986077994,
      "grad_norm": 44.25,
      "learning_rate": 2.002922852088301e-06,
      "loss": 0.0463,
      "mean_token_accuracy": 0.97826087474823,
      "num_tokens": 3764540.0,
      "step": 8642
    },
    {
      "epoch": 0.6647950157680178,
      "grad_norm": 23.0,
      "learning_rate": 2.0028074763479736e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 3764880.0,
      "step": 8643
    },
    {
      "epoch": 0.6648719329282363,
      "grad_norm": 6.28125,
      "learning_rate": 2.002692100607646e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 3765448.0,
      "step": 8644
    },
    {
      "epoch": 0.6649488500884547,
      "grad_norm": 6.625,
      "learning_rate": 2.0025767248673176e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 3765776.0,
      "step": 8645
    },
    {
      "epoch": 0.6650257672486731,
      "grad_norm": 5.40625,
      "learning_rate": 2.0024613491269903e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 3766346.0,
      "step": 8646
    },
    {
      "epoch": 0.6651026844088916,
      "grad_norm": 5.46875,
      "learning_rate": 2.0023459733866626e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9987623691558838,
      "num_tokens": 3767156.0,
      "step": 8647
    },
    {
      "epoch": 0.66517960156911,
      "grad_norm": 33.75,
      "learning_rate": 2.002230597646335e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 3767494.0,
      "step": 8648
    },
    {
      "epoch": 0.6652565187293286,
      "grad_norm": 60.0,
      "learning_rate": 2.0021152219060075e-06,
      "loss": 0.04,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3767800.0,
      "step": 8649
    },
    {
      "epoch": 0.665333435889547,
      "grad_norm": 5.90625,
      "learning_rate": 2.0019998461656798e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3768342.0,
      "step": 8650
    },
    {
      "epoch": 0.6654103530497654,
      "grad_norm": 8.5625,
      "learning_rate": 2.001884470425352e-06,
      "loss": 0.0069,
      "mean_token_accuracy": 0.9987623691558838,
      "num_tokens": 3769152.0,
      "step": 8651
    },
    {
      "epoch": 0.6654872702099839,
      "grad_norm": 19.125,
      "learning_rate": 2.0017690946850243e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3769486.0,
      "step": 8652
    },
    {
      "epoch": 0.6655641873702023,
      "grad_norm": 14.0,
      "learning_rate": 2.0016537189446966e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9987834692001343,
      "num_tokens": 3770310.0,
      "step": 8653
    },
    {
      "epoch": 0.6656411045304207,
      "grad_norm": 16.875,
      "learning_rate": 2.001538343204369e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3770622.0,
      "step": 8654
    },
    {
      "epoch": 0.6657180216906392,
      "grad_norm": 23.5,
      "learning_rate": 2.0014229674640415e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 3770916.0,
      "step": 8655
    },
    {
      "epoch": 0.6657949388508576,
      "grad_norm": 7.375,
      "learning_rate": 2.0013075917237133e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3771436.0,
      "step": 8656
    },
    {
      "epoch": 0.6658718560110761,
      "grad_norm": 15.5,
      "learning_rate": 2.001192215983386e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3771754.0,
      "step": 8657
    },
    {
      "epoch": 0.6659487731712945,
      "grad_norm": 4.34375,
      "learning_rate": 2.0010768402430583e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 3772248.0,
      "step": 8658
    },
    {
      "epoch": 0.6660256903315129,
      "grad_norm": 6.875,
      "learning_rate": 2.0009614645027305e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 3772540.0,
      "step": 8659
    },
    {
      "epoch": 0.6661026074917314,
      "grad_norm": 6.4375,
      "learning_rate": 2.000846088762403e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3772824.0,
      "step": 8660
    },
    {
      "epoch": 0.6661795246519499,
      "grad_norm": 11.625,
      "learning_rate": 2.0007307130220755e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3773146.0,
      "step": 8661
    },
    {
      "epoch": 0.6662564418121683,
      "grad_norm": 11.0625,
      "learning_rate": 2.0006153372817473e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3773424.0,
      "step": 8662
    },
    {
      "epoch": 0.6663333589723868,
      "grad_norm": 19.0,
      "learning_rate": 2.00049996154142e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3773748.0,
      "step": 8663
    },
    {
      "epoch": 0.6664102761326052,
      "grad_norm": 9.875,
      "learning_rate": 2.0003845858010922e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3774084.0,
      "step": 8664
    },
    {
      "epoch": 0.6664871932928236,
      "grad_norm": 27.25,
      "learning_rate": 2.0002692100607645e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3774410.0,
      "step": 8665
    },
    {
      "epoch": 0.6665641104530421,
      "grad_norm": 8.5625,
      "learning_rate": 2.000153834320437e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 3774946.0,
      "step": 8666
    },
    {
      "epoch": 0.6666410276132605,
      "grad_norm": 12.125,
      "learning_rate": 2.000038458580109e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 3775258.0,
      "step": 8667
    },
    {
      "epoch": 0.666717944773479,
      "grad_norm": 21.5,
      "learning_rate": 1.9999230828397817e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9916527271270752,
      "num_tokens": 3775858.0,
      "step": 8668
    },
    {
      "epoch": 0.6667948619336974,
      "grad_norm": 9.3125,
      "learning_rate": 1.999807707099454e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3776142.0,
      "step": 8669
    },
    {
      "epoch": 0.6668717790939158,
      "grad_norm": 8.0,
      "learning_rate": 1.9996923313591262e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 3776696.0,
      "step": 8670
    },
    {
      "epoch": 0.6669486962541343,
      "grad_norm": 45.5,
      "learning_rate": 1.999576955618799e-06,
      "loss": 0.0383,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3776980.0,
      "step": 8671
    },
    {
      "epoch": 0.6670256134143527,
      "grad_norm": 17.25,
      "learning_rate": 1.9994615798784707e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9813084006309509,
      "num_tokens": 3777302.0,
      "step": 8672
    },
    {
      "epoch": 0.6671025305745711,
      "grad_norm": 12.3125,
      "learning_rate": 1.999346204138143e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9896013736724854,
      "num_tokens": 3777880.0,
      "step": 8673
    },
    {
      "epoch": 0.6671794477347897,
      "grad_norm": 8.4375,
      "learning_rate": 1.9992308283978157e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3778190.0,
      "step": 8674
    },
    {
      "epoch": 0.6672563648950081,
      "grad_norm": 34.5,
      "learning_rate": 1.999115452657488e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3778510.0,
      "step": 8675
    },
    {
      "epoch": 0.6673332820552266,
      "grad_norm": 12.5625,
      "learning_rate": 1.99900007691716e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9952152967453003,
      "num_tokens": 3779558.0,
      "step": 8676
    },
    {
      "epoch": 0.667410199215445,
      "grad_norm": 4.1875,
      "learning_rate": 1.998884701176833e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 3780126.0,
      "step": 8677
    },
    {
      "epoch": 0.6674871163756634,
      "grad_norm": 19.25,
      "learning_rate": 1.9987693254365047e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9830028414726257,
      "num_tokens": 3780480.0,
      "step": 8678
    },
    {
      "epoch": 0.6675640335358819,
      "grad_norm": 26.75,
      "learning_rate": 1.9986539496961774e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 3780790.0,
      "step": 8679
    },
    {
      "epoch": 0.6676409506961003,
      "grad_norm": 15.5,
      "learning_rate": 1.9985385739558496e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3781112.0,
      "step": 8680
    },
    {
      "epoch": 0.6677178678563187,
      "grad_norm": 13.0,
      "learning_rate": 1.998423198215522e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 3782156.0,
      "step": 8681
    },
    {
      "epoch": 0.6677947850165372,
      "grad_norm": 26.375,
      "learning_rate": 1.9983078224751946e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3782500.0,
      "step": 8682
    },
    {
      "epoch": 0.6678717021767556,
      "grad_norm": 12.4375,
      "learning_rate": 1.9981924467348664e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 3782836.0,
      "step": 8683
    },
    {
      "epoch": 0.667948619336974,
      "grad_norm": 10.8125,
      "learning_rate": 1.9980770709945387e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3783160.0,
      "step": 8684
    },
    {
      "epoch": 0.6680255364971925,
      "grad_norm": 9.6875,
      "learning_rate": 1.9979616952542114e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9961977005004883,
      "num_tokens": 3783424.0,
      "step": 8685
    },
    {
      "epoch": 0.6681024536574109,
      "grad_norm": 11.3125,
      "learning_rate": 1.9978463195138836e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3783980.0,
      "step": 8686
    },
    {
      "epoch": 0.6681793708176295,
      "grad_norm": 13.875,
      "learning_rate": 1.997730943773556e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3784300.0,
      "step": 8687
    },
    {
      "epoch": 0.6682562879778479,
      "grad_norm": 5.4375,
      "learning_rate": 1.9976155680332286e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3784848.0,
      "step": 8688
    },
    {
      "epoch": 0.6683332051380663,
      "grad_norm": 18.375,
      "learning_rate": 1.9975001922929004e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 3785168.0,
      "step": 8689
    },
    {
      "epoch": 0.6684101222982848,
      "grad_norm": 16.875,
      "learning_rate": 1.997384816552573e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 3785464.0,
      "step": 8690
    },
    {
      "epoch": 0.6684870394585032,
      "grad_norm": 10.875,
      "learning_rate": 1.9972694408122453e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3785760.0,
      "step": 8691
    },
    {
      "epoch": 0.6685639566187216,
      "grad_norm": 4.59375,
      "learning_rate": 1.9971540650719176e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3786542.0,
      "step": 8692
    },
    {
      "epoch": 0.6686408737789401,
      "grad_norm": 7.21875,
      "learning_rate": 1.9970386893315903e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3787076.0,
      "step": 8693
    },
    {
      "epoch": 0.6687177909391585,
      "grad_norm": 13.6875,
      "learning_rate": 1.996923313591262e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3787396.0,
      "step": 8694
    },
    {
      "epoch": 0.668794708099377,
      "grad_norm": 8.3125,
      "learning_rate": 1.9968079378509344e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3787720.0,
      "step": 8695
    },
    {
      "epoch": 0.6688716252595954,
      "grad_norm": 7.25,
      "learning_rate": 1.996692562110607e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 3788316.0,
      "step": 8696
    },
    {
      "epoch": 0.6689485424198138,
      "grad_norm": 5.25,
      "learning_rate": 1.9965771863702793e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 3788884.0,
      "step": 8697
    },
    {
      "epoch": 0.6690254595800323,
      "grad_norm": 8.9375,
      "learning_rate": 1.9964618106299516e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3789202.0,
      "step": 8698
    },
    {
      "epoch": 0.6691023767402507,
      "grad_norm": 4.09375,
      "learning_rate": 1.996346434889624e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3789530.0,
      "step": 8699
    },
    {
      "epoch": 0.6691792939004692,
      "grad_norm": 12.625,
      "learning_rate": 1.996231059149296e-06,
      "loss": 0.0352,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3789762.0,
      "step": 8700
    },
    {
      "epoch": 0.6692562110606877,
      "grad_norm": 3.203125,
      "learning_rate": 1.9961156834089688e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3790280.0,
      "step": 8701
    },
    {
      "epoch": 0.6693331282209061,
      "grad_norm": 8.875,
      "learning_rate": 1.996000307668641e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3790562.0,
      "step": 8702
    },
    {
      "epoch": 0.6694100453811245,
      "grad_norm": 4.09375,
      "learning_rate": 1.9958849319283133e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9962476491928101,
      "num_tokens": 3791096.0,
      "step": 8703
    },
    {
      "epoch": 0.669486962541343,
      "grad_norm": 4.1875,
      "learning_rate": 1.9957695561879855e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3791638.0,
      "step": 8704
    },
    {
      "epoch": 0.6695638797015614,
      "grad_norm": 5.78125,
      "learning_rate": 1.9956541804476578e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9933554530143738,
      "num_tokens": 3791940.0,
      "step": 8705
    },
    {
      "epoch": 0.6696407968617799,
      "grad_norm": 3.421875,
      "learning_rate": 1.99553880470733e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3792464.0,
      "step": 8706
    },
    {
      "epoch": 0.6697177140219983,
      "grad_norm": 7.9375,
      "learning_rate": 1.9954234289670027e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 3793038.0,
      "step": 8707
    },
    {
      "epoch": 0.6697946311822167,
      "grad_norm": 9.5625,
      "learning_rate": 1.995308053226675e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 3793410.0,
      "step": 8708
    },
    {
      "epoch": 0.6698715483424352,
      "grad_norm": 4.03125,
      "learning_rate": 1.9951926774863472e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3793950.0,
      "step": 8709
    },
    {
      "epoch": 0.6699484655026536,
      "grad_norm": 6.90625,
      "learning_rate": 1.9950773017460195e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3794256.0,
      "step": 8710
    },
    {
      "epoch": 0.670025382662872,
      "grad_norm": 2.546875,
      "learning_rate": 1.9949619260056918e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3794798.0,
      "step": 8711
    },
    {
      "epoch": 0.6701022998230906,
      "grad_norm": 9.0,
      "learning_rate": 1.994846550265364e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 3795154.0,
      "step": 8712
    },
    {
      "epoch": 0.670179216983309,
      "grad_norm": 3.609375,
      "learning_rate": 1.9947311745250367e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 3795952.0,
      "step": 8713
    },
    {
      "epoch": 0.6702561341435275,
      "grad_norm": 13.4375,
      "learning_rate": 1.994615798784709e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3796296.0,
      "step": 8714
    },
    {
      "epoch": 0.6703330513037459,
      "grad_norm": 12.875,
      "learning_rate": 1.9945004230443812e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3796618.0,
      "step": 8715
    },
    {
      "epoch": 0.6704099684639643,
      "grad_norm": 22.5,
      "learning_rate": 1.9943850473040535e-06,
      "loss": 0.0393,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3796876.0,
      "step": 8716
    },
    {
      "epoch": 0.6704868856241828,
      "grad_norm": 9.0625,
      "learning_rate": 1.9942696715637257e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3797396.0,
      "step": 8717
    },
    {
      "epoch": 0.6705638027844012,
      "grad_norm": 12.9375,
      "learning_rate": 1.9941542958233984e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 3797736.0,
      "step": 8718
    },
    {
      "epoch": 0.6706407199446196,
      "grad_norm": 5.59375,
      "learning_rate": 1.9940389200830707e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3798332.0,
      "step": 8719
    },
    {
      "epoch": 0.6707176371048381,
      "grad_norm": 11.625,
      "learning_rate": 1.993923544342743e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3798644.0,
      "step": 8720
    },
    {
      "epoch": 0.6707945542650565,
      "grad_norm": 3.265625,
      "learning_rate": 1.993808168602415e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3799214.0,
      "step": 8721
    },
    {
      "epoch": 0.6708714714252749,
      "grad_norm": 6.0625,
      "learning_rate": 1.9936927928620874e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 3799548.0,
      "step": 8722
    },
    {
      "epoch": 0.6709483885854934,
      "grad_norm": 5.75,
      "learning_rate": 1.9935774171217597e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3800110.0,
      "step": 8723
    },
    {
      "epoch": 0.6710253057457118,
      "grad_norm": 9.0625,
      "learning_rate": 1.9934620413814324e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3800370.0,
      "step": 8724
    },
    {
      "epoch": 0.6711022229059304,
      "grad_norm": 3.828125,
      "learning_rate": 1.9933466656411046e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3800916.0,
      "step": 8725
    },
    {
      "epoch": 0.6711791400661488,
      "grad_norm": 3.5625,
      "learning_rate": 1.993231289900777e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3801458.0,
      "step": 8726
    },
    {
      "epoch": 0.6712560572263672,
      "grad_norm": 4.84375,
      "learning_rate": 1.993115914160449e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3802026.0,
      "step": 8727
    },
    {
      "epoch": 0.6713329743865857,
      "grad_norm": 9.5,
      "learning_rate": 1.9930005384201214e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3802348.0,
      "step": 8728
    },
    {
      "epoch": 0.6714098915468041,
      "grad_norm": 7.21875,
      "learning_rate": 1.992885162679794e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 3802688.0,
      "step": 8729
    },
    {
      "epoch": 0.6714868087070225,
      "grad_norm": 8.8125,
      "learning_rate": 1.9927697869394664e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 3802950.0,
      "step": 8730
    },
    {
      "epoch": 0.671563725867241,
      "grad_norm": 4.21875,
      "learning_rate": 1.9926544111991386e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 3803446.0,
      "step": 8731
    },
    {
      "epoch": 0.6716406430274594,
      "grad_norm": 16.25,
      "learning_rate": 1.992539035458811e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 3803804.0,
      "step": 8732
    },
    {
      "epoch": 0.6717175601876779,
      "grad_norm": 25.25,
      "learning_rate": 1.992423659718483e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3804140.0,
      "step": 8733
    },
    {
      "epoch": 0.6717944773478963,
      "grad_norm": 4.15625,
      "learning_rate": 1.9923082839781554e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3804680.0,
      "step": 8734
    },
    {
      "epoch": 0.6718713945081147,
      "grad_norm": 11.1875,
      "learning_rate": 1.992192908237828e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 3805042.0,
      "step": 8735
    },
    {
      "epoch": 0.6719483116683332,
      "grad_norm": 7.28125,
      "learning_rate": 1.9920775324975003e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3805598.0,
      "step": 8736
    },
    {
      "epoch": 0.6720252288285516,
      "grad_norm": 4.8125,
      "learning_rate": 1.9919621567571726e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 3806114.0,
      "step": 8737
    },
    {
      "epoch": 0.67210214598877,
      "grad_norm": 5.28125,
      "learning_rate": 1.991846781016845e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3806678.0,
      "step": 8738
    },
    {
      "epoch": 0.6721790631489886,
      "grad_norm": 9.0,
      "learning_rate": 1.991731405276517e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3806988.0,
      "step": 8739
    },
    {
      "epoch": 0.672255980309207,
      "grad_norm": 4.125,
      "learning_rate": 1.9916160295361898e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3807562.0,
      "step": 8740
    },
    {
      "epoch": 0.6723328974694255,
      "grad_norm": 11.625,
      "learning_rate": 1.991500653795862e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 3807910.0,
      "step": 8741
    },
    {
      "epoch": 0.6724098146296439,
      "grad_norm": 23.75,
      "learning_rate": 1.9913852780555343e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 3808240.0,
      "step": 8742
    },
    {
      "epoch": 0.6724867317898623,
      "grad_norm": 14.5,
      "learning_rate": 1.9912699023152066e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 3808514.0,
      "step": 8743
    },
    {
      "epoch": 0.6725636489500808,
      "grad_norm": 17.25,
      "learning_rate": 1.991154526574879e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 3809110.0,
      "step": 8744
    },
    {
      "epoch": 0.6726405661102992,
      "grad_norm": 10.5,
      "learning_rate": 1.991039150834551e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 3809400.0,
      "step": 8745
    },
    {
      "epoch": 0.6727174832705176,
      "grad_norm": 5.8125,
      "learning_rate": 1.9909237750942238e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 3809950.0,
      "step": 8746
    },
    {
      "epoch": 0.6727944004307361,
      "grad_norm": 6.15625,
      "learning_rate": 1.990808399353896e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 3810466.0,
      "step": 8747
    },
    {
      "epoch": 0.6728713175909545,
      "grad_norm": 14.1875,
      "learning_rate": 1.9906930236135683e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 3811028.0,
      "step": 8748
    },
    {
      "epoch": 0.6729482347511729,
      "grad_norm": 8.375,
      "learning_rate": 1.9905776478732405e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3811314.0,
      "step": 8749
    },
    {
      "epoch": 0.6730251519113915,
      "grad_norm": 12.0,
      "learning_rate": 1.990462272132913e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 3811648.0,
      "step": 8750
    },
    {
      "epoch": 0.6731020690716099,
      "grad_norm": 19.625,
      "learning_rate": 1.9903468963925855e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3811970.0,
      "step": 8751
    },
    {
      "epoch": 0.6731789862318284,
      "grad_norm": 16.25,
      "learning_rate": 1.9902315206522577e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9828178882598877,
      "num_tokens": 3812262.0,
      "step": 8752
    },
    {
      "epoch": 0.6732559033920468,
      "grad_norm": 11.3125,
      "learning_rate": 1.99011614491193e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9983022212982178,
      "num_tokens": 3812852.0,
      "step": 8753
    },
    {
      "epoch": 0.6733328205522652,
      "grad_norm": 13.9375,
      "learning_rate": 1.9900007691716022e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3813182.0,
      "step": 8754
    },
    {
      "epoch": 0.6734097377124837,
      "grad_norm": 3.671875,
      "learning_rate": 1.9898853934312745e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3813738.0,
      "step": 8755
    },
    {
      "epoch": 0.6734866548727021,
      "grad_norm": 4.3125,
      "learning_rate": 1.9897700176909468e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3814280.0,
      "step": 8756
    },
    {
      "epoch": 0.6735635720329205,
      "grad_norm": 3.59375,
      "learning_rate": 1.9896546419506194e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3814824.0,
      "step": 8757
    },
    {
      "epoch": 0.673640489193139,
      "grad_norm": 12.125,
      "learning_rate": 1.9895392662102917e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 3815098.0,
      "step": 8758
    },
    {
      "epoch": 0.6737174063533574,
      "grad_norm": 12.0,
      "learning_rate": 1.989423890469964e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3815384.0,
      "step": 8759
    },
    {
      "epoch": 0.6737943235135759,
      "grad_norm": 15.5,
      "learning_rate": 1.9893085147296362e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 3815706.0,
      "step": 8760
    },
    {
      "epoch": 0.6738712406737943,
      "grad_norm": 6.3125,
      "learning_rate": 1.9891931389893085e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3816252.0,
      "step": 8761
    },
    {
      "epoch": 0.6739481578340127,
      "grad_norm": 16.75,
      "learning_rate": 1.9890777632489807e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9966555237770081,
      "num_tokens": 3816554.0,
      "step": 8762
    },
    {
      "epoch": 0.6740250749942313,
      "grad_norm": 10.8125,
      "learning_rate": 1.9889623875086534e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 3817380.0,
      "step": 8763
    },
    {
      "epoch": 0.6741019921544497,
      "grad_norm": 11.6875,
      "learning_rate": 1.9888470117683257e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3817936.0,
      "step": 8764
    },
    {
      "epoch": 0.6741789093146681,
      "grad_norm": 1.9140625,
      "learning_rate": 1.988731636027998e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3818478.0,
      "step": 8765
    },
    {
      "epoch": 0.6742558264748866,
      "grad_norm": 14.75,
      "learning_rate": 1.98861626028767e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3818802.0,
      "step": 8766
    },
    {
      "epoch": 0.674332743635105,
      "grad_norm": 43.5,
      "learning_rate": 1.9885008845473424e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 3819110.0,
      "step": 8767
    },
    {
      "epoch": 0.6744096607953234,
      "grad_norm": 10.625,
      "learning_rate": 1.988385508807015e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 3819646.0,
      "step": 8768
    },
    {
      "epoch": 0.6744865779555419,
      "grad_norm": 3.296875,
      "learning_rate": 1.9882701330666874e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 3820428.0,
      "step": 8769
    },
    {
      "epoch": 0.6745634951157603,
      "grad_norm": 12.0625,
      "learning_rate": 1.9881547573263592e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 3820988.0,
      "step": 8770
    },
    {
      "epoch": 0.6746404122759788,
      "grad_norm": 7.40625,
      "learning_rate": 1.988039381586032e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.992337167263031,
      "num_tokens": 3821250.0,
      "step": 8771
    },
    {
      "epoch": 0.6747173294361972,
      "grad_norm": 5.0625,
      "learning_rate": 1.987924005845704e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3821824.0,
      "step": 8772
    },
    {
      "epoch": 0.6747942465964156,
      "grad_norm": 4.5,
      "learning_rate": 1.9878086301053764e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 3822356.0,
      "step": 8773
    },
    {
      "epoch": 0.6748711637566341,
      "grad_norm": 6.59375,
      "learning_rate": 1.987693254365049e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3822896.0,
      "step": 8774
    },
    {
      "epoch": 0.6749480809168525,
      "grad_norm": 17.375,
      "learning_rate": 1.987577878624721e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 3823256.0,
      "step": 8775
    },
    {
      "epoch": 0.675024998077071,
      "grad_norm": 34.75,
      "learning_rate": 1.9874625028843936e-06,
      "loss": 0.047,
      "mean_token_accuracy": 0.9787985682487488,
      "num_tokens": 3823540.0,
      "step": 8776
    },
    {
      "epoch": 0.6751019152372895,
      "grad_norm": 26.75,
      "learning_rate": 1.987347127144066e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3823794.0,
      "step": 8777
    },
    {
      "epoch": 0.6751788323975079,
      "grad_norm": 29.5,
      "learning_rate": 1.987231751403738e-06,
      "loss": 0.0443,
      "mean_token_accuracy": 0.9786477088928223,
      "num_tokens": 3824076.0,
      "step": 8778
    },
    {
      "epoch": 0.6752557495577264,
      "grad_norm": 5.875,
      "learning_rate": 1.987116375663411e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3824656.0,
      "step": 8779
    },
    {
      "epoch": 0.6753326667179448,
      "grad_norm": 20.625,
      "learning_rate": 1.987000999923083e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3824952.0,
      "step": 8780
    },
    {
      "epoch": 0.6754095838781632,
      "grad_norm": 6.59375,
      "learning_rate": 1.986885624182755e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3825502.0,
      "step": 8781
    },
    {
      "epoch": 0.6754865010383817,
      "grad_norm": 14.875,
      "learning_rate": 1.9867702484424276e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9858757257461548,
      "num_tokens": 3825858.0,
      "step": 8782
    },
    {
      "epoch": 0.6755634181986001,
      "grad_norm": 3.859375,
      "learning_rate": 1.9866548727021e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 3826662.0,
      "step": 8783
    },
    {
      "epoch": 0.6756403353588185,
      "grad_norm": 6.5625,
      "learning_rate": 1.986539496961772e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3827236.0,
      "step": 8784
    },
    {
      "epoch": 0.675717252519037,
      "grad_norm": 3.71875,
      "learning_rate": 1.986424121221445e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 3827776.0,
      "step": 8785
    },
    {
      "epoch": 0.6757941696792554,
      "grad_norm": 17.375,
      "learning_rate": 1.9863087454811166e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3828054.0,
      "step": 8786
    },
    {
      "epoch": 0.6758710868394738,
      "grad_norm": 10.3125,
      "learning_rate": 1.9861933697407893e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 3828376.0,
      "step": 8787
    },
    {
      "epoch": 0.6759480039996923,
      "grad_norm": 10.0,
      "learning_rate": 1.9860779940004616e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3828688.0,
      "step": 8788
    },
    {
      "epoch": 0.6760249211599108,
      "grad_norm": 8.375,
      "learning_rate": 1.985962618260134e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3829022.0,
      "step": 8789
    },
    {
      "epoch": 0.6761018383201293,
      "grad_norm": 20.875,
      "learning_rate": 1.9858472425198065e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 3829380.0,
      "step": 8790
    },
    {
      "epoch": 0.6761787554803477,
      "grad_norm": 8.3125,
      "learning_rate": 1.9857318667794788e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3829730.0,
      "step": 8791
    },
    {
      "epoch": 0.6762556726405661,
      "grad_norm": 3.953125,
      "learning_rate": 1.9856164910391506e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3830252.0,
      "step": 8792
    },
    {
      "epoch": 0.6763325898007846,
      "grad_norm": 9.4375,
      "learning_rate": 1.9855011152988233e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3830552.0,
      "step": 8793
    },
    {
      "epoch": 0.676409506961003,
      "grad_norm": 20.5,
      "learning_rate": 1.9853857395584955e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3830884.0,
      "step": 8794
    },
    {
      "epoch": 0.6764864241212214,
      "grad_norm": 22.125,
      "learning_rate": 1.985270363818168e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3831144.0,
      "step": 8795
    },
    {
      "epoch": 0.6765633412814399,
      "grad_norm": 9.1875,
      "learning_rate": 1.9851549880778405e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3831468.0,
      "step": 8796
    },
    {
      "epoch": 0.6766402584416583,
      "grad_norm": 9.4375,
      "learning_rate": 1.9850396123375123e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3831812.0,
      "step": 8797
    },
    {
      "epoch": 0.6767171756018768,
      "grad_norm": 8.75,
      "learning_rate": 1.984924236597185e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3832142.0,
      "step": 8798
    },
    {
      "epoch": 0.6767940927620952,
      "grad_norm": 9.625,
      "learning_rate": 1.9848088608568572e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 3832700.0,
      "step": 8799
    },
    {
      "epoch": 0.6768710099223136,
      "grad_norm": 4.09375,
      "learning_rate": 1.9846934851165295e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 3833488.0,
      "step": 8800
    },
    {
      "epoch": 0.6769479270825322,
      "grad_norm": 7.90625,
      "learning_rate": 1.984578109376202e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 3833818.0,
      "step": 8801
    },
    {
      "epoch": 0.6770248442427506,
      "grad_norm": 11.4375,
      "learning_rate": 1.984462733635874e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3834144.0,
      "step": 8802
    },
    {
      "epoch": 0.677101761402969,
      "grad_norm": 16.75,
      "learning_rate": 1.9843473578955463e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3834404.0,
      "step": 8803
    },
    {
      "epoch": 0.6771786785631875,
      "grad_norm": 12.625,
      "learning_rate": 1.984231982155219e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3834720.0,
      "step": 8804
    },
    {
      "epoch": 0.6772555957234059,
      "grad_norm": 5.46875,
      "learning_rate": 1.9841166064148912e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3835304.0,
      "step": 8805
    },
    {
      "epoch": 0.6773325128836243,
      "grad_norm": 20.875,
      "learning_rate": 1.9840012306745635e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9929078221321106,
      "num_tokens": 3835588.0,
      "step": 8806
    },
    {
      "epoch": 0.6774094300438428,
      "grad_norm": 14.0625,
      "learning_rate": 1.983885854934236e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3835866.0,
      "step": 8807
    },
    {
      "epoch": 0.6774863472040612,
      "grad_norm": 3.84375,
      "learning_rate": 1.983770479193908e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3836178.0,
      "step": 8808
    },
    {
      "epoch": 0.6775632643642797,
      "grad_norm": 4.5,
      "learning_rate": 1.9836551034535807e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 3836702.0,
      "step": 8809
    },
    {
      "epoch": 0.6776401815244981,
      "grad_norm": 16.5,
      "learning_rate": 1.983539727713253e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3836998.0,
      "step": 8810
    },
    {
      "epoch": 0.6777170986847165,
      "grad_norm": 10.5625,
      "learning_rate": 1.983424351972925e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 3837338.0,
      "step": 8811
    },
    {
      "epoch": 0.677794015844935,
      "grad_norm": 3.953125,
      "learning_rate": 1.9833089762325975e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 3837894.0,
      "step": 8812
    },
    {
      "epoch": 0.6778709330051534,
      "grad_norm": 13.0625,
      "learning_rate": 1.9831936004922697e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3838172.0,
      "step": 8813
    },
    {
      "epoch": 0.6779478501653718,
      "grad_norm": 5.125,
      "learning_rate": 1.983078224751942e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3838482.0,
      "step": 8814
    },
    {
      "epoch": 0.6780247673255904,
      "grad_norm": 12.6875,
      "learning_rate": 1.9829628490116146e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 3838806.0,
      "step": 8815
    },
    {
      "epoch": 0.6781016844858088,
      "grad_norm": 17.25,
      "learning_rate": 1.982847473271287e-06,
      "loss": 0.0407,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 3839038.0,
      "step": 8816
    },
    {
      "epoch": 0.6781786016460273,
      "grad_norm": 7.34375,
      "learning_rate": 1.982732097530959e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 3839298.0,
      "step": 8817
    },
    {
      "epoch": 0.6782555188062457,
      "grad_norm": 4.0,
      "learning_rate": 1.982616721790632e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 3839846.0,
      "step": 8818
    },
    {
      "epoch": 0.6783324359664641,
      "grad_norm": 4.1875,
      "learning_rate": 1.9825013460503037e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 3840342.0,
      "step": 8819
    },
    {
      "epoch": 0.6784093531266826,
      "grad_norm": 3.890625,
      "learning_rate": 1.982385970309976e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9971292018890381,
      "num_tokens": 3841390.0,
      "step": 8820
    },
    {
      "epoch": 0.678486270286901,
      "grad_norm": 13.0625,
      "learning_rate": 1.9822705945696486e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3841962.0,
      "step": 8821
    },
    {
      "epoch": 0.6785631874471194,
      "grad_norm": 6.84375,
      "learning_rate": 1.982155218829321e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3842524.0,
      "step": 8822
    },
    {
      "epoch": 0.6786401046073379,
      "grad_norm": 3.140625,
      "learning_rate": 1.982039843088993e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3843070.0,
      "step": 8823
    },
    {
      "epoch": 0.6787170217675563,
      "grad_norm": 11.6875,
      "learning_rate": 1.9819244673486654e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9960784316062927,
      "num_tokens": 3843326.0,
      "step": 8824
    },
    {
      "epoch": 0.6787939389277747,
      "grad_norm": 15.625,
      "learning_rate": 1.9818090916083377e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 3843658.0,
      "step": 8825
    },
    {
      "epoch": 0.6788708560879932,
      "grad_norm": 3.859375,
      "learning_rate": 1.9816937158680103e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3843956.0,
      "step": 8826
    },
    {
      "epoch": 0.6789477732482117,
      "grad_norm": 3.515625,
      "learning_rate": 1.9815783401276826e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3844504.0,
      "step": 8827
    },
    {
      "epoch": 0.6790246904084302,
      "grad_norm": 33.25,
      "learning_rate": 1.981462964387355e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9804560542106628,
      "num_tokens": 3844812.0,
      "step": 8828
    },
    {
      "epoch": 0.6791016075686486,
      "grad_norm": 14.4375,
      "learning_rate": 1.9813475886470275e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.985029935836792,
      "num_tokens": 3845148.0,
      "step": 8829
    },
    {
      "epoch": 0.679178524728867,
      "grad_norm": 12.25,
      "learning_rate": 1.9812322129066994e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 3845712.0,
      "step": 8830
    },
    {
      "epoch": 0.6792554418890855,
      "grad_norm": 18.875,
      "learning_rate": 1.9811168371663716e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3846052.0,
      "step": 8831
    },
    {
      "epoch": 0.6793323590493039,
      "grad_norm": 4.125,
      "learning_rate": 1.9810014614260443e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 3846872.0,
      "step": 8832
    },
    {
      "epoch": 0.6794092762095223,
      "grad_norm": 6.53125,
      "learning_rate": 1.9808860856857166e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9959431886672974,
      "num_tokens": 3847366.0,
      "step": 8833
    },
    {
      "epoch": 0.6794861933697408,
      "grad_norm": 23.5,
      "learning_rate": 1.980770709945389e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3847652.0,
      "step": 8834
    },
    {
      "epoch": 0.6795631105299592,
      "grad_norm": 25.75,
      "learning_rate": 1.980655334205061e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3847958.0,
      "step": 8835
    },
    {
      "epoch": 0.6796400276901777,
      "grad_norm": 16.125,
      "learning_rate": 1.9805399584647333e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3848244.0,
      "step": 8836
    },
    {
      "epoch": 0.6797169448503961,
      "grad_norm": 38.25,
      "learning_rate": 1.980424582724406e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.984375,
      "num_tokens": 3848568.0,
      "step": 8837
    },
    {
      "epoch": 0.6797938620106145,
      "grad_norm": 3.375,
      "learning_rate": 1.9803092069840783e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3849114.0,
      "step": 8838
    },
    {
      "epoch": 0.679870779170833,
      "grad_norm": 19.875,
      "learning_rate": 1.9801938312437505e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 3849372.0,
      "step": 8839
    },
    {
      "epoch": 0.6799476963310515,
      "grad_norm": 13.625,
      "learning_rate": 1.980078455503423e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 3849658.0,
      "step": 8840
    },
    {
      "epoch": 0.6800246134912699,
      "grad_norm": 4.9375,
      "learning_rate": 1.979963079763095e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 3849968.0,
      "step": 8841
    },
    {
      "epoch": 0.6801015306514884,
      "grad_norm": 13.6875,
      "learning_rate": 1.9798477040227673e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 3850318.0,
      "step": 8842
    },
    {
      "epoch": 0.6801784478117068,
      "grad_norm": 20.0,
      "learning_rate": 1.97973232828244e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 3850656.0,
      "step": 8843
    },
    {
      "epoch": 0.6802553649719253,
      "grad_norm": 17.375,
      "learning_rate": 1.9796169525421123e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 3850980.0,
      "step": 8844
    },
    {
      "epoch": 0.6803322821321437,
      "grad_norm": 8.0,
      "learning_rate": 1.9795015768017845e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3851324.0,
      "step": 8845
    },
    {
      "epoch": 0.6804091992923621,
      "grad_norm": 12.25,
      "learning_rate": 1.9793862010614568e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 3851684.0,
      "step": 8846
    },
    {
      "epoch": 0.6804861164525806,
      "grad_norm": 17.75,
      "learning_rate": 1.979270825321129e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 3852044.0,
      "step": 8847
    },
    {
      "epoch": 0.680563033612799,
      "grad_norm": 12.5625,
      "learning_rate": 1.9791554495808017e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 3852406.0,
      "step": 8848
    },
    {
      "epoch": 0.6806399507730174,
      "grad_norm": 11.1875,
      "learning_rate": 1.979040073840474e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3852978.0,
      "step": 8849
    },
    {
      "epoch": 0.6807168679332359,
      "grad_norm": 13.5625,
      "learning_rate": 1.9789246981001462e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3853264.0,
      "step": 8850
    },
    {
      "epoch": 0.6807937850934543,
      "grad_norm": 8.9375,
      "learning_rate": 1.9788093223598185e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 3853852.0,
      "step": 8851
    },
    {
      "epoch": 0.6808707022536727,
      "grad_norm": 8.75,
      "learning_rate": 1.9786939466194907e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3854426.0,
      "step": 8852
    },
    {
      "epoch": 0.6809476194138913,
      "grad_norm": 7.4375,
      "learning_rate": 1.978578570879163e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 3855026.0,
      "step": 8853
    },
    {
      "epoch": 0.6810245365741097,
      "grad_norm": 17.5,
      "learning_rate": 1.9784631951388357e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3855342.0,
      "step": 8854
    },
    {
      "epoch": 0.6811014537343282,
      "grad_norm": 6.5,
      "learning_rate": 1.978347819398508e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 3855910.0,
      "step": 8855
    },
    {
      "epoch": 0.6811783708945466,
      "grad_norm": 34.25,
      "learning_rate": 1.97823244365818e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.978723406791687,
      "num_tokens": 3856194.0,
      "step": 8856
    },
    {
      "epoch": 0.681255288054765,
      "grad_norm": 11.5,
      "learning_rate": 1.9781170679178525e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 3856542.0,
      "step": 8857
    },
    {
      "epoch": 0.6813322052149835,
      "grad_norm": 10.0625,
      "learning_rate": 1.9780016921775247e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3857092.0,
      "step": 8858
    },
    {
      "epoch": 0.6814091223752019,
      "grad_norm": 8.9375,
      "learning_rate": 1.9778863164371974e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3857634.0,
      "step": 8859
    },
    {
      "epoch": 0.6814860395354203,
      "grad_norm": 5.5625,
      "learning_rate": 1.9777709406968697e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3858180.0,
      "step": 8860
    },
    {
      "epoch": 0.6815629566956388,
      "grad_norm": 10.6875,
      "learning_rate": 1.977655564956542e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3858490.0,
      "step": 8861
    },
    {
      "epoch": 0.6816398738558572,
      "grad_norm": 29.75,
      "learning_rate": 1.977540189216214e-06,
      "loss": 0.0421,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 3858768.0,
      "step": 8862
    },
    {
      "epoch": 0.6817167910160757,
      "grad_norm": 5.15625,
      "learning_rate": 1.9774248134758864e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 3859092.0,
      "step": 8863
    },
    {
      "epoch": 0.6817937081762941,
      "grad_norm": 24.625,
      "learning_rate": 1.9773094377355587e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 3859432.0,
      "step": 8864
    },
    {
      "epoch": 0.6818706253365125,
      "grad_norm": 7.8125,
      "learning_rate": 1.9771940619952314e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3859748.0,
      "step": 8865
    },
    {
      "epoch": 0.6819475424967311,
      "grad_norm": 5.90625,
      "learning_rate": 1.9770786862549036e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3860312.0,
      "step": 8866
    },
    {
      "epoch": 0.6820244596569495,
      "grad_norm": 6.09375,
      "learning_rate": 1.976963310514576e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3860898.0,
      "step": 8867
    },
    {
      "epoch": 0.6821013768171679,
      "grad_norm": 3.59375,
      "learning_rate": 1.976847934774248e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3861478.0,
      "step": 8868
    },
    {
      "epoch": 0.6821782939773864,
      "grad_norm": 4.53125,
      "learning_rate": 1.9767325590339204e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 3862264.0,
      "step": 8869
    },
    {
      "epoch": 0.6822552111376048,
      "grad_norm": 8.1875,
      "learning_rate": 1.9766171832935927e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 3862552.0,
      "step": 8870
    },
    {
      "epoch": 0.6823321282978232,
      "grad_norm": 6.15625,
      "learning_rate": 1.9765018075532653e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 3862864.0,
      "step": 8871
    },
    {
      "epoch": 0.6824090454580417,
      "grad_norm": 9.625,
      "learning_rate": 1.9763864318129376e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 3863388.0,
      "step": 8872
    },
    {
      "epoch": 0.6824859626182601,
      "grad_norm": 3.640625,
      "learning_rate": 1.97627105607261e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3863968.0,
      "step": 8873
    },
    {
      "epoch": 0.6825628797784786,
      "grad_norm": 10.0,
      "learning_rate": 1.976155680332282e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 3864272.0,
      "step": 8874
    },
    {
      "epoch": 0.682639796938697,
      "grad_norm": 6.4375,
      "learning_rate": 1.9760403045919544e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 3864602.0,
      "step": 8875
    },
    {
      "epoch": 0.6827167140989154,
      "grad_norm": 13.5625,
      "learning_rate": 1.975924928851627e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3864948.0,
      "step": 8876
    },
    {
      "epoch": 0.682793631259134,
      "grad_norm": 3.28125,
      "learning_rate": 1.9758095531112993e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3865470.0,
      "step": 8877
    },
    {
      "epoch": 0.6828705484193524,
      "grad_norm": 7.625,
      "learning_rate": 1.975694177370971e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 3865790.0,
      "step": 8878
    },
    {
      "epoch": 0.6829474655795708,
      "grad_norm": 4.28125,
      "learning_rate": 1.975578801630644e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3866348.0,
      "step": 8879
    },
    {
      "epoch": 0.6830243827397893,
      "grad_norm": 9.8125,
      "learning_rate": 1.975463425890316e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3866606.0,
      "step": 8880
    },
    {
      "epoch": 0.6831012999000077,
      "grad_norm": 4.21875,
      "learning_rate": 1.9753480501499883e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3867164.0,
      "step": 8881
    },
    {
      "epoch": 0.6831782170602262,
      "grad_norm": 10.5625,
      "learning_rate": 1.975232674409661e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3867482.0,
      "step": 8882
    },
    {
      "epoch": 0.6832551342204446,
      "grad_norm": 3.453125,
      "learning_rate": 1.9751172986693333e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3868028.0,
      "step": 8883
    },
    {
      "epoch": 0.683332051380663,
      "grad_norm": 8.3125,
      "learning_rate": 1.9750019229290055e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3868312.0,
      "step": 8884
    },
    {
      "epoch": 0.6834089685408815,
      "grad_norm": 11.3125,
      "learning_rate": 1.974886547188678e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 3868638.0,
      "step": 8885
    },
    {
      "epoch": 0.6834858857010999,
      "grad_norm": 4.25,
      "learning_rate": 1.97477117144835e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 3869196.0,
      "step": 8886
    },
    {
      "epoch": 0.6835628028613183,
      "grad_norm": 2.828125,
      "learning_rate": 1.9746557957080227e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 3869780.0,
      "step": 8887
    },
    {
      "epoch": 0.6836397200215368,
      "grad_norm": 4.25,
      "learning_rate": 1.974540419967695e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3870130.0,
      "step": 8888
    },
    {
      "epoch": 0.6837166371817552,
      "grad_norm": 9.3125,
      "learning_rate": 1.974425044227367e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3870714.0,
      "step": 8889
    },
    {
      "epoch": 0.6837935543419736,
      "grad_norm": 10.1875,
      "learning_rate": 1.9743096684870395e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 3871014.0,
      "step": 8890
    },
    {
      "epoch": 0.6838704715021922,
      "grad_norm": 3.15625,
      "learning_rate": 1.9741942927467118e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3871576.0,
      "step": 8891
    },
    {
      "epoch": 0.6839473886624106,
      "grad_norm": 7.9375,
      "learning_rate": 1.974078917006384e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9972222447395325,
      "num_tokens": 3871938.0,
      "step": 8892
    },
    {
      "epoch": 0.6840243058226291,
      "grad_norm": 22.75,
      "learning_rate": 1.9739635412660567e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3872222.0,
      "step": 8893
    },
    {
      "epoch": 0.6841012229828475,
      "grad_norm": 15.8125,
      "learning_rate": 1.973848165525729e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 3872572.0,
      "step": 8894
    },
    {
      "epoch": 0.6841781401430659,
      "grad_norm": 28.875,
      "learning_rate": 1.9737327897854012e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 3872868.0,
      "step": 8895
    },
    {
      "epoch": 0.6842550573032844,
      "grad_norm": 20.0,
      "learning_rate": 1.9736174140450735e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 3873138.0,
      "step": 8896
    },
    {
      "epoch": 0.6843319744635028,
      "grad_norm": 3.71875,
      "learning_rate": 1.9735020383047457e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9940476417541504,
      "num_tokens": 3873476.0,
      "step": 8897
    },
    {
      "epoch": 0.6844088916237212,
      "grad_norm": 5.75,
      "learning_rate": 1.9733866625644184e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3874060.0,
      "step": 8898
    },
    {
      "epoch": 0.6844858087839397,
      "grad_norm": 8.375,
      "learning_rate": 1.9732712868240907e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 3874630.0,
      "step": 8899
    },
    {
      "epoch": 0.6845627259441581,
      "grad_norm": 11.6875,
      "learning_rate": 1.9731559110837625e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 3874966.0,
      "step": 8900
    },
    {
      "epoch": 0.6846396431043766,
      "grad_norm": 17.0,
      "learning_rate": 1.973040535343435e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9882698059082031,
      "num_tokens": 3875308.0,
      "step": 8901
    },
    {
      "epoch": 0.684716560264595,
      "grad_norm": 15.625,
      "learning_rate": 1.9729251596031075e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3875566.0,
      "step": 8902
    },
    {
      "epoch": 0.6847934774248134,
      "grad_norm": 6.375,
      "learning_rate": 1.9728097838627797e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3875912.0,
      "step": 8903
    },
    {
      "epoch": 0.684870394585032,
      "grad_norm": 13.0625,
      "learning_rate": 1.9726944081224524e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3876224.0,
      "step": 8904
    },
    {
      "epoch": 0.6849473117452504,
      "grad_norm": 4.1875,
      "learning_rate": 1.9725790323821242e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3876508.0,
      "step": 8905
    },
    {
      "epoch": 0.6850242289054688,
      "grad_norm": 8.5625,
      "learning_rate": 1.972463656641797e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3876832.0,
      "step": 8906
    },
    {
      "epoch": 0.6851011460656873,
      "grad_norm": 4.78125,
      "learning_rate": 1.972348280901469e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3877138.0,
      "step": 8907
    },
    {
      "epoch": 0.6851780632259057,
      "grad_norm": 17.625,
      "learning_rate": 1.9722329051611414e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3877466.0,
      "step": 8908
    },
    {
      "epoch": 0.6852549803861241,
      "grad_norm": 9.1875,
      "learning_rate": 1.972117529420814e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 3877780.0,
      "step": 8909
    },
    {
      "epoch": 0.6853318975463426,
      "grad_norm": 11.0,
      "learning_rate": 1.9720021536804864e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 3878118.0,
      "step": 8910
    },
    {
      "epoch": 0.685408814706561,
      "grad_norm": 7.65625,
      "learning_rate": 1.971886777940158e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3878440.0,
      "step": 8911
    },
    {
      "epoch": 0.6854857318667795,
      "grad_norm": 2.875,
      "learning_rate": 1.971771402199831e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 3879220.0,
      "step": 8912
    },
    {
      "epoch": 0.6855626490269979,
      "grad_norm": 9.375,
      "learning_rate": 1.971656026459503e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 3879548.0,
      "step": 8913
    },
    {
      "epoch": 0.6856395661872163,
      "grad_norm": 38.75,
      "learning_rate": 1.9715406507191754e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3879876.0,
      "step": 8914
    },
    {
      "epoch": 0.6857164833474348,
      "grad_norm": 9.375,
      "learning_rate": 1.971425274978848e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3880158.0,
      "step": 8915
    },
    {
      "epoch": 0.6857934005076533,
      "grad_norm": 5.375,
      "learning_rate": 1.97130989923852e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 3880472.0,
      "step": 8916
    },
    {
      "epoch": 0.6858703176678717,
      "grad_norm": 7.625,
      "learning_rate": 1.9711945234981926e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 3881256.0,
      "step": 8917
    },
    {
      "epoch": 0.6859472348280902,
      "grad_norm": 30.125,
      "learning_rate": 1.971079147757865e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 3881620.0,
      "step": 8918
    },
    {
      "epoch": 0.6860241519883086,
      "grad_norm": 6.5625,
      "learning_rate": 1.970963772017537e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3881926.0,
      "step": 8919
    },
    {
      "epoch": 0.6861010691485271,
      "grad_norm": 11.375,
      "learning_rate": 1.97084839627721e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3882236.0,
      "step": 8920
    },
    {
      "epoch": 0.6861779863087455,
      "grad_norm": 5.46875,
      "learning_rate": 1.970733020536882e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 3882550.0,
      "step": 8921
    },
    {
      "epoch": 0.6862549034689639,
      "grad_norm": 5.40625,
      "learning_rate": 1.970617644796554e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 3883072.0,
      "step": 8922
    },
    {
      "epoch": 0.6863318206291824,
      "grad_norm": 36.25,
      "learning_rate": 1.9705022690562266e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 3883348.0,
      "step": 8923
    },
    {
      "epoch": 0.6864087377894008,
      "grad_norm": 9.8125,
      "learning_rate": 1.970386893315899e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 3883704.0,
      "step": 8924
    },
    {
      "epoch": 0.6864856549496192,
      "grad_norm": 4.84375,
      "learning_rate": 1.970271517575571e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 3884220.0,
      "step": 8925
    },
    {
      "epoch": 0.6865625721098377,
      "grad_norm": 5.15625,
      "learning_rate": 1.9701561418352438e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3884800.0,
      "step": 8926
    },
    {
      "epoch": 0.6866394892700561,
      "grad_norm": 5.59375,
      "learning_rate": 1.9700407660949156e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3885082.0,
      "step": 8927
    },
    {
      "epoch": 0.6867164064302745,
      "grad_norm": 11.4375,
      "learning_rate": 1.9699253903545883e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 3885672.0,
      "step": 8928
    },
    {
      "epoch": 0.686793323590493,
      "grad_norm": 10.6875,
      "learning_rate": 1.9698100146142605e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 3885990.0,
      "step": 8929
    },
    {
      "epoch": 0.6868702407507115,
      "grad_norm": 3.65625,
      "learning_rate": 1.969694638873933e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3886532.0,
      "step": 8930
    },
    {
      "epoch": 0.68694715791093,
      "grad_norm": 7.78125,
      "learning_rate": 1.969579263133605e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 3886834.0,
      "step": 8931
    },
    {
      "epoch": 0.6870240750711484,
      "grad_norm": 14.375,
      "learning_rate": 1.9694638873932777e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3887380.0,
      "step": 8932
    },
    {
      "epoch": 0.6871009922313668,
      "grad_norm": 3.234375,
      "learning_rate": 1.9693485116529496e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 3888424.0,
      "step": 8933
    },
    {
      "epoch": 0.6871779093915853,
      "grad_norm": 4.125,
      "learning_rate": 1.9692331359126223e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 3889004.0,
      "step": 8934
    },
    {
      "epoch": 0.6872548265518037,
      "grad_norm": 4.78125,
      "learning_rate": 1.9691177601722945e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3889524.0,
      "step": 8935
    },
    {
      "epoch": 0.6873317437120221,
      "grad_norm": 8.75,
      "learning_rate": 1.9690023844319668e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 3889830.0,
      "step": 8936
    },
    {
      "epoch": 0.6874086608722406,
      "grad_norm": 27.25,
      "learning_rate": 1.9688870086916395e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3890110.0,
      "step": 8937
    },
    {
      "epoch": 0.687485578032459,
      "grad_norm": 18.5,
      "learning_rate": 1.9687716329513113e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 3890682.0,
      "step": 8938
    },
    {
      "epoch": 0.6875624951926775,
      "grad_norm": 19.25,
      "learning_rate": 1.9686562572109835e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 3891026.0,
      "step": 8939
    },
    {
      "epoch": 0.6876394123528959,
      "grad_norm": 7.46875,
      "learning_rate": 1.9685408814706562e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 3891600.0,
      "step": 8940
    },
    {
      "epoch": 0.6877163295131143,
      "grad_norm": 6.8125,
      "learning_rate": 1.9684255057303285e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3891950.0,
      "step": 8941
    },
    {
      "epoch": 0.6877932466733329,
      "grad_norm": 5.4375,
      "learning_rate": 1.9683101299900007e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 3892262.0,
      "step": 8942
    },
    {
      "epoch": 0.6878701638335513,
      "grad_norm": 9.0,
      "learning_rate": 1.968194754249673e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 3892596.0,
      "step": 8943
    },
    {
      "epoch": 0.6879470809937697,
      "grad_norm": 21.125,
      "learning_rate": 1.9680793785093453e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 3892916.0,
      "step": 8944
    },
    {
      "epoch": 0.6880239981539882,
      "grad_norm": 8.3125,
      "learning_rate": 1.967964002769018e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3893224.0,
      "step": 8945
    },
    {
      "epoch": 0.6881009153142066,
      "grad_norm": 15.9375,
      "learning_rate": 1.96784862702869e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 3893530.0,
      "step": 8946
    },
    {
      "epoch": 0.6881778324744251,
      "grad_norm": 6.84375,
      "learning_rate": 1.9677332512883625e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 3894088.0,
      "step": 8947
    },
    {
      "epoch": 0.6882547496346435,
      "grad_norm": 10.3125,
      "learning_rate": 1.967617875548035e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 3894426.0,
      "step": 8948
    },
    {
      "epoch": 0.6883316667948619,
      "grad_norm": 22.125,
      "learning_rate": 1.967502499807707e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3894748.0,
      "step": 8949
    },
    {
      "epoch": 0.6884085839550804,
      "grad_norm": 32.25,
      "learning_rate": 1.9673871240673792e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3895070.0,
      "step": 8950
    },
    {
      "epoch": 0.6884855011152988,
      "grad_norm": 16.125,
      "learning_rate": 1.967271748327052e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.997032642364502,
      "num_tokens": 3895408.0,
      "step": 8951
    },
    {
      "epoch": 0.6885624182755172,
      "grad_norm": 7.0625,
      "learning_rate": 1.967156372586724e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 3895694.0,
      "step": 8952
    },
    {
      "epoch": 0.6886393354357357,
      "grad_norm": 32.25,
      "learning_rate": 1.9670409968463964e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3895976.0,
      "step": 8953
    },
    {
      "epoch": 0.6887162525959541,
      "grad_norm": 68.0,
      "learning_rate": 1.9669256211060687e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3896292.0,
      "step": 8954
    },
    {
      "epoch": 0.6887931697561726,
      "grad_norm": 3.25,
      "learning_rate": 1.966810245365741e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3896834.0,
      "step": 8955
    },
    {
      "epoch": 0.6888700869163911,
      "grad_norm": 5.03125,
      "learning_rate": 1.9666948696254136e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 3897632.0,
      "step": 8956
    },
    {
      "epoch": 0.6889470040766095,
      "grad_norm": 14.125,
      "learning_rate": 1.966579493885086e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3897964.0,
      "step": 8957
    },
    {
      "epoch": 0.689023921236828,
      "grad_norm": 7.21875,
      "learning_rate": 1.966464118144758e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3898224.0,
      "step": 8958
    },
    {
      "epoch": 0.6891008383970464,
      "grad_norm": 11.9375,
      "learning_rate": 1.966348742404431e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 3898504.0,
      "step": 8959
    },
    {
      "epoch": 0.6891777555572648,
      "grad_norm": 2.796875,
      "learning_rate": 1.9662333666641027e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3899052.0,
      "step": 8960
    },
    {
      "epoch": 0.6892546727174833,
      "grad_norm": 17.5,
      "learning_rate": 1.966117990923775e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3899374.0,
      "step": 8961
    },
    {
      "epoch": 0.6893315898777017,
      "grad_norm": 11.0,
      "learning_rate": 1.9660026151834476e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3899696.0,
      "step": 8962
    },
    {
      "epoch": 0.6894085070379201,
      "grad_norm": 5.5625,
      "learning_rate": 1.96588723944312e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 3900272.0,
      "step": 8963
    },
    {
      "epoch": 0.6894854241981386,
      "grad_norm": 6.875,
      "learning_rate": 1.965771863702792e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3900580.0,
      "step": 8964
    },
    {
      "epoch": 0.689562341358357,
      "grad_norm": 17.125,
      "learning_rate": 1.9656564879624644e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 3901154.0,
      "step": 8965
    },
    {
      "epoch": 0.6896392585185755,
      "grad_norm": 2.0625,
      "learning_rate": 1.9655411122221366e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 3901696.0,
      "step": 8966
    },
    {
      "epoch": 0.689716175678794,
      "grad_norm": 9.75,
      "learning_rate": 1.9654257364818093e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3901994.0,
      "step": 8967
    },
    {
      "epoch": 0.6897930928390124,
      "grad_norm": 69.5,
      "learning_rate": 1.9653103607414816e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 3902326.0,
      "step": 8968
    },
    {
      "epoch": 0.6898700099992309,
      "grad_norm": 20.25,
      "learning_rate": 1.965194985001154e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 3902870.0,
      "step": 8969
    },
    {
      "epoch": 0.6899469271594493,
      "grad_norm": 3.8125,
      "learning_rate": 1.9650796092608265e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 3903192.0,
      "step": 8970
    },
    {
      "epoch": 0.6900238443196677,
      "grad_norm": 3.375,
      "learning_rate": 1.9649642335204983e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9963833689689636,
      "num_tokens": 3903746.0,
      "step": 8971
    },
    {
      "epoch": 0.6901007614798862,
      "grad_norm": 25.75,
      "learning_rate": 1.9648488577801706e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3904048.0,
      "step": 8972
    },
    {
      "epoch": 0.6901776786401046,
      "grad_norm": 34.0,
      "learning_rate": 1.9647334820398433e-06,
      "loss": 0.0422,
      "mean_token_accuracy": 0.9825327396392822,
      "num_tokens": 3904278.0,
      "step": 8973
    },
    {
      "epoch": 0.690254595800323,
      "grad_norm": 10.0,
      "learning_rate": 1.9646181062995155e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3904826.0,
      "step": 8974
    },
    {
      "epoch": 0.6903315129605415,
      "grad_norm": 6.75,
      "learning_rate": 1.964502730559188e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3905344.0,
      "step": 8975
    },
    {
      "epoch": 0.6904084301207599,
      "grad_norm": 22.875,
      "learning_rate": 1.96438735481886e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 3905654.0,
      "step": 8976
    },
    {
      "epoch": 0.6904853472809784,
      "grad_norm": 14.5,
      "learning_rate": 1.9642719790785323e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9825072884559631,
      "num_tokens": 3905998.0,
      "step": 8977
    },
    {
      "epoch": 0.6905622644411968,
      "grad_norm": 7.75,
      "learning_rate": 1.964156603338205e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 3906346.0,
      "step": 8978
    },
    {
      "epoch": 0.6906391816014152,
      "grad_norm": 38.0,
      "learning_rate": 1.9640412275978773e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 3906674.0,
      "step": 8979
    },
    {
      "epoch": 0.6907160987616338,
      "grad_norm": 4.46875,
      "learning_rate": 1.9639258518575495e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 3907198.0,
      "step": 8980
    },
    {
      "epoch": 0.6907930159218522,
      "grad_norm": 25.625,
      "learning_rate": 1.9638104761172218e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9889094233512878,
      "num_tokens": 3907740.0,
      "step": 8981
    },
    {
      "epoch": 0.6908699330820706,
      "grad_norm": 34.5,
      "learning_rate": 1.963695100376894e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9937499761581421,
      "num_tokens": 3908064.0,
      "step": 8982
    },
    {
      "epoch": 0.6909468502422891,
      "grad_norm": 15.25,
      "learning_rate": 1.9635797246365663e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3908362.0,
      "step": 8983
    },
    {
      "epoch": 0.6910237674025075,
      "grad_norm": 5.0625,
      "learning_rate": 1.963464348896239e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9961758852005005,
      "num_tokens": 3908886.0,
      "step": 8984
    },
    {
      "epoch": 0.691100684562726,
      "grad_norm": 11.25,
      "learning_rate": 1.9633489731559112e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3909140.0,
      "step": 8985
    },
    {
      "epoch": 0.6911776017229444,
      "grad_norm": 4.84375,
      "learning_rate": 1.9632335974155835e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9967741966247559,
      "num_tokens": 3909452.0,
      "step": 8986
    },
    {
      "epoch": 0.6912545188831628,
      "grad_norm": 4.21875,
      "learning_rate": 1.9631182216752557e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 3910234.0,
      "step": 8987
    },
    {
      "epoch": 0.6913314360433813,
      "grad_norm": 3.453125,
      "learning_rate": 1.963002845934928e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 3910780.0,
      "step": 8988
    },
    {
      "epoch": 0.6914083532035997,
      "grad_norm": 14.0,
      "learning_rate": 1.9628874701946003e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 3911082.0,
      "step": 8989
    },
    {
      "epoch": 0.6914852703638181,
      "grad_norm": 23.375,
      "learning_rate": 1.962772094454273e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 3911420.0,
      "step": 8990
    },
    {
      "epoch": 0.6915621875240366,
      "grad_norm": 20.125,
      "learning_rate": 1.962656718713945e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 3911980.0,
      "step": 8991
    },
    {
      "epoch": 0.691639104684255,
      "grad_norm": 16.875,
      "learning_rate": 1.9625413429736175e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3912554.0,
      "step": 8992
    },
    {
      "epoch": 0.6917160218444734,
      "grad_norm": 7.8125,
      "learning_rate": 1.9624259672332897e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3913134.0,
      "step": 8993
    },
    {
      "epoch": 0.691792939004692,
      "grad_norm": 27.5,
      "learning_rate": 1.962310591492962e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 3913394.0,
      "step": 8994
    },
    {
      "epoch": 0.6918698561649104,
      "grad_norm": 4.28125,
      "learning_rate": 1.9621952157526347e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9949748516082764,
      "num_tokens": 3914192.0,
      "step": 8995
    },
    {
      "epoch": 0.6919467733251289,
      "grad_norm": 10.875,
      "learning_rate": 1.962079840012307e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 3914494.0,
      "step": 8996
    },
    {
      "epoch": 0.6920236904853473,
      "grad_norm": 2.75,
      "learning_rate": 1.961964464271979e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 3915278.0,
      "step": 8997
    },
    {
      "epoch": 0.6921006076455657,
      "grad_norm": 2.828125,
      "learning_rate": 1.9618490885316514e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 3915864.0,
      "step": 8998
    },
    {
      "epoch": 0.6921775248057842,
      "grad_norm": 5.75,
      "learning_rate": 1.9617337127913237e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 3916444.0,
      "step": 8999
    },
    {
      "epoch": 0.6922544419660026,
      "grad_norm": 11.625,
      "learning_rate": 1.961618337050996e-06,
      "loss": 0.0171,
      "step": 9000
    },
    {
      "epoch": 0.6922544419660026,
      "eval_loss": 0.02149086631834507,
      "eval_mean_token_accuracy": 0.991448097229004,
      "eval_num_tokens": 3916962.0,
      "eval_runtime": 21.0836,
      "eval_samples_per_second": 9.486,
      "eval_steps_per_second": 1.186,
      "step": 9000
    },
    {
      "epoch": 0.692331359126221,
      "grad_norm": 20.125,
      "learning_rate": 1.9615029613106686e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9943260252475739,
      "num_tokens": 3917504.0,
      "step": 9001
    },
    {
      "epoch": 0.6924082762864395,
      "grad_norm": 11.75,
      "learning_rate": 1.961387585570341e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3918074.0,
      "step": 9002
    },
    {
      "epoch": 0.6924851934466579,
      "grad_norm": 25.625,
      "learning_rate": 1.961272209830013e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3918332.0,
      "step": 9003
    },
    {
      "epoch": 0.6925621106068764,
      "grad_norm": 3.0625,
      "learning_rate": 1.9611568340896854e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 3918890.0,
      "step": 9004
    },
    {
      "epoch": 0.6926390277670949,
      "grad_norm": 8.125,
      "learning_rate": 1.9610414583493577e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 3919406.0,
      "step": 9005
    },
    {
      "epoch": 0.6927159449273133,
      "grad_norm": 3.578125,
      "learning_rate": 1.9609260826090303e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3919952.0,
      "step": 9006
    },
    {
      "epoch": 0.6927928620875318,
      "grad_norm": 24.25,
      "learning_rate": 1.9608107068687026e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3920292.0,
      "step": 9007
    },
    {
      "epoch": 0.6928697792477502,
      "grad_norm": 33.25,
      "learning_rate": 1.9606953311283744e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3920570.0,
      "step": 9008
    },
    {
      "epoch": 0.6929466964079686,
      "grad_norm": 3.09375,
      "learning_rate": 1.960579955388047e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 3921116.0,
      "step": 9009
    },
    {
      "epoch": 0.6930236135681871,
      "grad_norm": 14.125,
      "learning_rate": 1.9604645796477194e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 3921692.0,
      "step": 9010
    },
    {
      "epoch": 0.6931005307284055,
      "grad_norm": 21.625,
      "learning_rate": 1.9603492039073916e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 3922270.0,
      "step": 9011
    },
    {
      "epoch": 0.6931774478886239,
      "grad_norm": 16.375,
      "learning_rate": 1.9602338281670643e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3922572.0,
      "step": 9012
    },
    {
      "epoch": 0.6932543650488424,
      "grad_norm": 27.75,
      "learning_rate": 1.9601184524267366e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3923136.0,
      "step": 9013
    },
    {
      "epoch": 0.6933312822090608,
      "grad_norm": 7.90625,
      "learning_rate": 1.960003076686409e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9965635538101196,
      "num_tokens": 3923428.0,
      "step": 9014
    },
    {
      "epoch": 0.6934081993692793,
      "grad_norm": 4.03125,
      "learning_rate": 1.959887700946081e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3924186.0,
      "step": 9015
    },
    {
      "epoch": 0.6934851165294977,
      "grad_norm": 1.6875,
      "learning_rate": 1.9597723252057534e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3924704.0,
      "step": 9016
    },
    {
      "epoch": 0.6935620336897161,
      "grad_norm": 3.546875,
      "learning_rate": 1.959656949465426e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 3925198.0,
      "step": 9017
    },
    {
      "epoch": 0.6936389508499347,
      "grad_norm": 17.5,
      "learning_rate": 1.9595415737250983e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3925480.0,
      "step": 9018
    },
    {
      "epoch": 0.6937158680101531,
      "grad_norm": 3.90625,
      "learning_rate": 1.95942619798477e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3925788.0,
      "step": 9019
    },
    {
      "epoch": 0.6937927851703715,
      "grad_norm": 26.75,
      "learning_rate": 1.959310822244443e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 3926128.0,
      "step": 9020
    },
    {
      "epoch": 0.69386970233059,
      "grad_norm": 6.03125,
      "learning_rate": 1.959195446504115e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 3926464.0,
      "step": 9021
    },
    {
      "epoch": 0.6939466194908084,
      "grad_norm": 4.53125,
      "learning_rate": 1.9590800707637873e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3926748.0,
      "step": 9022
    },
    {
      "epoch": 0.6940235366510269,
      "grad_norm": 16.75,
      "learning_rate": 1.95896469502346e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9915110468864441,
      "num_tokens": 3927338.0,
      "step": 9023
    },
    {
      "epoch": 0.6941004538112453,
      "grad_norm": 7.5625,
      "learning_rate": 1.9588493192831323e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 3927626.0,
      "step": 9024
    },
    {
      "epoch": 0.6941773709714637,
      "grad_norm": 11.5625,
      "learning_rate": 1.9587339435428045e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 3928210.0,
      "step": 9025
    },
    {
      "epoch": 0.6942542881316822,
      "grad_norm": 11.8125,
      "learning_rate": 1.9586185678024768e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9936224222183228,
      "num_tokens": 3928996.0,
      "step": 9026
    },
    {
      "epoch": 0.6943312052919006,
      "grad_norm": 10.3125,
      "learning_rate": 1.958503192062149e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3929318.0,
      "step": 9027
    },
    {
      "epoch": 0.694408122452119,
      "grad_norm": 28.875,
      "learning_rate": 1.9583878163218217e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3929666.0,
      "step": 9028
    },
    {
      "epoch": 0.6944850396123375,
      "grad_norm": 22.125,
      "learning_rate": 1.958272440581494e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 3929986.0,
      "step": 9029
    },
    {
      "epoch": 0.6945619567725559,
      "grad_norm": 17.625,
      "learning_rate": 1.958157064841166e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3930286.0,
      "step": 9030
    },
    {
      "epoch": 0.6946388739327743,
      "grad_norm": 39.75,
      "learning_rate": 1.9580416891008385e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 3930580.0,
      "step": 9031
    },
    {
      "epoch": 0.6947157910929929,
      "grad_norm": 37.25,
      "learning_rate": 1.9579263133605108e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 3930912.0,
      "step": 9032
    },
    {
      "epoch": 0.6947927082532113,
      "grad_norm": 16.25,
      "learning_rate": 1.957810937620183e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3931198.0,
      "step": 9033
    },
    {
      "epoch": 0.6948696254134298,
      "grad_norm": 39.0,
      "learning_rate": 1.9576955618798557e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9842271208763123,
      "num_tokens": 3931516.0,
      "step": 9034
    },
    {
      "epoch": 0.6949465425736482,
      "grad_norm": 4.1875,
      "learning_rate": 1.957580186139528e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3932038.0,
      "step": 9035
    },
    {
      "epoch": 0.6950234597338666,
      "grad_norm": 4.40625,
      "learning_rate": 1.9574648103992e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3932586.0,
      "step": 9036
    },
    {
      "epoch": 0.6951003768940851,
      "grad_norm": 4.59375,
      "learning_rate": 1.9573494346588725e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9948365092277527,
      "num_tokens": 3933168.0,
      "step": 9037
    },
    {
      "epoch": 0.6951772940543035,
      "grad_norm": 15.5,
      "learning_rate": 1.9572340589185447e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 3933428.0,
      "step": 9038
    },
    {
      "epoch": 0.6952542112145219,
      "grad_norm": 11.5625,
      "learning_rate": 1.957118683178217e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 3933738.0,
      "step": 9039
    },
    {
      "epoch": 0.6953311283747404,
      "grad_norm": 22.25,
      "learning_rate": 1.9570033074378897e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3934016.0,
      "step": 9040
    },
    {
      "epoch": 0.6954080455349588,
      "grad_norm": 4.625,
      "learning_rate": 1.9568879316975615e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3934588.0,
      "step": 9041
    },
    {
      "epoch": 0.6954849626951773,
      "grad_norm": 9.1875,
      "learning_rate": 1.956772555957234e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9966158866882324,
      "num_tokens": 3935180.0,
      "step": 9042
    },
    {
      "epoch": 0.6955618798553957,
      "grad_norm": 18.625,
      "learning_rate": 1.9566571802169064e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 3935546.0,
      "step": 9043
    },
    {
      "epoch": 0.6956387970156142,
      "grad_norm": 16.75,
      "learning_rate": 1.9565418044765787e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3935854.0,
      "step": 9044
    },
    {
      "epoch": 0.6957157141758327,
      "grad_norm": 9.9375,
      "learning_rate": 1.9564264287362514e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 3936156.0,
      "step": 9045
    },
    {
      "epoch": 0.6957926313360511,
      "grad_norm": 5.1875,
      "learning_rate": 1.956311052995923e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 3936690.0,
      "step": 9046
    },
    {
      "epoch": 0.6958695484962695,
      "grad_norm": 5.4375,
      "learning_rate": 1.9561956772555955e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3937246.0,
      "step": 9047
    },
    {
      "epoch": 0.695946465656488,
      "grad_norm": 12.1875,
      "learning_rate": 1.956080301515268e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3937500.0,
      "step": 9048
    },
    {
      "epoch": 0.6960233828167064,
      "grad_norm": 4.5,
      "learning_rate": 1.9559649257749404e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 3938086.0,
      "step": 9049
    },
    {
      "epoch": 0.6961002999769248,
      "grad_norm": 35.0,
      "learning_rate": 1.9558495500346127e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 3938408.0,
      "step": 9050
    },
    {
      "epoch": 0.6961772171371433,
      "grad_norm": 18.75,
      "learning_rate": 1.9557341742942853e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.990328848361969,
      "num_tokens": 3938926.0,
      "step": 9051
    },
    {
      "epoch": 0.6962541342973617,
      "grad_norm": 21.125,
      "learning_rate": 1.955618798553957e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3939472.0,
      "step": 9052
    },
    {
      "epoch": 0.6963310514575802,
      "grad_norm": 1.9375,
      "learning_rate": 1.95550342281363e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 3939996.0,
      "step": 9053
    },
    {
      "epoch": 0.6964079686177986,
      "grad_norm": 3.75,
      "learning_rate": 1.955388047073302e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 3940546.0,
      "step": 9054
    },
    {
      "epoch": 0.696484885778017,
      "grad_norm": 3.0625,
      "learning_rate": 1.9552726713329744e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 3941088.0,
      "step": 9055
    },
    {
      "epoch": 0.6965618029382356,
      "grad_norm": 4.625,
      "learning_rate": 1.955157295592647e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 3941914.0,
      "step": 9056
    },
    {
      "epoch": 0.696638720098454,
      "grad_norm": 31.875,
      "learning_rate": 1.955041919852319e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 3942274.0,
      "step": 9057
    },
    {
      "epoch": 0.6967156372586724,
      "grad_norm": 81.5,
      "learning_rate": 1.954926544111991e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.9781249761581421,
      "num_tokens": 3942598.0,
      "step": 9058
    },
    {
      "epoch": 0.6967925544188909,
      "grad_norm": 20.125,
      "learning_rate": 1.954811168371664e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9960159659385681,
      "num_tokens": 3942850.0,
      "step": 9059
    },
    {
      "epoch": 0.6968694715791093,
      "grad_norm": 4.65625,
      "learning_rate": 1.954695792631336e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 3943346.0,
      "step": 9060
    },
    {
      "epoch": 0.6969463887393278,
      "grad_norm": 5.53125,
      "learning_rate": 1.9545804168910084e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9939637780189514,
      "num_tokens": 3943844.0,
      "step": 9061
    },
    {
      "epoch": 0.6970233058995462,
      "grad_norm": 21.25,
      "learning_rate": 1.954465041150681e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 3944366.0,
      "step": 9062
    },
    {
      "epoch": 0.6971002230597646,
      "grad_norm": 2.734375,
      "learning_rate": 1.954349665410353e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9962476491928101,
      "num_tokens": 3944900.0,
      "step": 9063
    },
    {
      "epoch": 0.6971771402199831,
      "grad_norm": 17.125,
      "learning_rate": 1.9542342896700256e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 3945446.0,
      "step": 9064
    },
    {
      "epoch": 0.6972540573802015,
      "grad_norm": 15.6875,
      "learning_rate": 1.954118913929698e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3945728.0,
      "step": 9065
    },
    {
      "epoch": 0.6973309745404199,
      "grad_norm": 23.625,
      "learning_rate": 1.95400353818937e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 3946010.0,
      "step": 9066
    },
    {
      "epoch": 0.6974078917006384,
      "grad_norm": 19.25,
      "learning_rate": 1.9538881624490427e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 3946292.0,
      "step": 9067
    },
    {
      "epoch": 0.6974848088608568,
      "grad_norm": 11.8125,
      "learning_rate": 1.9537727867087146e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3946864.0,
      "step": 9068
    },
    {
      "epoch": 0.6975617260210754,
      "grad_norm": 5.6875,
      "learning_rate": 1.953657410968387e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 3947660.0,
      "step": 9069
    },
    {
      "epoch": 0.6976386431812938,
      "grad_norm": 4.46875,
      "learning_rate": 1.9535420352280595e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 3948680.0,
      "step": 9070
    },
    {
      "epoch": 0.6977155603415122,
      "grad_norm": 6.53125,
      "learning_rate": 1.9534266594877318e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3949242.0,
      "step": 9071
    },
    {
      "epoch": 0.6977924775017307,
      "grad_norm": 5.90625,
      "learning_rate": 1.953311283747404e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3949540.0,
      "step": 9072
    },
    {
      "epoch": 0.6978693946619491,
      "grad_norm": 2.984375,
      "learning_rate": 1.9531959080070767e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 3950084.0,
      "step": 9073
    },
    {
      "epoch": 0.6979463118221675,
      "grad_norm": 5.125,
      "learning_rate": 1.9530805322667486e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 3950646.0,
      "step": 9074
    },
    {
      "epoch": 0.698023228982386,
      "grad_norm": 42.75,
      "learning_rate": 1.9529651565264212e-06,
      "loss": 0.0466,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 3950900.0,
      "step": 9075
    },
    {
      "epoch": 0.6981001461426044,
      "grad_norm": 37.25,
      "learning_rate": 1.9528497807860935e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 3951222.0,
      "step": 9076
    },
    {
      "epoch": 0.6981770633028228,
      "grad_norm": 6.9375,
      "learning_rate": 1.9527344050457658e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 3951574.0,
      "step": 9077
    },
    {
      "epoch": 0.6982539804630413,
      "grad_norm": 30.875,
      "learning_rate": 1.9526190293054384e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3951882.0,
      "step": 9078
    },
    {
      "epoch": 0.6983308976232597,
      "grad_norm": 16.25,
      "learning_rate": 1.9525036535651103e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 3952114.0,
      "step": 9079
    },
    {
      "epoch": 0.6984078147834782,
      "grad_norm": 10.0,
      "learning_rate": 1.9523882778247825e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 3952404.0,
      "step": 9080
    },
    {
      "epoch": 0.6984847319436966,
      "grad_norm": 22.125,
      "learning_rate": 1.952272902084455e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 3952968.0,
      "step": 9081
    },
    {
      "epoch": 0.698561649103915,
      "grad_norm": 8.125,
      "learning_rate": 1.9521575263441275e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 3953332.0,
      "step": 9082
    },
    {
      "epoch": 0.6986385662641336,
      "grad_norm": 5.84375,
      "learning_rate": 1.9520421506037997e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 3953904.0,
      "step": 9083
    },
    {
      "epoch": 0.698715483424352,
      "grad_norm": 14.0,
      "learning_rate": 1.951926774863472e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 3954476.0,
      "step": 9084
    },
    {
      "epoch": 0.6987924005845704,
      "grad_norm": 13.875,
      "learning_rate": 1.9518113991231442e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 3954788.0,
      "step": 9085
    },
    {
      "epoch": 0.6988693177447889,
      "grad_norm": 17.0,
      "learning_rate": 1.951696023382817e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 3955096.0,
      "step": 9086
    },
    {
      "epoch": 0.6989462349050073,
      "grad_norm": 19.125,
      "learning_rate": 1.951580647642489e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.992094874382019,
      "num_tokens": 3955350.0,
      "step": 9087
    },
    {
      "epoch": 0.6990231520652258,
      "grad_norm": 8.5,
      "learning_rate": 1.9514652719021614e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3955636.0,
      "step": 9088
    },
    {
      "epoch": 0.6991000692254442,
      "grad_norm": 3.421875,
      "learning_rate": 1.9513498961618337e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9959431886672974,
      "num_tokens": 3956130.0,
      "step": 9089
    },
    {
      "epoch": 0.6991769863856626,
      "grad_norm": 35.0,
      "learning_rate": 1.951234520421506e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3956426.0,
      "step": 9090
    },
    {
      "epoch": 0.6992539035458811,
      "grad_norm": 7.21875,
      "learning_rate": 1.9511191446811782e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 3956756.0,
      "step": 9091
    },
    {
      "epoch": 0.6993308207060995,
      "grad_norm": 5.46875,
      "learning_rate": 1.951003768940851e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3957040.0,
      "step": 9092
    },
    {
      "epoch": 0.6994077378663179,
      "grad_norm": 26.125,
      "learning_rate": 1.950888393200523e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 3957418.0,
      "step": 9093
    },
    {
      "epoch": 0.6994846550265365,
      "grad_norm": 26.25,
      "learning_rate": 1.9507730174601954e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 3957752.0,
      "step": 9094
    },
    {
      "epoch": 0.6995615721867549,
      "grad_norm": 21.625,
      "learning_rate": 1.9506576417198677e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 3958272.0,
      "step": 9095
    },
    {
      "epoch": 0.6996384893469733,
      "grad_norm": 14.25,
      "learning_rate": 1.95054226597954e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 3958868.0,
      "step": 9096
    },
    {
      "epoch": 0.6997154065071918,
      "grad_norm": 14.5,
      "learning_rate": 1.950426890239212e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9914236664772034,
      "num_tokens": 3959452.0,
      "step": 9097
    },
    {
      "epoch": 0.6997923236674102,
      "grad_norm": 28.75,
      "learning_rate": 1.950311514498885e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 3959796.0,
      "step": 9098
    },
    {
      "epoch": 0.6998692408276287,
      "grad_norm": 29.625,
      "learning_rate": 1.950196138758557e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 3960152.0,
      "step": 9099
    },
    {
      "epoch": 0.6999461579878471,
      "grad_norm": 29.875,
      "learning_rate": 1.9500807630182294e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 3960484.0,
      "step": 9100
    },
    {
      "epoch": 0.7000230751480655,
      "grad_norm": 2.40625,
      "learning_rate": 1.9499653872779016e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3961002.0,
      "step": 9101
    },
    {
      "epoch": 0.700099992308284,
      "grad_norm": 30.5,
      "learning_rate": 1.949850011537574e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3961338.0,
      "step": 9102
    },
    {
      "epoch": 0.7001769094685024,
      "grad_norm": 9.5625,
      "learning_rate": 1.9497346357972466e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 3961662.0,
      "step": 9103
    },
    {
      "epoch": 0.7002538266287208,
      "grad_norm": 8.0625,
      "learning_rate": 1.949619260056919e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3962008.0,
      "step": 9104
    },
    {
      "epoch": 0.7003307437889393,
      "grad_norm": 4.59375,
      "learning_rate": 1.949503884316591e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 3962296.0,
      "step": 9105
    },
    {
      "epoch": 0.7004076609491577,
      "grad_norm": 11.0,
      "learning_rate": 1.9493885085762634e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 3962578.0,
      "step": 9106
    },
    {
      "epoch": 0.7004845781093763,
      "grad_norm": 6.65625,
      "learning_rate": 1.9492731328359356e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 3962924.0,
      "step": 9107
    },
    {
      "epoch": 0.7005614952695947,
      "grad_norm": 5.46875,
      "learning_rate": 1.949157757095608e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 3963460.0,
      "step": 9108
    },
    {
      "epoch": 0.7006384124298131,
      "grad_norm": 28.625,
      "learning_rate": 1.9490423813552806e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9842271208763123,
      "num_tokens": 3963778.0,
      "step": 9109
    },
    {
      "epoch": 0.7007153295900316,
      "grad_norm": 2.140625,
      "learning_rate": 1.948927005614953e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 3964326.0,
      "step": 9110
    },
    {
      "epoch": 0.70079224675025,
      "grad_norm": 65.5,
      "learning_rate": 1.948811629874625e-06,
      "loss": 0.0461,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 3964630.0,
      "step": 9111
    },
    {
      "epoch": 0.7008691639104684,
      "grad_norm": 12.8125,
      "learning_rate": 1.9486962541342973e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9930070042610168,
      "num_tokens": 3964918.0,
      "step": 9112
    },
    {
      "epoch": 0.7009460810706869,
      "grad_norm": 18.25,
      "learning_rate": 1.9485808783939696e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 3965464.0,
      "step": 9113
    },
    {
      "epoch": 0.7010229982309053,
      "grad_norm": 38.75,
      "learning_rate": 1.9484655026536423e-06,
      "loss": 0.0399,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 3965744.0,
      "step": 9114
    },
    {
      "epoch": 0.7010999153911237,
      "grad_norm": 3.671875,
      "learning_rate": 1.9483501269133145e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 3966068.0,
      "step": 9115
    },
    {
      "epoch": 0.7011768325513422,
      "grad_norm": 22.125,
      "learning_rate": 1.9482347511729868e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 3966418.0,
      "step": 9116
    },
    {
      "epoch": 0.7012537497115606,
      "grad_norm": 7.25,
      "learning_rate": 1.948119375432659e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9895833134651184,
      "num_tokens": 3966708.0,
      "step": 9117
    },
    {
      "epoch": 0.7013306668717791,
      "grad_norm": 10.625,
      "learning_rate": 1.9480039996923313e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 3967050.0,
      "step": 9118
    },
    {
      "epoch": 0.7014075840319975,
      "grad_norm": 4.59375,
      "learning_rate": 1.9478886239520036e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 3967630.0,
      "step": 9119
    },
    {
      "epoch": 0.701484501192216,
      "grad_norm": 3.890625,
      "learning_rate": 1.9477732482116762e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 3968200.0,
      "step": 9120
    },
    {
      "epoch": 0.7015614183524345,
      "grad_norm": 19.625,
      "learning_rate": 1.9476578724713485e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3968484.0,
      "step": 9121
    },
    {
      "epoch": 0.7016383355126529,
      "grad_norm": 15.75,
      "learning_rate": 1.9475424967310208e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3968792.0,
      "step": 9122
    },
    {
      "epoch": 0.7017152526728713,
      "grad_norm": 25.625,
      "learning_rate": 1.947427120990693e-06,
      "loss": 0.0398,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 3969050.0,
      "step": 9123
    },
    {
      "epoch": 0.7017921698330898,
      "grad_norm": 49.0,
      "learning_rate": 1.9473117452503653e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 3969364.0,
      "step": 9124
    },
    {
      "epoch": 0.7018690869933082,
      "grad_norm": 4.96875,
      "learning_rate": 1.947196369510038e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3969954.0,
      "step": 9125
    },
    {
      "epoch": 0.7019460041535267,
      "grad_norm": 21.625,
      "learning_rate": 1.9470809937697102e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3970280.0,
      "step": 9126
    },
    {
      "epoch": 0.7020229213137451,
      "grad_norm": 5.90625,
      "learning_rate": 1.9469656180293825e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3970854.0,
      "step": 9127
    },
    {
      "epoch": 0.7020998384739635,
      "grad_norm": 24.25,
      "learning_rate": 1.9468502422890547e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 3971162.0,
      "step": 9128
    },
    {
      "epoch": 0.702176755634182,
      "grad_norm": 17.875,
      "learning_rate": 1.946734866548727e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 3971488.0,
      "step": 9129
    },
    {
      "epoch": 0.7022536727944004,
      "grad_norm": 4.625,
      "learning_rate": 1.9466194908083992e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3971838.0,
      "step": 9130
    },
    {
      "epoch": 0.7023305899546188,
      "grad_norm": 4.625,
      "learning_rate": 1.946504115068072e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 3972652.0,
      "step": 9131
    },
    {
      "epoch": 0.7024075071148373,
      "grad_norm": 26.25,
      "learning_rate": 1.946388739327744e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 3972988.0,
      "step": 9132
    },
    {
      "epoch": 0.7024844242750558,
      "grad_norm": 19.0,
      "learning_rate": 1.9462733635874164e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 3973352.0,
      "step": 9133
    },
    {
      "epoch": 0.7025613414352742,
      "grad_norm": 9.375,
      "learning_rate": 1.9461579878470887e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9885057210922241,
      "num_tokens": 3973614.0,
      "step": 9134
    },
    {
      "epoch": 0.7026382585954927,
      "grad_norm": 7.40625,
      "learning_rate": 1.946042612106761e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 3973912.0,
      "step": 9135
    },
    {
      "epoch": 0.7027151757557111,
      "grad_norm": 4.46875,
      "learning_rate": 1.9459272363664336e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3974234.0,
      "step": 9136
    },
    {
      "epoch": 0.7027920929159296,
      "grad_norm": 29.125,
      "learning_rate": 1.945811860626106e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 3974540.0,
      "step": 9137
    },
    {
      "epoch": 0.702869010076148,
      "grad_norm": 25.75,
      "learning_rate": 1.945696484885778e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 3974876.0,
      "step": 9138
    },
    {
      "epoch": 0.7029459272363664,
      "grad_norm": 32.0,
      "learning_rate": 1.9455811091454504e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 3975160.0,
      "step": 9139
    },
    {
      "epoch": 0.7030228443965849,
      "grad_norm": 20.0,
      "learning_rate": 1.9454657334051227e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 3975440.0,
      "step": 9140
    },
    {
      "epoch": 0.7030997615568033,
      "grad_norm": 3.84375,
      "learning_rate": 1.945350357664795e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 3975724.0,
      "step": 9141
    },
    {
      "epoch": 0.7031766787170217,
      "grad_norm": 13.6875,
      "learning_rate": 1.9452349819244676e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 3976256.0,
      "step": 9142
    },
    {
      "epoch": 0.7032535958772402,
      "grad_norm": 2.78125,
      "learning_rate": 1.94511960618414e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 3976828.0,
      "step": 9143
    },
    {
      "epoch": 0.7033305130374586,
      "grad_norm": 14.0,
      "learning_rate": 1.945004230443812e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 3977396.0,
      "step": 9144
    },
    {
      "epoch": 0.7034074301976772,
      "grad_norm": 3.8125,
      "learning_rate": 1.9448888547034844e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 3977936.0,
      "step": 9145
    },
    {
      "epoch": 0.7034843473578956,
      "grad_norm": 45.25,
      "learning_rate": 1.9447734789631566e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 3978286.0,
      "step": 9146
    },
    {
      "epoch": 0.703561264518114,
      "grad_norm": 31.875,
      "learning_rate": 1.944658103222829e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 3978652.0,
      "step": 9147
    },
    {
      "epoch": 0.7036381816783325,
      "grad_norm": 5.9375,
      "learning_rate": 1.9445427274825016e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 3978966.0,
      "step": 9148
    },
    {
      "epoch": 0.7037150988385509,
      "grad_norm": 9.625,
      "learning_rate": 1.9444273517421734e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 3979530.0,
      "step": 9149
    },
    {
      "epoch": 0.7037920159987693,
      "grad_norm": 16.75,
      "learning_rate": 1.944311976001846e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 3979800.0,
      "step": 9150
    },
    {
      "epoch": 0.7038689331589878,
      "grad_norm": 11.75,
      "learning_rate": 1.9441966002615184e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3980108.0,
      "step": 9151
    },
    {
      "epoch": 0.7039458503192062,
      "grad_norm": 17.25,
      "learning_rate": 1.9440812245211906e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9910233616828918,
      "num_tokens": 3980666.0,
      "step": 9152
    },
    {
      "epoch": 0.7040227674794246,
      "grad_norm": 15.0,
      "learning_rate": 1.9439658487808633e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9968253970146179,
      "num_tokens": 3980982.0,
      "step": 9153
    },
    {
      "epoch": 0.7040996846396431,
      "grad_norm": 21.25,
      "learning_rate": 1.9438504730405356e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 3981264.0,
      "step": 9154
    },
    {
      "epoch": 0.7041766017998615,
      "grad_norm": 25.25,
      "learning_rate": 1.9437350973002074e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 3981604.0,
      "step": 9155
    },
    {
      "epoch": 0.70425351896008,
      "grad_norm": 11.25,
      "learning_rate": 1.94361972155988e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 3982172.0,
      "step": 9156
    },
    {
      "epoch": 0.7043304361202984,
      "grad_norm": 12.875,
      "learning_rate": 1.9435043458195523e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3982516.0,
      "step": 9157
    },
    {
      "epoch": 0.7044073532805168,
      "grad_norm": 8.875,
      "learning_rate": 1.9433889700792246e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 3982802.0,
      "step": 9158
    },
    {
      "epoch": 0.7044842704407354,
      "grad_norm": 21.75,
      "learning_rate": 1.9432735943388973e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 3983080.0,
      "step": 9159
    },
    {
      "epoch": 0.7045611876009538,
      "grad_norm": 5.0625,
      "learning_rate": 1.943158218598569e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 3983374.0,
      "step": 9160
    },
    {
      "epoch": 0.7046381047611722,
      "grad_norm": 37.25,
      "learning_rate": 1.9430428428582418e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 3983656.0,
      "step": 9161
    },
    {
      "epoch": 0.7047150219213907,
      "grad_norm": 3.75,
      "learning_rate": 1.942927467117914e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 3984176.0,
      "step": 9162
    },
    {
      "epoch": 0.7047919390816091,
      "grad_norm": 21.25,
      "learning_rate": 1.9428120913775863e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 3984454.0,
      "step": 9163
    },
    {
      "epoch": 0.7048688562418276,
      "grad_norm": 17.75,
      "learning_rate": 1.942696715637259e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 3985038.0,
      "step": 9164
    },
    {
      "epoch": 0.704945773402046,
      "grad_norm": 16.25,
      "learning_rate": 1.9425813398969312e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 3985612.0,
      "step": 9165
    },
    {
      "epoch": 0.7050226905622644,
      "grad_norm": 15.0625,
      "learning_rate": 1.942465964156603e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3985912.0,
      "step": 9166
    },
    {
      "epoch": 0.7050996077224829,
      "grad_norm": 26.375,
      "learning_rate": 1.9423505884162758e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 3986256.0,
      "step": 9167
    },
    {
      "epoch": 0.7051765248827013,
      "grad_norm": 33.5,
      "learning_rate": 1.942235212675948e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 3986536.0,
      "step": 9168
    },
    {
      "epoch": 0.7052534420429197,
      "grad_norm": 29.375,
      "learning_rate": 1.9421198369356203e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 3986834.0,
      "step": 9169
    },
    {
      "epoch": 0.7053303592031382,
      "grad_norm": 47.5,
      "learning_rate": 1.942004461195293e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9808306694030762,
      "num_tokens": 3987148.0,
      "step": 9170
    },
    {
      "epoch": 0.7054072763633567,
      "grad_norm": 3.84375,
      "learning_rate": 1.941889085454965e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 3987688.0,
      "step": 9171
    },
    {
      "epoch": 0.7054841935235752,
      "grad_norm": 23.75,
      "learning_rate": 1.9417737097146375e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 3988046.0,
      "step": 9172
    },
    {
      "epoch": 0.7055611106837936,
      "grad_norm": 8.4375,
      "learning_rate": 1.9416583339743097e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 3988580.0,
      "step": 9173
    },
    {
      "epoch": 0.705638027844012,
      "grad_norm": 5.6875,
      "learning_rate": 1.941542958233982e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 3988902.0,
      "step": 9174
    },
    {
      "epoch": 0.7057149450042305,
      "grad_norm": 6.40625,
      "learning_rate": 1.9414275824936547e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 3989712.0,
      "step": 9175
    },
    {
      "epoch": 0.7057918621644489,
      "grad_norm": 8.375,
      "learning_rate": 1.941312206753327e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 3990230.0,
      "step": 9176
    },
    {
      "epoch": 0.7058687793246673,
      "grad_norm": 6.625,
      "learning_rate": 1.9411968310129988e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 3990820.0,
      "step": 9177
    },
    {
      "epoch": 0.7059456964848858,
      "grad_norm": 8.6875,
      "learning_rate": 1.9410814552726714e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 3991376.0,
      "step": 9178
    },
    {
      "epoch": 0.7060226136451042,
      "grad_norm": 11.6875,
      "learning_rate": 1.9409660795323437e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 3991726.0,
      "step": 9179
    },
    {
      "epoch": 0.7060995308053226,
      "grad_norm": 36.0,
      "learning_rate": 1.940850703792016e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 3992028.0,
      "step": 9180
    },
    {
      "epoch": 0.7061764479655411,
      "grad_norm": 12.0625,
      "learning_rate": 1.9407353280516886e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 3992352.0,
      "step": 9181
    },
    {
      "epoch": 0.7062533651257595,
      "grad_norm": 29.875,
      "learning_rate": 1.9406199523113605e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 3992668.0,
      "step": 9182
    },
    {
      "epoch": 0.706330282285978,
      "grad_norm": 17.625,
      "learning_rate": 1.940504576571033e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 3992984.0,
      "step": 9183
    },
    {
      "epoch": 0.7064071994461965,
      "grad_norm": 15.5,
      "learning_rate": 1.9403892008307054e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 3993318.0,
      "step": 9184
    },
    {
      "epoch": 0.7064841166064149,
      "grad_norm": 18.75,
      "learning_rate": 1.9402738250903777e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 3993614.0,
      "step": 9185
    },
    {
      "epoch": 0.7065610337666334,
      "grad_norm": 6.0625,
      "learning_rate": 1.9401584493500504e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 3994108.0,
      "step": 9186
    },
    {
      "epoch": 0.7066379509268518,
      "grad_norm": 18.0,
      "learning_rate": 1.940043073609722e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 3994400.0,
      "step": 9187
    },
    {
      "epoch": 0.7067148680870702,
      "grad_norm": 23.75,
      "learning_rate": 1.9399276978693944e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 3994744.0,
      "step": 9188
    },
    {
      "epoch": 0.7067917852472887,
      "grad_norm": 21.75,
      "learning_rate": 1.939812322129067e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 1.0,
      "num_tokens": 3995076.0,
      "step": 9189
    },
    {
      "epoch": 0.7068687024075071,
      "grad_norm": 13.4375,
      "learning_rate": 1.9396969463887394e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 3995648.0,
      "step": 9190
    },
    {
      "epoch": 0.7069456195677256,
      "grad_norm": 25.625,
      "learning_rate": 1.9395815706484116e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 3995904.0,
      "step": 9191
    },
    {
      "epoch": 0.707022536727944,
      "grad_norm": 12.875,
      "learning_rate": 1.9394661949080843e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 3996474.0,
      "step": 9192
    },
    {
      "epoch": 0.7070994538881624,
      "grad_norm": 6.15625,
      "learning_rate": 1.939350819167756e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 3996782.0,
      "step": 9193
    },
    {
      "epoch": 0.7071763710483809,
      "grad_norm": 16.5,
      "learning_rate": 1.939235443427429e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 3997334.0,
      "step": 9194
    },
    {
      "epoch": 0.7072532882085993,
      "grad_norm": 47.5,
      "learning_rate": 1.939120067687101e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9918699264526367,
      "num_tokens": 3997706.0,
      "step": 9195
    },
    {
      "epoch": 0.7073302053688177,
      "grad_norm": 13.4375,
      "learning_rate": 1.9390046919467734e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 3998256.0,
      "step": 9196
    },
    {
      "epoch": 0.7074071225290363,
      "grad_norm": 21.875,
      "learning_rate": 1.9388893162064456e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 3998574.0,
      "step": 9197
    },
    {
      "epoch": 0.7074840396892547,
      "grad_norm": 5.75,
      "learning_rate": 1.938773940466118e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 3999138.0,
      "step": 9198
    },
    {
      "epoch": 0.7075609568494731,
      "grad_norm": 8.4375,
      "learning_rate": 1.93865856472579e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 3999686.0,
      "step": 9199
    },
    {
      "epoch": 0.7076378740096916,
      "grad_norm": 6.3125,
      "learning_rate": 1.938543188985463e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 3999970.0,
      "step": 9200
    },
    {
      "epoch": 0.70771479116991,
      "grad_norm": 12.9375,
      "learning_rate": 1.938427813245135e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4000516.0,
      "step": 9201
    },
    {
      "epoch": 0.7077917083301285,
      "grad_norm": 2.671875,
      "learning_rate": 1.9383124375048073e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4001034.0,
      "step": 9202
    },
    {
      "epoch": 0.7078686254903469,
      "grad_norm": 14.75,
      "learning_rate": 1.93819706176448e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4001376.0,
      "step": 9203
    },
    {
      "epoch": 0.7079455426505653,
      "grad_norm": 21.25,
      "learning_rate": 1.938081686024152e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.9825327396392822,
      "num_tokens": 4001606.0,
      "step": 9204
    },
    {
      "epoch": 0.7080224598107838,
      "grad_norm": 18.375,
      "learning_rate": 1.937966310283824e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 4002178.0,
      "step": 9205
    },
    {
      "epoch": 0.7080993769710022,
      "grad_norm": 4.46875,
      "learning_rate": 1.937850934543497e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 4002734.0,
      "step": 9206
    },
    {
      "epoch": 0.7081762941312206,
      "grad_norm": 8.1875,
      "learning_rate": 1.937735558803169e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 4003026.0,
      "step": 9207
    },
    {
      "epoch": 0.7082532112914391,
      "grad_norm": 8.3125,
      "learning_rate": 1.9376201830628413e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 4003542.0,
      "step": 9208
    },
    {
      "epoch": 0.7083301284516575,
      "grad_norm": 27.5,
      "learning_rate": 1.9375048073225136e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 4003852.0,
      "step": 9209
    },
    {
      "epoch": 0.7084070456118761,
      "grad_norm": 30.625,
      "learning_rate": 1.937389431582186e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 4004206.0,
      "step": 9210
    },
    {
      "epoch": 0.7084839627720945,
      "grad_norm": 18.75,
      "learning_rate": 1.9372740558418585e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9849849939346313,
      "num_tokens": 4004540.0,
      "step": 9211
    },
    {
      "epoch": 0.7085608799323129,
      "grad_norm": 13.25,
      "learning_rate": 1.9371586801015308e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 4004902.0,
      "step": 9212
    },
    {
      "epoch": 0.7086377970925314,
      "grad_norm": 4.0625,
      "learning_rate": 1.937043304361203e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4005424.0,
      "step": 9213
    },
    {
      "epoch": 0.7087147142527498,
      "grad_norm": 4.53125,
      "learning_rate": 1.9369279286208757e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 4005980.0,
      "step": 9214
    },
    {
      "epoch": 0.7087916314129682,
      "grad_norm": 15.125,
      "learning_rate": 1.9368125528805475e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4006292.0,
      "step": 9215
    },
    {
      "epoch": 0.7088685485731867,
      "grad_norm": 6.75,
      "learning_rate": 1.93669717714022e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9961904883384705,
      "num_tokens": 4006818.0,
      "step": 9216
    },
    {
      "epoch": 0.7089454657334051,
      "grad_norm": 4.96875,
      "learning_rate": 1.9365818013998925e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 4007638.0,
      "step": 9217
    },
    {
      "epoch": 0.7090223828936235,
      "grad_norm": 4.8125,
      "learning_rate": 1.9364664256595647e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4008156.0,
      "step": 9218
    },
    {
      "epoch": 0.709099300053842,
      "grad_norm": 7.125,
      "learning_rate": 1.936351049919237e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4008720.0,
      "step": 9219
    },
    {
      "epoch": 0.7091762172140604,
      "grad_norm": 5.0625,
      "learning_rate": 1.9362356741789092e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 4009300.0,
      "step": 9220
    },
    {
      "epoch": 0.709253134374279,
      "grad_norm": 7.59375,
      "learning_rate": 1.9361202984385815e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4009582.0,
      "step": 9221
    },
    {
      "epoch": 0.7093300515344974,
      "grad_norm": 18.25,
      "learning_rate": 1.936004922698254e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4009922.0,
      "step": 9222
    },
    {
      "epoch": 0.7094069686947158,
      "grad_norm": 10.0625,
      "learning_rate": 1.9358895469579264e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4010246.0,
      "step": 9223
    },
    {
      "epoch": 0.7094838858549343,
      "grad_norm": 5.03125,
      "learning_rate": 1.9357741712175987e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 4010782.0,
      "step": 9224
    },
    {
      "epoch": 0.7095608030151527,
      "grad_norm": 7.21875,
      "learning_rate": 1.935658795477271e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4011356.0,
      "step": 9225
    },
    {
      "epoch": 0.7096377201753711,
      "grad_norm": 3.890625,
      "learning_rate": 1.9355434197369432e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4011918.0,
      "step": 9226
    },
    {
      "epoch": 0.7097146373355896,
      "grad_norm": 4.375,
      "learning_rate": 1.9354280439966155e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9974358677864075,
      "num_tokens": 4012700.0,
      "step": 9227
    },
    {
      "epoch": 0.709791554495808,
      "grad_norm": 22.625,
      "learning_rate": 1.935312668256288e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 4013032.0,
      "step": 9228
    },
    {
      "epoch": 0.7098684716560265,
      "grad_norm": 13.4375,
      "learning_rate": 1.9351972925159604e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4013356.0,
      "step": 9229
    },
    {
      "epoch": 0.7099453888162449,
      "grad_norm": 16.625,
      "learning_rate": 1.9350819167756327e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 4013648.0,
      "step": 9230
    },
    {
      "epoch": 0.7100223059764633,
      "grad_norm": 25.25,
      "learning_rate": 1.934966541035305e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 4013964.0,
      "step": 9231
    },
    {
      "epoch": 0.7100992231366818,
      "grad_norm": 20.125,
      "learning_rate": 1.934851165294977e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.982758641242981,
      "num_tokens": 4014200.0,
      "step": 9232
    },
    {
      "epoch": 0.7101761402969002,
      "grad_norm": 44.0,
      "learning_rate": 1.93473578955465e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 4014548.0,
      "step": 9233
    },
    {
      "epoch": 0.7102530574571186,
      "grad_norm": 7.28125,
      "learning_rate": 1.934620413814322e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9974358677864075,
      "num_tokens": 4015330.0,
      "step": 9234
    },
    {
      "epoch": 0.7103299746173372,
      "grad_norm": 7.5625,
      "learning_rate": 1.9345050380739944e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4015682.0,
      "step": 9235
    },
    {
      "epoch": 0.7104068917775556,
      "grad_norm": 30.875,
      "learning_rate": 1.9343896623336667e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4015986.0,
      "step": 9236
    },
    {
      "epoch": 0.710483808937774,
      "grad_norm": 5.03125,
      "learning_rate": 1.934274286593339e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4016556.0,
      "step": 9237
    },
    {
      "epoch": 0.7105607260979925,
      "grad_norm": 30.875,
      "learning_rate": 1.934158910853011e-06,
      "loss": 0.0302,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 4016862.0,
      "step": 9238
    },
    {
      "epoch": 0.7106376432582109,
      "grad_norm": 24.0,
      "learning_rate": 1.934043535112684e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 4017440.0,
      "step": 9239
    },
    {
      "epoch": 0.7107145604184294,
      "grad_norm": 31.125,
      "learning_rate": 1.933928159372356e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4017724.0,
      "step": 9240
    },
    {
      "epoch": 0.7107914775786478,
      "grad_norm": 43.75,
      "learning_rate": 1.9338127836320284e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.9814241528511047,
      "num_tokens": 4018048.0,
      "step": 9241
    },
    {
      "epoch": 0.7108683947388662,
      "grad_norm": 35.5,
      "learning_rate": 1.9336974078917006e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 4018406.0,
      "step": 9242
    },
    {
      "epoch": 0.7109453118990847,
      "grad_norm": 17.0,
      "learning_rate": 1.933582032151373e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 4018996.0,
      "step": 9243
    },
    {
      "epoch": 0.7110222290593031,
      "grad_norm": 15.5,
      "learning_rate": 1.9334666564110456e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 4019552.0,
      "step": 9244
    },
    {
      "epoch": 0.7110991462195215,
      "grad_norm": 10.5625,
      "learning_rate": 1.933351280670718e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 4019892.0,
      "step": 9245
    },
    {
      "epoch": 0.71117606337974,
      "grad_norm": 3.28125,
      "learning_rate": 1.93323590493039e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4020410.0,
      "step": 9246
    },
    {
      "epoch": 0.7112529805399584,
      "grad_norm": 6.25,
      "learning_rate": 1.9331205291900623e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 4021218.0,
      "step": 9247
    },
    {
      "epoch": 0.711329897700177,
      "grad_norm": 7.28125,
      "learning_rate": 1.9330051534497346e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 4021520.0,
      "step": 9248
    },
    {
      "epoch": 0.7114068148603954,
      "grad_norm": 7.0625,
      "learning_rate": 1.932889777709407e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4022072.0,
      "step": 9249
    },
    {
      "epoch": 0.7114837320206138,
      "grad_norm": 9.0,
      "learning_rate": 1.9327744019690795e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4022632.0,
      "step": 9250
    },
    {
      "epoch": 0.7115606491808323,
      "grad_norm": 7.84375,
      "learning_rate": 1.932659026228752e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4023206.0,
      "step": 9251
    },
    {
      "epoch": 0.7116375663410507,
      "grad_norm": 35.25,
      "learning_rate": 1.932543650488424e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4023538.0,
      "step": 9252
    },
    {
      "epoch": 0.7117144835012691,
      "grad_norm": 41.5,
      "learning_rate": 1.9324282747480963e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 4023906.0,
      "step": 9253
    },
    {
      "epoch": 0.7117914006614876,
      "grad_norm": 4.40625,
      "learning_rate": 1.9323128990077686e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 4024498.0,
      "step": 9254
    },
    {
      "epoch": 0.711868317821706,
      "grad_norm": 29.625,
      "learning_rate": 1.932197523267441e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4025014.0,
      "step": 9255
    },
    {
      "epoch": 0.7119452349819244,
      "grad_norm": 17.125,
      "learning_rate": 1.9320821475271135e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4025560.0,
      "step": 9256
    },
    {
      "epoch": 0.7120221521421429,
      "grad_norm": 10.9375,
      "learning_rate": 1.9319667717867858e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 4025830.0,
      "step": 9257
    },
    {
      "epoch": 0.7120990693023613,
      "grad_norm": 39.5,
      "learning_rate": 1.931851396046458e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4026110.0,
      "step": 9258
    },
    {
      "epoch": 0.7121759864625798,
      "grad_norm": 7.34375,
      "learning_rate": 1.9317360203061303e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4026632.0,
      "step": 9259
    },
    {
      "epoch": 0.7122529036227983,
      "grad_norm": 10.0625,
      "learning_rate": 1.9316206445658025e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 4027436.0,
      "step": 9260
    },
    {
      "epoch": 0.7123298207830167,
      "grad_norm": 5.96875,
      "learning_rate": 1.9315052688254752e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4028194.0,
      "step": 9261
    },
    {
      "epoch": 0.7124067379432352,
      "grad_norm": 11.5,
      "learning_rate": 1.9313898930851475e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 4028492.0,
      "step": 9262
    },
    {
      "epoch": 0.7124836551034536,
      "grad_norm": 3.015625,
      "learning_rate": 1.9312745173448193e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4029038.0,
      "step": 9263
    },
    {
      "epoch": 0.712560572263672,
      "grad_norm": 13.125,
      "learning_rate": 1.931159141604492e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 4029302.0,
      "step": 9264
    },
    {
      "epoch": 0.7126374894238905,
      "grad_norm": 9.1875,
      "learning_rate": 1.9310437658641643e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 4029822.0,
      "step": 9265
    },
    {
      "epoch": 0.7127144065841089,
      "grad_norm": 62.75,
      "learning_rate": 1.9309283901238365e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4030132.0,
      "step": 9266
    },
    {
      "epoch": 0.7127913237443274,
      "grad_norm": 22.5,
      "learning_rate": 1.930813014383509e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 4030702.0,
      "step": 9267
    },
    {
      "epoch": 0.7128682409045458,
      "grad_norm": 57.25,
      "learning_rate": 1.9306976386431815e-06,
      "loss": 0.0383,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4031034.0,
      "step": 9268
    },
    {
      "epoch": 0.7129451580647642,
      "grad_norm": 10.8125,
      "learning_rate": 1.9305822629028537e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4031368.0,
      "step": 9269
    },
    {
      "epoch": 0.7130220752249827,
      "grad_norm": 12.75,
      "learning_rate": 1.930466887162526e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4031912.0,
      "step": 9270
    },
    {
      "epoch": 0.7130989923852011,
      "grad_norm": 31.375,
      "learning_rate": 1.9303515114221982e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4032194.0,
      "step": 9271
    },
    {
      "epoch": 0.7131759095454195,
      "grad_norm": 32.25,
      "learning_rate": 1.930236135681871e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4032526.0,
      "step": 9272
    },
    {
      "epoch": 0.713252826705638,
      "grad_norm": 19.375,
      "learning_rate": 1.930120759941543e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4032782.0,
      "step": 9273
    },
    {
      "epoch": 0.7133297438658565,
      "grad_norm": 52.0,
      "learning_rate": 1.930005384201215e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 4033116.0,
      "step": 9274
    },
    {
      "epoch": 0.7134066610260749,
      "grad_norm": 14.625,
      "learning_rate": 1.9298900084608877e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4033452.0,
      "step": 9275
    },
    {
      "epoch": 0.7134835781862934,
      "grad_norm": 15.1875,
      "learning_rate": 1.92977463272056e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 4034028.0,
      "step": 9276
    },
    {
      "epoch": 0.7135604953465118,
      "grad_norm": 19.0,
      "learning_rate": 1.929659256980232e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 4034586.0,
      "step": 9277
    },
    {
      "epoch": 0.7136374125067303,
      "grad_norm": 40.5,
      "learning_rate": 1.929543881239905e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4034890.0,
      "step": 9278
    },
    {
      "epoch": 0.7137143296669487,
      "grad_norm": 30.375,
      "learning_rate": 1.929428505499577e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4035226.0,
      "step": 9279
    },
    {
      "epoch": 0.7137912468271671,
      "grad_norm": 4.0,
      "learning_rate": 1.9293131297592494e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4035794.0,
      "step": 9280
    },
    {
      "epoch": 0.7138681639873856,
      "grad_norm": 25.875,
      "learning_rate": 1.9291977540189217e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4036134.0,
      "step": 9281
    },
    {
      "epoch": 0.713945081147604,
      "grad_norm": 2.890625,
      "learning_rate": 1.929082378278594e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4036652.0,
      "step": 9282
    },
    {
      "epoch": 0.7140219983078224,
      "grad_norm": 40.25,
      "learning_rate": 1.9289670025382666e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4036958.0,
      "step": 9283
    },
    {
      "epoch": 0.7140989154680409,
      "grad_norm": 12.1875,
      "learning_rate": 1.928851626797939e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 4037478.0,
      "step": 9284
    },
    {
      "epoch": 0.7141758326282593,
      "grad_norm": 3.375,
      "learning_rate": 1.9287362510576107e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9980732202529907,
      "num_tokens": 4037998.0,
      "step": 9285
    },
    {
      "epoch": 0.7142527497884779,
      "grad_norm": 15.3125,
      "learning_rate": 1.9286208753172834e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 4038546.0,
      "step": 9286
    },
    {
      "epoch": 0.7143296669486963,
      "grad_norm": 23.375,
      "learning_rate": 1.9285054995769556e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 4038800.0,
      "step": 9287
    },
    {
      "epoch": 0.7144065841089147,
      "grad_norm": 10.3125,
      "learning_rate": 1.928390123836628e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4039138.0,
      "step": 9288
    },
    {
      "epoch": 0.7144835012691332,
      "grad_norm": 7.4375,
      "learning_rate": 1.9282747480963006e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 4039950.0,
      "step": 9289
    },
    {
      "epoch": 0.7145604184293516,
      "grad_norm": 4.28125,
      "learning_rate": 1.9281593723559724e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 4040736.0,
      "step": 9290
    },
    {
      "epoch": 0.71463733558957,
      "grad_norm": 28.25,
      "learning_rate": 1.928043996615645e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 4041042.0,
      "step": 9291
    },
    {
      "epoch": 0.7147142527497885,
      "grad_norm": 11.75,
      "learning_rate": 1.9279286208753173e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9950900077819824,
      "num_tokens": 4041654.0,
      "step": 9292
    },
    {
      "epoch": 0.7147911699100069,
      "grad_norm": 5.0,
      "learning_rate": 1.9278132451349896e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4042216.0,
      "step": 9293
    },
    {
      "epoch": 0.7148680870702254,
      "grad_norm": 7.25,
      "learning_rate": 1.9276978693946623e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 4042802.0,
      "step": 9294
    },
    {
      "epoch": 0.7149450042304438,
      "grad_norm": 7.125,
      "learning_rate": 1.9275824936543345e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4043324.0,
      "step": 9295
    },
    {
      "epoch": 0.7150219213906622,
      "grad_norm": 3.09375,
      "learning_rate": 1.9274671179140064e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 4043860.0,
      "step": 9296
    },
    {
      "epoch": 0.7150988385508807,
      "grad_norm": 19.25,
      "learning_rate": 1.927351742173679e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4044424.0,
      "step": 9297
    },
    {
      "epoch": 0.7151757557110991,
      "grad_norm": 20.375,
      "learning_rate": 1.9272363664333513e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 4044752.0,
      "step": 9298
    },
    {
      "epoch": 0.7152526728713176,
      "grad_norm": 8.25,
      "learning_rate": 1.9271209906930236e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 4045100.0,
      "step": 9299
    },
    {
      "epoch": 0.7153295900315361,
      "grad_norm": 14.9375,
      "learning_rate": 1.9270056149526963e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 4045412.0,
      "step": 9300
    },
    {
      "epoch": 0.7154065071917545,
      "grad_norm": 1.890625,
      "learning_rate": 1.926890239212368e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4045976.0,
      "step": 9301
    },
    {
      "epoch": 0.7154834243519729,
      "grad_norm": 8.125,
      "learning_rate": 1.9267748634720408e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9974358677864075,
      "num_tokens": 4046758.0,
      "step": 9302
    },
    {
      "epoch": 0.7155603415121914,
      "grad_norm": 17.5,
      "learning_rate": 1.926659487731713e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 4046992.0,
      "step": 9303
    },
    {
      "epoch": 0.7156372586724098,
      "grad_norm": 15.5,
      "learning_rate": 1.9265441119913853e-06,
      "loss": 0.0417,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4047270.0,
      "step": 9304
    },
    {
      "epoch": 0.7157141758326283,
      "grad_norm": 9.5,
      "learning_rate": 1.926428736251058e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9974293112754822,
      "num_tokens": 4048050.0,
      "step": 9305
    },
    {
      "epoch": 0.7157910929928467,
      "grad_norm": 17.625,
      "learning_rate": 1.9263133605107302e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4048372.0,
      "step": 9306
    },
    {
      "epoch": 0.7158680101530651,
      "grad_norm": 8.25,
      "learning_rate": 1.926197984770402e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 4048936.0,
      "step": 9307
    },
    {
      "epoch": 0.7159449273132836,
      "grad_norm": 5.90625,
      "learning_rate": 1.9260826090300747e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4049504.0,
      "step": 9308
    },
    {
      "epoch": 0.716021844473502,
      "grad_norm": 12.25,
      "learning_rate": 1.925967233289747e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 4049854.0,
      "step": 9309
    },
    {
      "epoch": 0.7160987616337204,
      "grad_norm": 9.875,
      "learning_rate": 1.9258518575494193e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4050160.0,
      "step": 9310
    },
    {
      "epoch": 0.716175678793939,
      "grad_norm": 6.71875,
      "learning_rate": 1.925736481809092e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4050684.0,
      "step": 9311
    },
    {
      "epoch": 0.7162525959541574,
      "grad_norm": 10.125,
      "learning_rate": 1.9256211060687638e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4050966.0,
      "step": 9312
    },
    {
      "epoch": 0.7163295131143759,
      "grad_norm": 39.75,
      "learning_rate": 1.9255057303284365e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4051322.0,
      "step": 9313
    },
    {
      "epoch": 0.7164064302745943,
      "grad_norm": 80.5,
      "learning_rate": 1.9253903545881087e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4051650.0,
      "step": 9314
    },
    {
      "epoch": 0.7164833474348127,
      "grad_norm": 6.8125,
      "learning_rate": 1.925274978847781e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 4051942.0,
      "step": 9315
    },
    {
      "epoch": 0.7165602645950312,
      "grad_norm": 26.375,
      "learning_rate": 1.9251596031074532e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 4052544.0,
      "step": 9316
    },
    {
      "epoch": 0.7166371817552496,
      "grad_norm": 12.375,
      "learning_rate": 1.925044227367126e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4052888.0,
      "step": 9317
    },
    {
      "epoch": 0.716714098915468,
      "grad_norm": 17.0,
      "learning_rate": 1.9249288516267977e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4053446.0,
      "step": 9318
    },
    {
      "epoch": 0.7167910160756865,
      "grad_norm": 43.75,
      "learning_rate": 1.9248134758864704e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4053774.0,
      "step": 9319
    },
    {
      "epoch": 0.7168679332359049,
      "grad_norm": 41.75,
      "learning_rate": 1.9246981001461427e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4054096.0,
      "step": 9320
    },
    {
      "epoch": 0.7169448503961233,
      "grad_norm": 7.90625,
      "learning_rate": 1.924582724405815e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4054670.0,
      "step": 9321
    },
    {
      "epoch": 0.7170217675563418,
      "grad_norm": 21.0,
      "learning_rate": 1.9244673486654876e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4055232.0,
      "step": 9322
    },
    {
      "epoch": 0.7170986847165602,
      "grad_norm": 10.3125,
      "learning_rate": 1.9243519729251595e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 4055594.0,
      "step": 9323
    },
    {
      "epoch": 0.7171756018767788,
      "grad_norm": 50.75,
      "learning_rate": 1.9242365971848317e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4055872.0,
      "step": 9324
    },
    {
      "epoch": 0.7172525190369972,
      "grad_norm": 8.625,
      "learning_rate": 1.9241212214445044e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 4056472.0,
      "step": 9325
    },
    {
      "epoch": 0.7173294361972156,
      "grad_norm": 10.625,
      "learning_rate": 1.9240058457041767e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4056780.0,
      "step": 9326
    },
    {
      "epoch": 0.7174063533574341,
      "grad_norm": 5.75,
      "learning_rate": 1.923890469963849e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4057326.0,
      "step": 9327
    },
    {
      "epoch": 0.7174832705176525,
      "grad_norm": 10.25,
      "learning_rate": 1.923775094223521e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 4057894.0,
      "step": 9328
    },
    {
      "epoch": 0.7175601876778709,
      "grad_norm": 7.6875,
      "learning_rate": 1.9236597184831934e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4058702.0,
      "step": 9329
    },
    {
      "epoch": 0.7176371048380894,
      "grad_norm": 46.25,
      "learning_rate": 1.923544342742866e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4058984.0,
      "step": 9330
    },
    {
      "epoch": 0.7177140219983078,
      "grad_norm": 6.75,
      "learning_rate": 1.9234289670025384e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4059544.0,
      "step": 9331
    },
    {
      "epoch": 0.7177909391585263,
      "grad_norm": 32.5,
      "learning_rate": 1.9233135912622106e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 4059850.0,
      "step": 9332
    },
    {
      "epoch": 0.7178678563187447,
      "grad_norm": 5.375,
      "learning_rate": 1.9231982155218833e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 4060638.0,
      "step": 9333
    },
    {
      "epoch": 0.7179447734789631,
      "grad_norm": 6.15625,
      "learning_rate": 1.923082839781555e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4061190.0,
      "step": 9334
    },
    {
      "epoch": 0.7180216906391816,
      "grad_norm": 2.453125,
      "learning_rate": 1.9229674640412274e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4061738.0,
      "step": 9335
    },
    {
      "epoch": 0.7180986077994,
      "grad_norm": 6.125,
      "learning_rate": 1.9228520883009e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4062254.0,
      "step": 9336
    },
    {
      "epoch": 0.7181755249596184,
      "grad_norm": 41.75,
      "learning_rate": 1.9227367125605723e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4062600.0,
      "step": 9337
    },
    {
      "epoch": 0.718252442119837,
      "grad_norm": 22.0,
      "learning_rate": 1.9226213368202446e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 4062960.0,
      "step": 9338
    },
    {
      "epoch": 0.7183293592800554,
      "grad_norm": 16.875,
      "learning_rate": 1.922505961079917e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 4063248.0,
      "step": 9339
    },
    {
      "epoch": 0.7184062764402738,
      "grad_norm": 15.5,
      "learning_rate": 1.922390585339589e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4063546.0,
      "step": 9340
    },
    {
      "epoch": 0.7184831936004923,
      "grad_norm": 12.375,
      "learning_rate": 1.922275209599262e-06,
      "loss": 0.0054,
      "mean_token_accuracy": 0.999053955078125,
      "num_tokens": 4064606.0,
      "step": 9341
    },
    {
      "epoch": 0.7185601107607107,
      "grad_norm": 19.75,
      "learning_rate": 1.922159833858934e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4064930.0,
      "step": 9342
    },
    {
      "epoch": 0.7186370279209292,
      "grad_norm": 6.71875,
      "learning_rate": 1.9220444581186063e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4065502.0,
      "step": 9343
    },
    {
      "epoch": 0.7187139450811476,
      "grad_norm": 12.8125,
      "learning_rate": 1.921929082378279e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4066020.0,
      "step": 9344
    },
    {
      "epoch": 0.718790862241366,
      "grad_norm": 10.8125,
      "learning_rate": 1.921813706637951e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 4066578.0,
      "step": 9345
    },
    {
      "epoch": 0.7188677794015845,
      "grad_norm": 5.59375,
      "learning_rate": 1.921698330897623e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4067102.0,
      "step": 9346
    },
    {
      "epoch": 0.7189446965618029,
      "grad_norm": 5.75,
      "learning_rate": 1.9215829551572958e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9974424839019775,
      "num_tokens": 4067886.0,
      "step": 9347
    },
    {
      "epoch": 0.7190216137220213,
      "grad_norm": 34.5,
      "learning_rate": 1.921467579416968e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 4068192.0,
      "step": 9348
    },
    {
      "epoch": 0.7190985308822399,
      "grad_norm": 4.03125,
      "learning_rate": 1.9213522036766403e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 4068788.0,
      "step": 9349
    },
    {
      "epoch": 0.7191754480424583,
      "grad_norm": 29.625,
      "learning_rate": 1.9212368279363125e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4069066.0,
      "step": 9350
    },
    {
      "epoch": 0.7192523652026768,
      "grad_norm": 20.25,
      "learning_rate": 1.921121452195985e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4069326.0,
      "step": 9351
    },
    {
      "epoch": 0.7193292823628952,
      "grad_norm": 4.34375,
      "learning_rate": 1.9210060764556575e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9987146258354187,
      "num_tokens": 4070106.0,
      "step": 9352
    },
    {
      "epoch": 0.7194061995231136,
      "grad_norm": 3.0,
      "learning_rate": 1.9208907007153297e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4070664.0,
      "step": 9353
    },
    {
      "epoch": 0.7194831166833321,
      "grad_norm": 26.875,
      "learning_rate": 1.920775324975002e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4070976.0,
      "step": 9354
    },
    {
      "epoch": 0.7195600338435505,
      "grad_norm": 8.4375,
      "learning_rate": 1.9206599492346743e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 4071572.0,
      "step": 9355
    },
    {
      "epoch": 0.7196369510037689,
      "grad_norm": 17.375,
      "learning_rate": 1.9205445734943465e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4071858.0,
      "step": 9356
    },
    {
      "epoch": 0.7197138681639874,
      "grad_norm": 2.84375,
      "learning_rate": 1.9204291977540188e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4072382.0,
      "step": 9357
    },
    {
      "epoch": 0.7197907853242058,
      "grad_norm": 16.5,
      "learning_rate": 1.9203138220136915e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9907235503196716,
      "num_tokens": 4072922.0,
      "step": 9358
    },
    {
      "epoch": 0.7198677024844242,
      "grad_norm": 8.3125,
      "learning_rate": 1.9201984462733637e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4073496.0,
      "step": 9359
    },
    {
      "epoch": 0.7199446196446427,
      "grad_norm": 6.75,
      "learning_rate": 1.920083070533036e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4074020.0,
      "step": 9360
    },
    {
      "epoch": 0.7200215368048611,
      "grad_norm": 22.25,
      "learning_rate": 1.9199676947927082e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4074360.0,
      "step": 9361
    },
    {
      "epoch": 0.7200984539650797,
      "grad_norm": 5.9375,
      "learning_rate": 1.9198523190523805e-06,
      "loss": 0.009,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4074902.0,
      "step": 9362
    },
    {
      "epoch": 0.7201753711252981,
      "grad_norm": 10.0625,
      "learning_rate": 1.919736943312053e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4075488.0,
      "step": 9363
    },
    {
      "epoch": 0.7202522882855165,
      "grad_norm": 31.625,
      "learning_rate": 1.9196215675717254e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 4075822.0,
      "step": 9364
    },
    {
      "epoch": 0.720329205445735,
      "grad_norm": 26.875,
      "learning_rate": 1.9195061918313977e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 4076150.0,
      "step": 9365
    },
    {
      "epoch": 0.7204061226059534,
      "grad_norm": 12.25,
      "learning_rate": 1.91939081609107e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4076468.0,
      "step": 9366
    },
    {
      "epoch": 0.7204830397661718,
      "grad_norm": 36.25,
      "learning_rate": 1.919275440350742e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4076774.0,
      "step": 9367
    },
    {
      "epoch": 0.7205599569263903,
      "grad_norm": 25.75,
      "learning_rate": 1.9191600646104145e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4077056.0,
      "step": 9368
    },
    {
      "epoch": 0.7206368740866087,
      "grad_norm": 23.625,
      "learning_rate": 1.919044688870087e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 4077364.0,
      "step": 9369
    },
    {
      "epoch": 0.7207137912468272,
      "grad_norm": 42.0,
      "learning_rate": 1.9189293131297594e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4077688.0,
      "step": 9370
    },
    {
      "epoch": 0.7207907084070456,
      "grad_norm": 28.0,
      "learning_rate": 1.9188139373894317e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4077988.0,
      "step": 9371
    },
    {
      "epoch": 0.720867625567264,
      "grad_norm": 30.125,
      "learning_rate": 1.918698561649104e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 4078542.0,
      "step": 9372
    },
    {
      "epoch": 0.7209445427274825,
      "grad_norm": 38.25,
      "learning_rate": 1.918583185908776e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 4079134.0,
      "step": 9373
    },
    {
      "epoch": 0.7210214598877009,
      "grad_norm": 26.25,
      "learning_rate": 1.9184678101684484e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9809523820877075,
      "num_tokens": 4079450.0,
      "step": 9374
    },
    {
      "epoch": 0.7210983770479193,
      "grad_norm": 31.125,
      "learning_rate": 1.918352434428121e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4080012.0,
      "step": 9375
    },
    {
      "epoch": 0.7211752942081379,
      "grad_norm": 22.375,
      "learning_rate": 1.9182370586877934e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4080558.0,
      "step": 9376
    },
    {
      "epoch": 0.7212522113683563,
      "grad_norm": 32.5,
      "learning_rate": 1.9181216829474656e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4080878.0,
      "step": 9377
    },
    {
      "epoch": 0.7213291285285747,
      "grad_norm": 22.0,
      "learning_rate": 1.918006307207138e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 4081198.0,
      "step": 9378
    },
    {
      "epoch": 0.7214060456887932,
      "grad_norm": 18.25,
      "learning_rate": 1.91789093146681e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 4081514.0,
      "step": 9379
    },
    {
      "epoch": 0.7214829628490116,
      "grad_norm": 4.25,
      "learning_rate": 1.917775555726483e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4082032.0,
      "step": 9380
    },
    {
      "epoch": 0.7215598800092301,
      "grad_norm": 5.15625,
      "learning_rate": 1.917660179986155e-06,
      "loss": 0.0068,
      "mean_token_accuracy": 0.9987775087356567,
      "num_tokens": 4082852.0,
      "step": 9381
    },
    {
      "epoch": 0.7216367971694485,
      "grad_norm": 6.34375,
      "learning_rate": 1.9175448042458273e-06,
      "loss": 0.0074,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 4083632.0,
      "step": 9382
    },
    {
      "epoch": 0.7217137143296669,
      "grad_norm": 13.0,
      "learning_rate": 1.9174294285054996e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4083954.0,
      "step": 9383
    },
    {
      "epoch": 0.7217906314898854,
      "grad_norm": 5.6875,
      "learning_rate": 1.917314052765172e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 4084740.0,
      "step": 9384
    },
    {
      "epoch": 0.7218675486501038,
      "grad_norm": 31.625,
      "learning_rate": 1.917198677024844e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4085024.0,
      "step": 9385
    },
    {
      "epoch": 0.7219444658103222,
      "grad_norm": 4.21875,
      "learning_rate": 1.917083301284517e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9975669384002686,
      "num_tokens": 4085848.0,
      "step": 9386
    },
    {
      "epoch": 0.7220213829705407,
      "grad_norm": 1.5390625,
      "learning_rate": 1.916967925544189e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 4086644.0,
      "step": 9387
    },
    {
      "epoch": 0.7220983001307592,
      "grad_norm": 11.5625,
      "learning_rate": 1.9168525498038613e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4087186.0,
      "step": 9388
    },
    {
      "epoch": 0.7221752172909777,
      "grad_norm": 11.5,
      "learning_rate": 1.9167371740635336e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 4087994.0,
      "step": 9389
    },
    {
      "epoch": 0.7222521344511961,
      "grad_norm": 31.5,
      "learning_rate": 1.916621798323206e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 4088510.0,
      "step": 9390
    },
    {
      "epoch": 0.7223290516114145,
      "grad_norm": 3.5625,
      "learning_rate": 1.9165064225828785e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 4089042.0,
      "step": 9391
    },
    {
      "epoch": 0.722405968771633,
      "grad_norm": 14.0625,
      "learning_rate": 1.9163910468425508e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4089370.0,
      "step": 9392
    },
    {
      "epoch": 0.7224828859318514,
      "grad_norm": 18.5,
      "learning_rate": 1.9162756711022226e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4089690.0,
      "step": 9393
    },
    {
      "epoch": 0.7225598030920698,
      "grad_norm": 25.125,
      "learning_rate": 1.9161602953618953e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4089976.0,
      "step": 9394
    },
    {
      "epoch": 0.7226367202522883,
      "grad_norm": 19.25,
      "learning_rate": 1.9160449196215675e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4090546.0,
      "step": 9395
    },
    {
      "epoch": 0.7227136374125067,
      "grad_norm": 26.25,
      "learning_rate": 1.91592954388124e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4090830.0,
      "step": 9396
    },
    {
      "epoch": 0.7227905545727252,
      "grad_norm": 13.375,
      "learning_rate": 1.9158141681409125e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 4091370.0,
      "step": 9397
    },
    {
      "epoch": 0.7228674717329436,
      "grad_norm": 2.609375,
      "learning_rate": 1.9156987924005847e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4091886.0,
      "step": 9398
    },
    {
      "epoch": 0.722944388893162,
      "grad_norm": 11.25,
      "learning_rate": 1.915583416660257e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 4092690.0,
      "step": 9399
    },
    {
      "epoch": 0.7230213060533806,
      "grad_norm": 22.5,
      "learning_rate": 1.9154680409199293e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4093270.0,
      "step": 9400
    },
    {
      "epoch": 0.723098223213599,
      "grad_norm": 55.5,
      "learning_rate": 1.9153526651796015e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 4093620.0,
      "step": 9401
    },
    {
      "epoch": 0.7231751403738174,
      "grad_norm": 27.25,
      "learning_rate": 1.915237289439274e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 4093952.0,
      "step": 9402
    },
    {
      "epoch": 0.7232520575340359,
      "grad_norm": 9.0,
      "learning_rate": 1.9151219136989465e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 4094502.0,
      "step": 9403
    },
    {
      "epoch": 0.7233289746942543,
      "grad_norm": 23.125,
      "learning_rate": 1.9150065379586183e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9987775087356567,
      "num_tokens": 4095322.0,
      "step": 9404
    },
    {
      "epoch": 0.7234058918544727,
      "grad_norm": 13.6875,
      "learning_rate": 1.914891162218291e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4095840.0,
      "step": 9405
    },
    {
      "epoch": 0.7234828090146912,
      "grad_norm": 11.25,
      "learning_rate": 1.9147757864779632e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9959595799446106,
      "num_tokens": 4096336.0,
      "step": 9406
    },
    {
      "epoch": 0.7235597261749096,
      "grad_norm": 39.75,
      "learning_rate": 1.9146604107376355e-06,
      "loss": 0.0452,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4096614.0,
      "step": 9407
    },
    {
      "epoch": 0.7236366433351281,
      "grad_norm": 29.125,
      "learning_rate": 1.914545034997308e-06,
      "loss": 0.0369,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 4096910.0,
      "step": 9408
    },
    {
      "epoch": 0.7237135604953465,
      "grad_norm": 16.75,
      "learning_rate": 1.9144296592569804e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4097192.0,
      "step": 9409
    },
    {
      "epoch": 0.7237904776555649,
      "grad_norm": 33.5,
      "learning_rate": 1.9143142835166527e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 4097530.0,
      "step": 9410
    },
    {
      "epoch": 0.7238673948157834,
      "grad_norm": 6.4375,
      "learning_rate": 1.914198907776325e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9980582594871521,
      "num_tokens": 4098046.0,
      "step": 9411
    },
    {
      "epoch": 0.7239443119760018,
      "grad_norm": 21.25,
      "learning_rate": 1.914083532035997e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 4098358.0,
      "step": 9412
    },
    {
      "epoch": 0.7240212291362202,
      "grad_norm": 31.625,
      "learning_rate": 1.91396815629567e-06,
      "loss": 0.0408,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4098654.0,
      "step": 9413
    },
    {
      "epoch": 0.7240981462964388,
      "grad_norm": 10.6875,
      "learning_rate": 1.913852780555342e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9974424839019775,
      "num_tokens": 4099438.0,
      "step": 9414
    },
    {
      "epoch": 0.7241750634566572,
      "grad_norm": 6.71875,
      "learning_rate": 1.913737404815014e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 4099994.0,
      "step": 9415
    },
    {
      "epoch": 0.7242519806168757,
      "grad_norm": 31.375,
      "learning_rate": 1.9136220290746867e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4100310.0,
      "step": 9416
    },
    {
      "epoch": 0.7243288977770941,
      "grad_norm": 10.375,
      "learning_rate": 1.913506653334359e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4100650.0,
      "step": 9417
    },
    {
      "epoch": 0.7244058149373125,
      "grad_norm": 28.875,
      "learning_rate": 1.913391277594031e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 4100960.0,
      "step": 9418
    },
    {
      "epoch": 0.724482732097531,
      "grad_norm": 7.25,
      "learning_rate": 1.913275901853704e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4101284.0,
      "step": 9419
    },
    {
      "epoch": 0.7245596492577494,
      "grad_norm": 4.03125,
      "learning_rate": 1.913160526113376e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4101854.0,
      "step": 9420
    },
    {
      "epoch": 0.7246365664179678,
      "grad_norm": 21.875,
      "learning_rate": 1.9130451503730484e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 4102198.0,
      "step": 9421
    },
    {
      "epoch": 0.7247134835781863,
      "grad_norm": 18.5,
      "learning_rate": 1.9129297746327206e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.991482138633728,
      "num_tokens": 4102786.0,
      "step": 9422
    },
    {
      "epoch": 0.7247904007384047,
      "grad_norm": 17.5,
      "learning_rate": 1.912814398892393e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9846153855323792,
      "num_tokens": 4103112.0,
      "step": 9423
    },
    {
      "epoch": 0.7248673178986231,
      "grad_norm": 19.375,
      "learning_rate": 1.912699023152065e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 4103428.0,
      "step": 9424
    },
    {
      "epoch": 0.7249442350588416,
      "grad_norm": 20.375,
      "learning_rate": 1.912583647411738e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4103752.0,
      "step": 9425
    },
    {
      "epoch": 0.72502115221906,
      "grad_norm": 25.75,
      "learning_rate": 1.9124682716714097e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9893333315849304,
      "num_tokens": 4104130.0,
      "step": 9426
    },
    {
      "epoch": 0.7250980693792786,
      "grad_norm": 44.25,
      "learning_rate": 1.9123528959310823e-06,
      "loss": 0.0432,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4104408.0,
      "step": 9427
    },
    {
      "epoch": 0.725174986539497,
      "grad_norm": 3.421875,
      "learning_rate": 1.9122375201907546e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4104930.0,
      "step": 9428
    },
    {
      "epoch": 0.7252519036997154,
      "grad_norm": 22.5,
      "learning_rate": 1.912122144450427e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4105236.0,
      "step": 9429
    },
    {
      "epoch": 0.7253288208599339,
      "grad_norm": 22.0,
      "learning_rate": 1.9120067687100995e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 4105812.0,
      "step": 9430
    },
    {
      "epoch": 0.7254057380201523,
      "grad_norm": 100.0,
      "learning_rate": 1.9118913929697714e-06,
      "loss": 0.0397,
      "mean_token_accuracy": 0.984375,
      "num_tokens": 4106136.0,
      "step": 9431
    },
    {
      "epoch": 0.7254826551803707,
      "grad_norm": 28.375,
      "learning_rate": 1.9117760172294436e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 4106442.0,
      "step": 9432
    },
    {
      "epoch": 0.7255595723405892,
      "grad_norm": 57.5,
      "learning_rate": 1.9116606414891163e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4106774.0,
      "step": 9433
    },
    {
      "epoch": 0.7256364895008076,
      "grad_norm": 43.25,
      "learning_rate": 1.9115452657487886e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4107098.0,
      "step": 9434
    },
    {
      "epoch": 0.7257134066610261,
      "grad_norm": 16.75,
      "learning_rate": 1.911429890008461e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 4107666.0,
      "step": 9435
    },
    {
      "epoch": 0.7257903238212445,
      "grad_norm": 7.71875,
      "learning_rate": 1.9113145142681335e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.997032642364502,
      "num_tokens": 4108004.0,
      "step": 9436
    },
    {
      "epoch": 0.7258672409814629,
      "grad_norm": 22.375,
      "learning_rate": 1.9111991385278054e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4108590.0,
      "step": 9437
    },
    {
      "epoch": 0.7259441581416815,
      "grad_norm": 6.03125,
      "learning_rate": 1.911083762787478e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4108912.0,
      "step": 9438
    },
    {
      "epoch": 0.7260210753018999,
      "grad_norm": 43.75,
      "learning_rate": 1.9109683870471503e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 4109272.0,
      "step": 9439
    },
    {
      "epoch": 0.7260979924621183,
      "grad_norm": 8.4375,
      "learning_rate": 1.9108530113068225e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4110030.0,
      "step": 9440
    },
    {
      "epoch": 0.7261749096223368,
      "grad_norm": 46.0,
      "learning_rate": 1.9107376355664952e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 4110328.0,
      "step": 9441
    },
    {
      "epoch": 0.7262518267825552,
      "grad_norm": 31.25,
      "learning_rate": 1.910622259826167e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4110626.0,
      "step": 9442
    },
    {
      "epoch": 0.7263287439427736,
      "grad_norm": 24.75,
      "learning_rate": 1.9105068840858393e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4110972.0,
      "step": 9443
    },
    {
      "epoch": 0.7264056611029921,
      "grad_norm": 25.5,
      "learning_rate": 1.910391508345512e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4111314.0,
      "step": 9444
    },
    {
      "epoch": 0.7264825782632105,
      "grad_norm": 4.15625,
      "learning_rate": 1.9102761326051843e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 4112334.0,
      "step": 9445
    },
    {
      "epoch": 0.726559495423429,
      "grad_norm": 9.25,
      "learning_rate": 1.9101607568648565e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9915966391563416,
      "num_tokens": 4112692.0,
      "step": 9446
    },
    {
      "epoch": 0.7266364125836474,
      "grad_norm": 17.625,
      "learning_rate": 1.910045381124529e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4113266.0,
      "step": 9447
    },
    {
      "epoch": 0.7267133297438658,
      "grad_norm": 53.25,
      "learning_rate": 1.909930005384201e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 4113598.0,
      "step": 9448
    },
    {
      "epoch": 0.7267902469040843,
      "grad_norm": 54.5,
      "learning_rate": 1.9098146296438737e-06,
      "loss": 0.0493,
      "mean_token_accuracy": 0.9784172773361206,
      "num_tokens": 4113878.0,
      "step": 9449
    },
    {
      "epoch": 0.7268671640643027,
      "grad_norm": 8.5625,
      "learning_rate": 1.909699253903546e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 4114398.0,
      "step": 9450
    },
    {
      "epoch": 0.7269440812245211,
      "grad_norm": 30.625,
      "learning_rate": 1.9095838781632182e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 4114700.0,
      "step": 9451
    },
    {
      "epoch": 0.7270209983847397,
      "grad_norm": 12.125,
      "learning_rate": 1.909468502422891e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4115012.0,
      "step": 9452
    },
    {
      "epoch": 0.7270979155449581,
      "grad_norm": 53.5,
      "learning_rate": 1.9093531266825628e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 4115270.0,
      "step": 9453
    },
    {
      "epoch": 0.7271748327051766,
      "grad_norm": 7.9375,
      "learning_rate": 1.909237750942235e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 4115560.0,
      "step": 9454
    },
    {
      "epoch": 0.727251749865395,
      "grad_norm": 27.5,
      "learning_rate": 1.9091223752019077e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4116146.0,
      "step": 9455
    },
    {
      "epoch": 0.7273286670256134,
      "grad_norm": 46.0,
      "learning_rate": 1.90900699946158e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4116402.0,
      "step": 9456
    },
    {
      "epoch": 0.7274055841858319,
      "grad_norm": 3.578125,
      "learning_rate": 1.908891623721252e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4116946.0,
      "step": 9457
    },
    {
      "epoch": 0.7274825013460503,
      "grad_norm": 27.5,
      "learning_rate": 1.9087762479809245e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4117532.0,
      "step": 9458
    },
    {
      "epoch": 0.7275594185062687,
      "grad_norm": 31.125,
      "learning_rate": 1.9086608722405967e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4117872.0,
      "step": 9459
    },
    {
      "epoch": 0.7276363356664872,
      "grad_norm": 40.5,
      "learning_rate": 1.9085454965002694e-06,
      "loss": 0.0418,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4118220.0,
      "step": 9460
    },
    {
      "epoch": 0.7277132528267056,
      "grad_norm": 42.5,
      "learning_rate": 1.9084301207599417e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 4118582.0,
      "step": 9461
    },
    {
      "epoch": 0.727790169986924,
      "grad_norm": 4.34375,
      "learning_rate": 1.908314745019614e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4118894.0,
      "step": 9462
    },
    {
      "epoch": 0.7278670871471425,
      "grad_norm": 45.0,
      "learning_rate": 1.9081993692792866e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4119250.0,
      "step": 9463
    },
    {
      "epoch": 0.727944004307361,
      "grad_norm": 9.625,
      "learning_rate": 1.9080839935389584e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4119548.0,
      "step": 9464
    },
    {
      "epoch": 0.7280209214675795,
      "grad_norm": 5.71875,
      "learning_rate": 1.9079686177986307e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 4120108.0,
      "step": 9465
    },
    {
      "epoch": 0.7280978386277979,
      "grad_norm": 62.5,
      "learning_rate": 1.9078532420583034e-06,
      "loss": 0.0382,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4120410.0,
      "step": 9466
    },
    {
      "epoch": 0.7281747557880163,
      "grad_norm": 45.75,
      "learning_rate": 1.9077378663179756e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4120692.0,
      "step": 9467
    },
    {
      "epoch": 0.7282516729482348,
      "grad_norm": 15.9375,
      "learning_rate": 1.907622490577648e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 4121040.0,
      "step": 9468
    },
    {
      "epoch": 0.7283285901084532,
      "grad_norm": 33.75,
      "learning_rate": 1.90750711483732e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4121350.0,
      "step": 9469
    },
    {
      "epoch": 0.7284055072686716,
      "grad_norm": 27.875,
      "learning_rate": 1.9073917390969924e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4121664.0,
      "step": 9470
    },
    {
      "epoch": 0.7284824244288901,
      "grad_norm": 37.0,
      "learning_rate": 1.9072763633566649e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4121982.0,
      "step": 9471
    },
    {
      "epoch": 0.7285593415891085,
      "grad_norm": 34.5,
      "learning_rate": 1.9071609876163373e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4122328.0,
      "step": 9472
    },
    {
      "epoch": 0.728636258749327,
      "grad_norm": 18.625,
      "learning_rate": 1.9070456118760096e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4122658.0,
      "step": 9473
    },
    {
      "epoch": 0.7287131759095454,
      "grad_norm": 11.0,
      "learning_rate": 1.906930236135682e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 4122890.0,
      "step": 9474
    },
    {
      "epoch": 0.7287900930697638,
      "grad_norm": 26.125,
      "learning_rate": 1.9068148603953541e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4123214.0,
      "step": 9475
    },
    {
      "epoch": 0.7288670102299823,
      "grad_norm": 20.25,
      "learning_rate": 1.9066994846550266e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4123560.0,
      "step": 9476
    },
    {
      "epoch": 0.7289439273902008,
      "grad_norm": 11.8125,
      "learning_rate": 1.9065841089146989e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 4124108.0,
      "step": 9477
    },
    {
      "epoch": 0.7290208445504192,
      "grad_norm": 15.0,
      "learning_rate": 1.9064687331743713e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4124660.0,
      "step": 9478
    },
    {
      "epoch": 0.7290977617106377,
      "grad_norm": 10.9375,
      "learning_rate": 1.9063533574340438e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4124938.0,
      "step": 9479
    },
    {
      "epoch": 0.7291746788708561,
      "grad_norm": 13.25,
      "learning_rate": 1.9062379816937158e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4125282.0,
      "step": 9480
    },
    {
      "epoch": 0.7292515960310745,
      "grad_norm": 29.0,
      "learning_rate": 1.906122605953388e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9814241528511047,
      "num_tokens": 4125606.0,
      "step": 9481
    },
    {
      "epoch": 0.729328513191293,
      "grad_norm": 8.625,
      "learning_rate": 1.9060072302130606e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 4126430.0,
      "step": 9482
    },
    {
      "epoch": 0.7294054303515114,
      "grad_norm": 11.125,
      "learning_rate": 1.905891854472733e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4126952.0,
      "step": 9483
    },
    {
      "epoch": 0.7294823475117299,
      "grad_norm": 18.25,
      "learning_rate": 1.9057764787324053e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4127710.0,
      "step": 9484
    },
    {
      "epoch": 0.7295592646719483,
      "grad_norm": 5.3125,
      "learning_rate": 1.9056611029920778e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4128278.0,
      "step": 9485
    },
    {
      "epoch": 0.7296361818321667,
      "grad_norm": 14.625,
      "learning_rate": 1.9055457272517498e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4128618.0,
      "step": 9486
    },
    {
      "epoch": 0.7297130989923852,
      "grad_norm": 34.5,
      "learning_rate": 1.9054303515114223e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4128930.0,
      "step": 9487
    },
    {
      "epoch": 0.7297900161526036,
      "grad_norm": 27.25,
      "learning_rate": 1.9053149757710945e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4129240.0,
      "step": 9488
    },
    {
      "epoch": 0.729866933312822,
      "grad_norm": 3.078125,
      "learning_rate": 1.905199600030767e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 4130020.0,
      "step": 9489
    },
    {
      "epoch": 0.7299438504730406,
      "grad_norm": 4.65625,
      "learning_rate": 1.9050842242904395e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4130568.0,
      "step": 9490
    },
    {
      "epoch": 0.730020767633259,
      "grad_norm": 10.125,
      "learning_rate": 1.9049688485501115e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4131112.0,
      "step": 9491
    },
    {
      "epoch": 0.7300976847934775,
      "grad_norm": 28.375,
      "learning_rate": 1.9048534728097838e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4131464.0,
      "step": 9492
    },
    {
      "epoch": 0.7301746019536959,
      "grad_norm": 9.4375,
      "learning_rate": 1.9047380970694563e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4131840.0,
      "step": 9493
    },
    {
      "epoch": 0.7302515191139143,
      "grad_norm": 21.0,
      "learning_rate": 1.9046227213291287e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 4132156.0,
      "step": 9494
    },
    {
      "epoch": 0.7303284362741328,
      "grad_norm": 12.375,
      "learning_rate": 1.904507345588801e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9971751570701599,
      "num_tokens": 4132512.0,
      "step": 9495
    },
    {
      "epoch": 0.7304053534343512,
      "grad_norm": 18.375,
      "learning_rate": 1.904391969848473e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 4132744.0,
      "step": 9496
    },
    {
      "epoch": 0.7304822705945696,
      "grad_norm": 12.3125,
      "learning_rate": 1.9042765941081455e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9964221715927124,
      "num_tokens": 4133304.0,
      "step": 9497
    },
    {
      "epoch": 0.7305591877547881,
      "grad_norm": 35.75,
      "learning_rate": 1.904161218367818e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9971014261245728,
      "num_tokens": 4133652.0,
      "step": 9498
    },
    {
      "epoch": 0.7306361049150065,
      "grad_norm": 15.5,
      "learning_rate": 1.9040458426274902e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 4133948.0,
      "step": 9499
    },
    {
      "epoch": 0.7307130220752249,
      "grad_norm": 46.0,
      "learning_rate": 1.9039304668871627e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4134224.0,
      "step": 9500
    },
    {
      "epoch": 0.7307899392354434,
      "grad_norm": 64.5,
      "learning_rate": 1.903815091146835e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 4134586.0,
      "step": 9501
    },
    {
      "epoch": 0.7308668563956618,
      "grad_norm": 6.28125,
      "learning_rate": 1.9036997154065072e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 4134920.0,
      "step": 9502
    },
    {
      "epoch": 0.7309437735558804,
      "grad_norm": 14.5625,
      "learning_rate": 1.9035843396661795e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 4135210.0,
      "step": 9503
    },
    {
      "epoch": 0.7310206907160988,
      "grad_norm": 31.75,
      "learning_rate": 1.903468963925852e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 4135526.0,
      "step": 9504
    },
    {
      "epoch": 0.7310976078763172,
      "grad_norm": 8.3125,
      "learning_rate": 1.9033535881855242e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 4136022.0,
      "step": 9505
    },
    {
      "epoch": 0.7311745250365357,
      "grad_norm": 24.125,
      "learning_rate": 1.9032382124451967e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4136300.0,
      "step": 9506
    },
    {
      "epoch": 0.7312514421967541,
      "grad_norm": 9.75,
      "learning_rate": 1.9031228367048687e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4136644.0,
      "step": 9507
    },
    {
      "epoch": 0.7313283593569725,
      "grad_norm": 40.75,
      "learning_rate": 1.9030074609645412e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4136982.0,
      "step": 9508
    },
    {
      "epoch": 0.731405276517191,
      "grad_norm": 5.75,
      "learning_rate": 1.9028920852242134e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 4137506.0,
      "step": 9509
    },
    {
      "epoch": 0.7314821936774094,
      "grad_norm": 6.9375,
      "learning_rate": 1.902776709483886e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9968652129173279,
      "num_tokens": 4137826.0,
      "step": 9510
    },
    {
      "epoch": 0.7315591108376279,
      "grad_norm": 15.125,
      "learning_rate": 1.9026613337435584e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 4138636.0,
      "step": 9511
    },
    {
      "epoch": 0.7316360279978463,
      "grad_norm": 5.21875,
      "learning_rate": 1.9025459580032306e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9968652129173279,
      "num_tokens": 4138956.0,
      "step": 9512
    },
    {
      "epoch": 0.7317129451580647,
      "grad_norm": 12.5,
      "learning_rate": 1.9024305822629027e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4139540.0,
      "step": 9513
    },
    {
      "epoch": 0.7317898623182832,
      "grad_norm": 45.0,
      "learning_rate": 1.9023152065225752e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4139876.0,
      "step": 9514
    },
    {
      "epoch": 0.7318667794785017,
      "grad_norm": 17.0,
      "learning_rate": 1.9021998307822476e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4140436.0,
      "step": 9515
    },
    {
      "epoch": 0.73194369663872,
      "grad_norm": 25.875,
      "learning_rate": 1.9020844550419199e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4140746.0,
      "step": 9516
    },
    {
      "epoch": 0.7320206137989386,
      "grad_norm": 35.0,
      "learning_rate": 1.9019690793015924e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 4141108.0,
      "step": 9517
    },
    {
      "epoch": 0.732097530959157,
      "grad_norm": 10.8125,
      "learning_rate": 1.9018537035612644e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 4141448.0,
      "step": 9518
    },
    {
      "epoch": 0.7321744481193755,
      "grad_norm": 15.4375,
      "learning_rate": 1.9017383278209369e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9972222447395325,
      "num_tokens": 4141810.0,
      "step": 9519
    },
    {
      "epoch": 0.7322513652795939,
      "grad_norm": 13.0625,
      "learning_rate": 1.9016229520806091e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4142356.0,
      "step": 9520
    },
    {
      "epoch": 0.7323282824398123,
      "grad_norm": 5.25,
      "learning_rate": 1.9015075763402816e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4142904.0,
      "step": 9521
    },
    {
      "epoch": 0.7324051996000308,
      "grad_norm": 6.96875,
      "learning_rate": 1.901392200599954e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 4143730.0,
      "step": 9522
    },
    {
      "epoch": 0.7324821167602492,
      "grad_norm": 2.3125,
      "learning_rate": 1.9012768248596263e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4144290.0,
      "step": 9523
    },
    {
      "epoch": 0.7325590339204676,
      "grad_norm": 5.09375,
      "learning_rate": 1.9011614491192984e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 4144882.0,
      "step": 9524
    },
    {
      "epoch": 0.7326359510806861,
      "grad_norm": 36.5,
      "learning_rate": 1.9010460733789708e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4145210.0,
      "step": 9525
    },
    {
      "epoch": 0.7327128682409045,
      "grad_norm": 4.09375,
      "learning_rate": 1.9009306976386433e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 4145800.0,
      "step": 9526
    },
    {
      "epoch": 0.7327897854011229,
      "grad_norm": 10.625,
      "learning_rate": 1.9008153218983156e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 4146116.0,
      "step": 9527
    },
    {
      "epoch": 0.7328667025613415,
      "grad_norm": 12.75,
      "learning_rate": 1.900699946157988e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9970674514770508,
      "num_tokens": 4146458.0,
      "step": 9528
    },
    {
      "epoch": 0.7329436197215599,
      "grad_norm": 66.0,
      "learning_rate": 1.90058457041766e-06,
      "loss": 0.0511,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4146736.0,
      "step": 9529
    },
    {
      "epoch": 0.7330205368817784,
      "grad_norm": 4.625,
      "learning_rate": 1.9004691946773326e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 4147292.0,
      "step": 9530
    },
    {
      "epoch": 0.7330974540419968,
      "grad_norm": 59.25,
      "learning_rate": 1.9003538189370048e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4147602.0,
      "step": 9531
    },
    {
      "epoch": 0.7331743712022152,
      "grad_norm": 11.75,
      "learning_rate": 1.9002384431966773e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4148188.0,
      "step": 9532
    },
    {
      "epoch": 0.7332512883624337,
      "grad_norm": 15.5,
      "learning_rate": 1.9001230674563498e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4148756.0,
      "step": 9533
    },
    {
      "epoch": 0.7333282055226521,
      "grad_norm": 14.8125,
      "learning_rate": 1.9000076917160218e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4149320.0,
      "step": 9534
    },
    {
      "epoch": 0.7334051226828705,
      "grad_norm": 7.4375,
      "learning_rate": 1.899892315975694e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4149900.0,
      "step": 9535
    },
    {
      "epoch": 0.733482039843089,
      "grad_norm": 8.0625,
      "learning_rate": 1.8997769402353665e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4150448.0,
      "step": 9536
    },
    {
      "epoch": 0.7335589570033074,
      "grad_norm": 7.90625,
      "learning_rate": 1.899661564495039e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 4151004.0,
      "step": 9537
    },
    {
      "epoch": 0.7336358741635259,
      "grad_norm": 16.125,
      "learning_rate": 1.8995461887547113e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4151312.0,
      "step": 9538
    },
    {
      "epoch": 0.7337127913237443,
      "grad_norm": 40.25,
      "learning_rate": 1.8994308130143837e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4151622.0,
      "step": 9539
    },
    {
      "epoch": 0.7337897084839627,
      "grad_norm": 12.1875,
      "learning_rate": 1.8993154372740558e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 4152190.0,
      "step": 9540
    },
    {
      "epoch": 0.7338666256441813,
      "grad_norm": 14.5625,
      "learning_rate": 1.8992000615337282e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4152760.0,
      "step": 9541
    },
    {
      "epoch": 0.7339435428043997,
      "grad_norm": 42.25,
      "learning_rate": 1.8990846857934005e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4153042.0,
      "step": 9542
    },
    {
      "epoch": 0.7340204599646181,
      "grad_norm": 5.84375,
      "learning_rate": 1.898969310053073e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4153584.0,
      "step": 9543
    },
    {
      "epoch": 0.7340973771248366,
      "grad_norm": 21.125,
      "learning_rate": 1.8988539343127454e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9867109656333923,
      "num_tokens": 4153886.0,
      "step": 9544
    },
    {
      "epoch": 0.734174294285055,
      "grad_norm": 28.375,
      "learning_rate": 1.8987385585724175e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4154220.0,
      "step": 9545
    },
    {
      "epoch": 0.7342512114452734,
      "grad_norm": 36.5,
      "learning_rate": 1.8986231828320897e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4154566.0,
      "step": 9546
    },
    {
      "epoch": 0.7343281286054919,
      "grad_norm": 9.25,
      "learning_rate": 1.8985078070917622e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4154902.0,
      "step": 9547
    },
    {
      "epoch": 0.7344050457657103,
      "grad_norm": 32.5,
      "learning_rate": 1.8983924313514347e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9890310764312744,
      "num_tokens": 4155450.0,
      "step": 9548
    },
    {
      "epoch": 0.7344819629259288,
      "grad_norm": 6.78125,
      "learning_rate": 1.898277055611107e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4155998.0,
      "step": 9549
    },
    {
      "epoch": 0.7345588800861472,
      "grad_norm": 31.625,
      "learning_rate": 1.8981616798707794e-06,
      "loss": 0.0395,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 4156286.0,
      "step": 9550
    },
    {
      "epoch": 0.7346357972463656,
      "grad_norm": 6.5625,
      "learning_rate": 1.8980463041304515e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4156832.0,
      "step": 9551
    },
    {
      "epoch": 0.7347127144065841,
      "grad_norm": 7.03125,
      "learning_rate": 1.897930928390124e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4157148.0,
      "step": 9552
    },
    {
      "epoch": 0.7347896315668025,
      "grad_norm": 67.5,
      "learning_rate": 1.8978155526497962e-06,
      "loss": 0.0561,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 4157458.0,
      "step": 9553
    },
    {
      "epoch": 0.734866548727021,
      "grad_norm": 47.25,
      "learning_rate": 1.8977001769094687e-06,
      "loss": 0.0448,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4157762.0,
      "step": 9554
    },
    {
      "epoch": 0.7349434658872395,
      "grad_norm": 27.125,
      "learning_rate": 1.897584801169141e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 4158278.0,
      "step": 9555
    },
    {
      "epoch": 0.7350203830474579,
      "grad_norm": 15.375,
      "learning_rate": 1.8974694254288132e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 4158868.0,
      "step": 9556
    },
    {
      "epoch": 0.7350973002076764,
      "grad_norm": 4.375,
      "learning_rate": 1.8973540496884854e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4159408.0,
      "step": 9557
    },
    {
      "epoch": 0.7351742173678948,
      "grad_norm": 4.1875,
      "learning_rate": 1.897238673948158e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4159950.0,
      "step": 9558
    },
    {
      "epoch": 0.7352511345281132,
      "grad_norm": 58.75,
      "learning_rate": 1.8971232982078302e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4160294.0,
      "step": 9559
    },
    {
      "epoch": 0.7353280516883317,
      "grad_norm": 10.4375,
      "learning_rate": 1.8970079224675026e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4160812.0,
      "step": 9560
    },
    {
      "epoch": 0.7354049688485501,
      "grad_norm": 43.5,
      "learning_rate": 1.8968925467271747e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4161120.0,
      "step": 9561
    },
    {
      "epoch": 0.7354818860087685,
      "grad_norm": 11.375,
      "learning_rate": 1.8967771709868471e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4161660.0,
      "step": 9562
    },
    {
      "epoch": 0.735558803168987,
      "grad_norm": 19.125,
      "learning_rate": 1.8966617952465194e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4161972.0,
      "step": 9563
    },
    {
      "epoch": 0.7356357203292054,
      "grad_norm": 22.875,
      "learning_rate": 1.8965464195061919e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 4162520.0,
      "step": 9564
    },
    {
      "epoch": 0.7357126374894238,
      "grad_norm": 13.0,
      "learning_rate": 1.8964310437658643e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4163302.0,
      "step": 9565
    },
    {
      "epoch": 0.7357895546496424,
      "grad_norm": 51.0,
      "learning_rate": 1.8963156680255366e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 4163560.0,
      "step": 9566
    },
    {
      "epoch": 0.7358664718098608,
      "grad_norm": 22.625,
      "learning_rate": 1.8962002922852086e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4163820.0,
      "step": 9567
    },
    {
      "epoch": 0.7359433889700793,
      "grad_norm": 9.9375,
      "learning_rate": 1.8960849165448811e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 4164168.0,
      "step": 9568
    },
    {
      "epoch": 0.7360203061302977,
      "grad_norm": 36.25,
      "learning_rate": 1.8959695408045536e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9965928196907043,
      "num_tokens": 4164756.0,
      "step": 9569
    },
    {
      "epoch": 0.7360972232905161,
      "grad_norm": 57.5,
      "learning_rate": 1.8958541650642258e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4165086.0,
      "step": 9570
    },
    {
      "epoch": 0.7361741404507346,
      "grad_norm": 9.3125,
      "learning_rate": 1.8957387893238983e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 4165402.0,
      "step": 9571
    },
    {
      "epoch": 0.736251057610953,
      "grad_norm": 12.625,
      "learning_rate": 1.8956234135835704e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 4165694.0,
      "step": 9572
    },
    {
      "epoch": 0.7363279747711714,
      "grad_norm": 70.5,
      "learning_rate": 1.8955080378432428e-06,
      "loss": 0.0426,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 4166028.0,
      "step": 9573
    },
    {
      "epoch": 0.7364048919313899,
      "grad_norm": 15.5,
      "learning_rate": 1.895392662102915e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 4166608.0,
      "step": 9574
    },
    {
      "epoch": 0.7364818090916083,
      "grad_norm": 27.5,
      "learning_rate": 1.8952772863625876e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4166932.0,
      "step": 9575
    },
    {
      "epoch": 0.7365587262518268,
      "grad_norm": 14.3125,
      "learning_rate": 1.89516191062226e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 4167476.0,
      "step": 9576
    },
    {
      "epoch": 0.7366356434120452,
      "grad_norm": 5.4375,
      "learning_rate": 1.8950465348819323e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 4168052.0,
      "step": 9577
    },
    {
      "epoch": 0.7367125605722636,
      "grad_norm": 9.6875,
      "learning_rate": 1.8949311591416043e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 4168592.0,
      "step": 9578
    },
    {
      "epoch": 0.7367894777324822,
      "grad_norm": 21.75,
      "learning_rate": 1.8948157834012768e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 4168942.0,
      "step": 9579
    },
    {
      "epoch": 0.7368663948927006,
      "grad_norm": 9.75,
      "learning_rate": 1.8947004076609493e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9858155846595764,
      "num_tokens": 4169226.0,
      "step": 9580
    },
    {
      "epoch": 0.736943312052919,
      "grad_norm": 36.5,
      "learning_rate": 1.8945850319206215e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 4169484.0,
      "step": 9581
    },
    {
      "epoch": 0.7370202292131375,
      "grad_norm": 41.0,
      "learning_rate": 1.894469656180294e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4169816.0,
      "step": 9582
    },
    {
      "epoch": 0.7370971463733559,
      "grad_norm": 7.1875,
      "learning_rate": 1.894354280439966e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4170362.0,
      "step": 9583
    },
    {
      "epoch": 0.7371740635335743,
      "grad_norm": 19.375,
      "learning_rate": 1.8942389046996385e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 4170722.0,
      "step": 9584
    },
    {
      "epoch": 0.7372509806937928,
      "grad_norm": 13.3125,
      "learning_rate": 1.8941235289593108e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 4170984.0,
      "step": 9585
    },
    {
      "epoch": 0.7373278978540112,
      "grad_norm": 25.25,
      "learning_rate": 1.8940081532189832e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4171294.0,
      "step": 9586
    },
    {
      "epoch": 0.7374048150142297,
      "grad_norm": 35.75,
      "learning_rate": 1.8938927774786557e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4171574.0,
      "step": 9587
    },
    {
      "epoch": 0.7374817321744481,
      "grad_norm": 5.875,
      "learning_rate": 1.893777401738328e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4171858.0,
      "step": 9588
    },
    {
      "epoch": 0.7375586493346665,
      "grad_norm": 25.5,
      "learning_rate": 1.893662025998e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4172134.0,
      "step": 9589
    },
    {
      "epoch": 0.737635566494885,
      "grad_norm": 20.375,
      "learning_rate": 1.8935466502576725e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4172444.0,
      "step": 9590
    },
    {
      "epoch": 0.7377124836551034,
      "grad_norm": 26.25,
      "learning_rate": 1.893431274517345e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 4172774.0,
      "step": 9591
    },
    {
      "epoch": 0.7377894008153218,
      "grad_norm": 46.25,
      "learning_rate": 1.8933158987770172e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9830028414726257,
      "num_tokens": 4173128.0,
      "step": 9592
    },
    {
      "epoch": 0.7378663179755404,
      "grad_norm": 14.625,
      "learning_rate": 1.8932005230366897e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4173410.0,
      "step": 9593
    },
    {
      "epoch": 0.7379432351357588,
      "grad_norm": 15.0,
      "learning_rate": 1.8930851472963617e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4173716.0,
      "step": 9594
    },
    {
      "epoch": 0.7380201522959773,
      "grad_norm": 40.0,
      "learning_rate": 1.8929697715560342e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4173994.0,
      "step": 9595
    },
    {
      "epoch": 0.7380970694561957,
      "grad_norm": 32.75,
      "learning_rate": 1.8928543958157065e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 4174528.0,
      "step": 9596
    },
    {
      "epoch": 0.7381739866164141,
      "grad_norm": 36.75,
      "learning_rate": 1.892739020075379e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4174806.0,
      "step": 9597
    },
    {
      "epoch": 0.7382509037766326,
      "grad_norm": 38.25,
      "learning_rate": 1.8926236443350514e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 4175392.0,
      "step": 9598
    },
    {
      "epoch": 0.738327820936851,
      "grad_norm": 35.5,
      "learning_rate": 1.8925082685947234e-06,
      "loss": 0.0434,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4175724.0,
      "step": 9599
    },
    {
      "epoch": 0.7384047380970694,
      "grad_norm": 15.625,
      "learning_rate": 1.8923928928543957e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4176288.0,
      "step": 9600
    },
    {
      "epoch": 0.7384816552572879,
      "grad_norm": 15.9375,
      "learning_rate": 1.8922775171140682e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 4176596.0,
      "step": 9601
    },
    {
      "epoch": 0.7385585724175063,
      "grad_norm": 36.25,
      "learning_rate": 1.8921621413737406e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4176936.0,
      "step": 9602
    },
    {
      "epoch": 0.7386354895777247,
      "grad_norm": 43.5,
      "learning_rate": 1.892046765633413e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4177500.0,
      "step": 9603
    },
    {
      "epoch": 0.7387124067379433,
      "grad_norm": 14.4375,
      "learning_rate": 1.8919313898930854e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4178048.0,
      "step": 9604
    },
    {
      "epoch": 0.7387893238981617,
      "grad_norm": 25.375,
      "learning_rate": 1.8918160141527574e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 4178330.0,
      "step": 9605
    },
    {
      "epoch": 0.7388662410583802,
      "grad_norm": 10.625,
      "learning_rate": 1.8917006384124299e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 4178632.0,
      "step": 9606
    },
    {
      "epoch": 0.7389431582185986,
      "grad_norm": 8.9375,
      "learning_rate": 1.8915852626721021e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4179174.0,
      "step": 9607
    },
    {
      "epoch": 0.739020075378817,
      "grad_norm": 3.0,
      "learning_rate": 1.8914698869317746e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4179744.0,
      "step": 9608
    },
    {
      "epoch": 0.7390969925390355,
      "grad_norm": 5.03125,
      "learning_rate": 1.8913545111914469e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4180290.0,
      "step": 9609
    },
    {
      "epoch": 0.7391739096992539,
      "grad_norm": 4.59375,
      "learning_rate": 1.8912391354511191e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 4181070.0,
      "step": 9610
    },
    {
      "epoch": 0.7392508268594723,
      "grad_norm": 14.4375,
      "learning_rate": 1.8911237597107914e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 4181642.0,
      "step": 9611
    },
    {
      "epoch": 0.7393277440196908,
      "grad_norm": 14.125,
      "learning_rate": 1.8910083839704639e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4182222.0,
      "step": 9612
    },
    {
      "epoch": 0.7394046611799092,
      "grad_norm": 21.25,
      "learning_rate": 1.8908930082301361e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4182504.0,
      "step": 9613
    },
    {
      "epoch": 0.7394815783401277,
      "grad_norm": 30.25,
      "learning_rate": 1.8907776324898086e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4182782.0,
      "step": 9614
    },
    {
      "epoch": 0.7395584955003461,
      "grad_norm": 7.59375,
      "learning_rate": 1.890662256749481e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4183356.0,
      "step": 9615
    },
    {
      "epoch": 0.7396354126605645,
      "grad_norm": 10.5,
      "learning_rate": 1.890546881009153e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4183674.0,
      "step": 9616
    },
    {
      "epoch": 0.739712329820783,
      "grad_norm": 24.75,
      "learning_rate": 1.8904315052688254e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4184010.0,
      "step": 9617
    },
    {
      "epoch": 0.7397892469810015,
      "grad_norm": 20.125,
      "learning_rate": 1.8903161295284978e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4184558.0,
      "step": 9618
    },
    {
      "epoch": 0.7398661641412199,
      "grad_norm": 29.125,
      "learning_rate": 1.8902007537881703e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 4185094.0,
      "step": 9619
    },
    {
      "epoch": 0.7399430813014384,
      "grad_norm": 54.75,
      "learning_rate": 1.8900853780478426e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4185450.0,
      "step": 9620
    },
    {
      "epoch": 0.7400199984616568,
      "grad_norm": 6.03125,
      "learning_rate": 1.8899700023075146e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4185992.0,
      "step": 9621
    },
    {
      "epoch": 0.7400969156218753,
      "grad_norm": 20.125,
      "learning_rate": 1.889854626567187e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4186324.0,
      "step": 9622
    },
    {
      "epoch": 0.7401738327820937,
      "grad_norm": 9.4375,
      "learning_rate": 1.8897392508268595e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 4187084.0,
      "step": 9623
    },
    {
      "epoch": 0.7402507499423121,
      "grad_norm": 26.5,
      "learning_rate": 1.8896238750865318e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4187404.0,
      "step": 9624
    },
    {
      "epoch": 0.7403276671025306,
      "grad_norm": 6.25,
      "learning_rate": 1.8895084993462043e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 4187966.0,
      "step": 9625
    },
    {
      "epoch": 0.740404584262749,
      "grad_norm": 72.5,
      "learning_rate": 1.8893931236058767e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4188312.0,
      "step": 9626
    },
    {
      "epoch": 0.7404815014229674,
      "grad_norm": 25.25,
      "learning_rate": 1.8892777478655488e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4188622.0,
      "step": 9627
    },
    {
      "epoch": 0.7405584185831859,
      "grad_norm": 7.1875,
      "learning_rate": 1.889162372125221e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 4189436.0,
      "step": 9628
    },
    {
      "epoch": 0.7406353357434043,
      "grad_norm": 34.5,
      "learning_rate": 1.8890469963848935e-06,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4189692.0,
      "step": 9629
    },
    {
      "epoch": 0.7407122529036227,
      "grad_norm": 17.0,
      "learning_rate": 1.888931620644566e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4190026.0,
      "step": 9630
    },
    {
      "epoch": 0.7407891700638413,
      "grad_norm": 56.25,
      "learning_rate": 1.8888162449042382e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4190358.0,
      "step": 9631
    },
    {
      "epoch": 0.7408660872240597,
      "grad_norm": 30.375,
      "learning_rate": 1.8887008691639103e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 4190644.0,
      "step": 9632
    },
    {
      "epoch": 0.7409430043842782,
      "grad_norm": 18.5,
      "learning_rate": 1.8885854934235828e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4190966.0,
      "step": 9633
    },
    {
      "epoch": 0.7410199215444966,
      "grad_norm": 12.3125,
      "learning_rate": 1.8884701176832552e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4191226.0,
      "step": 9634
    },
    {
      "epoch": 0.741096838704715,
      "grad_norm": 20.625,
      "learning_rate": 1.8883547419429275e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 4191768.0,
      "step": 9635
    },
    {
      "epoch": 0.7411737558649335,
      "grad_norm": 18.5,
      "learning_rate": 1.8882393662026e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4192104.0,
      "step": 9636
    },
    {
      "epoch": 0.7412506730251519,
      "grad_norm": 21.75,
      "learning_rate": 1.888123990462272e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 4192438.0,
      "step": 9637
    },
    {
      "epoch": 0.7413275901853703,
      "grad_norm": 14.625,
      "learning_rate": 1.8880086147219445e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 4192746.0,
      "step": 9638
    },
    {
      "epoch": 0.7414045073455888,
      "grad_norm": 83.0,
      "learning_rate": 1.8878932389816167e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 4193112.0,
      "step": 9639
    },
    {
      "epoch": 0.7414814245058072,
      "grad_norm": 48.0,
      "learning_rate": 1.8877778632412892e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 4193472.0,
      "step": 9640
    },
    {
      "epoch": 0.7415583416660257,
      "grad_norm": 26.375,
      "learning_rate": 1.8876624875009617e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4193808.0,
      "step": 9641
    },
    {
      "epoch": 0.7416352588262441,
      "grad_norm": 27.375,
      "learning_rate": 1.887547111760634e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4194154.0,
      "step": 9642
    },
    {
      "epoch": 0.7417121759864626,
      "grad_norm": 51.5,
      "learning_rate": 1.887431736020306e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4194518.0,
      "step": 9643
    },
    {
      "epoch": 0.7417890931466811,
      "grad_norm": 38.75,
      "learning_rate": 1.8873163602799784e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 4195120.0,
      "step": 9644
    },
    {
      "epoch": 0.7418660103068995,
      "grad_norm": 12.4375,
      "learning_rate": 1.887200984539651e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 4195644.0,
      "step": 9645
    },
    {
      "epoch": 0.7419429274671179,
      "grad_norm": 44.25,
      "learning_rate": 1.8870856087993232e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 4195962.0,
      "step": 9646
    },
    {
      "epoch": 0.7420198446273364,
      "grad_norm": 43.75,
      "learning_rate": 1.8869702330589956e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4196290.0,
      "step": 9647
    },
    {
      "epoch": 0.7420967617875548,
      "grad_norm": 9.0625,
      "learning_rate": 1.8868548573186677e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 4196602.0,
      "step": 9648
    },
    {
      "epoch": 0.7421736789477732,
      "grad_norm": 13.9375,
      "learning_rate": 1.8867394815783402e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4196900.0,
      "step": 9649
    },
    {
      "epoch": 0.7422505961079917,
      "grad_norm": 5.21875,
      "learning_rate": 1.8866241058380124e-06,
      "loss": 0.0068,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 4197704.0,
      "step": 9650
    },
    {
      "epoch": 0.7423275132682101,
      "grad_norm": 6.375,
      "learning_rate": 1.8865087300976849e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4198278.0,
      "step": 9651
    },
    {
      "epoch": 0.7424044304284286,
      "grad_norm": 4.25,
      "learning_rate": 1.8863933543573574e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 4199114.0,
      "step": 9652
    },
    {
      "epoch": 0.742481347588647,
      "grad_norm": 47.5,
      "learning_rate": 1.8862779786170296e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4199396.0,
      "step": 9653
    },
    {
      "epoch": 0.7425582647488654,
      "grad_norm": 68.0,
      "learning_rate": 1.8861626028767017e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4199724.0,
      "step": 9654
    },
    {
      "epoch": 0.742635181909084,
      "grad_norm": 12.625,
      "learning_rate": 1.8860472271363741e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 4200014.0,
      "step": 9655
    },
    {
      "epoch": 0.7427120990693024,
      "grad_norm": 19.5,
      "learning_rate": 1.8859318513960466e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 4200336.0,
      "step": 9656
    },
    {
      "epoch": 0.7427890162295208,
      "grad_norm": 46.75,
      "learning_rate": 1.8858164756557189e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 4200644.0,
      "step": 9657
    },
    {
      "epoch": 0.7428659333897393,
      "grad_norm": 13.1875,
      "learning_rate": 1.8857010999153913e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 4201164.0,
      "step": 9658
    },
    {
      "epoch": 0.7429428505499577,
      "grad_norm": 21.625,
      "learning_rate": 1.8855857241750634e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4201504.0,
      "step": 9659
    },
    {
      "epoch": 0.7430197677101762,
      "grad_norm": 7.0,
      "learning_rate": 1.8854703484347358e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 4201860.0,
      "step": 9660
    },
    {
      "epoch": 0.7430966848703946,
      "grad_norm": 12.1875,
      "learning_rate": 1.8853549726944081e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4202178.0,
      "step": 9661
    },
    {
      "epoch": 0.743173602030613,
      "grad_norm": 7.15625,
      "learning_rate": 1.8852395969540806e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4202462.0,
      "step": 9662
    },
    {
      "epoch": 0.7432505191908315,
      "grad_norm": 19.25,
      "learning_rate": 1.885124221213753e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4203010.0,
      "step": 9663
    },
    {
      "epoch": 0.7433274363510499,
      "grad_norm": 16.375,
      "learning_rate": 1.8850088454734253e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 4203818.0,
      "step": 9664
    },
    {
      "epoch": 0.7434043535112683,
      "grad_norm": 21.75,
      "learning_rate": 1.8848934697330974e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4204124.0,
      "step": 9665
    },
    {
      "epoch": 0.7434812706714868,
      "grad_norm": 26.125,
      "learning_rate": 1.8847780939927698e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4204378.0,
      "step": 9666
    },
    {
      "epoch": 0.7435581878317052,
      "grad_norm": 16.375,
      "learning_rate": 1.8846627182524423e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4204928.0,
      "step": 9667
    },
    {
      "epoch": 0.7436351049919236,
      "grad_norm": 16.875,
      "learning_rate": 1.8845473425121145e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 4205258.0,
      "step": 9668
    },
    {
      "epoch": 0.7437120221521422,
      "grad_norm": 19.75,
      "learning_rate": 1.884431966771787e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4205568.0,
      "step": 9669
    },
    {
      "epoch": 0.7437889393123606,
      "grad_norm": 113.0,
      "learning_rate": 1.884316591031459e-06,
      "loss": 0.0523,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 4205892.0,
      "step": 9670
    },
    {
      "epoch": 0.7438658564725791,
      "grad_norm": 28.375,
      "learning_rate": 1.8842012152911315e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 4206218.0,
      "step": 9671
    },
    {
      "epoch": 0.7439427736327975,
      "grad_norm": 7.625,
      "learning_rate": 1.8840858395508038e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 4206794.0,
      "step": 9672
    },
    {
      "epoch": 0.7440196907930159,
      "grad_norm": 24.75,
      "learning_rate": 1.8839704638104763e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4207134.0,
      "step": 9673
    },
    {
      "epoch": 0.7440966079532344,
      "grad_norm": 17.375,
      "learning_rate": 1.8838550880701485e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 4207450.0,
      "step": 9674
    },
    {
      "epoch": 0.7441735251134528,
      "grad_norm": 9.3125,
      "learning_rate": 1.8837397123298208e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4207786.0,
      "step": 9675
    },
    {
      "epoch": 0.7442504422736712,
      "grad_norm": 12.8125,
      "learning_rate": 1.883624336589493e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 4208358.0,
      "step": 9676
    },
    {
      "epoch": 0.7443273594338897,
      "grad_norm": 4.375,
      "learning_rate": 1.8835089608491655e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 4208948.0,
      "step": 9677
    },
    {
      "epoch": 0.7444042765941081,
      "grad_norm": 30.25,
      "learning_rate": 1.8833935851088378e-06,
      "loss": 0.0407,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4209230.0,
      "step": 9678
    },
    {
      "epoch": 0.7444811937543266,
      "grad_norm": 16.625,
      "learning_rate": 1.8832782093685102e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 4209564.0,
      "step": 9679
    },
    {
      "epoch": 0.744558110914545,
      "grad_norm": 17.125,
      "learning_rate": 1.8831628336281827e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4209874.0,
      "step": 9680
    },
    {
      "epoch": 0.7446350280747634,
      "grad_norm": 24.25,
      "learning_rate": 1.8830474578878548e-06,
      "loss": 0.0544,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4210150.0,
      "step": 9681
    },
    {
      "epoch": 0.744711945234982,
      "grad_norm": 20.75,
      "learning_rate": 1.882932082147527e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 4210666.0,
      "step": 9682
    },
    {
      "epoch": 0.7447888623952004,
      "grad_norm": 7.6875,
      "learning_rate": 1.8828167064071995e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4211210.0,
      "step": 9683
    },
    {
      "epoch": 0.7448657795554188,
      "grad_norm": 6.8125,
      "learning_rate": 1.882701330666872e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4211778.0,
      "step": 9684
    },
    {
      "epoch": 0.7449426967156373,
      "grad_norm": 10.0,
      "learning_rate": 1.8825859549265442e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 4212036.0,
      "step": 9685
    },
    {
      "epoch": 0.7450196138758557,
      "grad_norm": 58.25,
      "learning_rate": 1.8824705791862163e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 4212402.0,
      "step": 9686
    },
    {
      "epoch": 0.7450965310360741,
      "grad_norm": 55.5,
      "learning_rate": 1.8823552034458887e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4212678.0,
      "step": 9687
    },
    {
      "epoch": 0.7451734481962926,
      "grad_norm": 24.75,
      "learning_rate": 1.8822398277055612e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4212984.0,
      "step": 9688
    },
    {
      "epoch": 0.745250365356511,
      "grad_norm": 21.0,
      "learning_rate": 1.8821244519652335e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 4213318.0,
      "step": 9689
    },
    {
      "epoch": 0.7453272825167295,
      "grad_norm": 5.90625,
      "learning_rate": 1.882009076224906e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4213600.0,
      "step": 9690
    },
    {
      "epoch": 0.7454041996769479,
      "grad_norm": 21.75,
      "learning_rate": 1.8818937004845784e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4214146.0,
      "step": 9691
    },
    {
      "epoch": 0.7454811168371663,
      "grad_norm": 23.875,
      "learning_rate": 1.8817783247442504e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9947826266288757,
      "num_tokens": 4214722.0,
      "step": 9692
    },
    {
      "epoch": 0.7455580339973849,
      "grad_norm": 74.0,
      "learning_rate": 1.8816629490039227e-06,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4215032.0,
      "step": 9693
    },
    {
      "epoch": 0.7456349511576033,
      "grad_norm": 13.0625,
      "learning_rate": 1.8815475732635952e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4215354.0,
      "step": 9694
    },
    {
      "epoch": 0.7457118683178217,
      "grad_norm": 63.5,
      "learning_rate": 1.8814321975232676e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4215632.0,
      "step": 9695
    },
    {
      "epoch": 0.7457887854780402,
      "grad_norm": 10.4375,
      "learning_rate": 1.8813168217829399e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4216178.0,
      "step": 9696
    },
    {
      "epoch": 0.7458657026382586,
      "grad_norm": 9.125,
      "learning_rate": 1.881201446042612e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 4216998.0,
      "step": 9697
    },
    {
      "epoch": 0.7459426197984771,
      "grad_norm": 25.5,
      "learning_rate": 1.8810860703022844e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4217344.0,
      "step": 9698
    },
    {
      "epoch": 0.7460195369586955,
      "grad_norm": 7.03125,
      "learning_rate": 1.8809706945619569e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4217908.0,
      "step": 9699
    },
    {
      "epoch": 0.7460964541189139,
      "grad_norm": 30.5,
      "learning_rate": 1.8808553188216291e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4218480.0,
      "step": 9700
    },
    {
      "epoch": 0.7461733712791324,
      "grad_norm": 26.75,
      "learning_rate": 1.8807399430813016e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4218820.0,
      "step": 9701
    },
    {
      "epoch": 0.7462502884393508,
      "grad_norm": 51.5,
      "learning_rate": 1.8806245673409737e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 4219128.0,
      "step": 9702
    },
    {
      "epoch": 0.7463272055995692,
      "grad_norm": 28.75,
      "learning_rate": 1.8805091916006461e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4219424.0,
      "step": 9703
    },
    {
      "epoch": 0.7464041227597877,
      "grad_norm": 4.59375,
      "learning_rate": 1.8803938158603184e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4219942.0,
      "step": 9704
    },
    {
      "epoch": 0.7464810399200061,
      "grad_norm": 5.125,
      "learning_rate": 1.8802784401199909e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4220264.0,
      "step": 9705
    },
    {
      "epoch": 0.7465579570802245,
      "grad_norm": 2.765625,
      "learning_rate": 1.8801630643796633e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4220812.0,
      "step": 9706
    },
    {
      "epoch": 0.7466348742404431,
      "grad_norm": 8.875,
      "learning_rate": 1.8800476886393356e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 4221596.0,
      "step": 9707
    },
    {
      "epoch": 0.7467117914006615,
      "grad_norm": 29.75,
      "learning_rate": 1.8799323128990076e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4221936.0,
      "step": 9708
    },
    {
      "epoch": 0.74678870856088,
      "grad_norm": 25.25,
      "learning_rate": 1.87981693715868e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 4222286.0,
      "step": 9709
    },
    {
      "epoch": 0.7468656257210984,
      "grad_norm": 4.75,
      "learning_rate": 1.8797015614183526e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4222848.0,
      "step": 9710
    },
    {
      "epoch": 0.7469425428813168,
      "grad_norm": 32.5,
      "learning_rate": 1.8795861856780248e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4223124.0,
      "step": 9711
    },
    {
      "epoch": 0.7470194600415353,
      "grad_norm": 2.640625,
      "learning_rate": 1.8794708099376973e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4223676.0,
      "step": 9712
    },
    {
      "epoch": 0.7470963772017537,
      "grad_norm": 14.8125,
      "learning_rate": 1.8793554341973693e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4224238.0,
      "step": 9713
    },
    {
      "epoch": 0.7471732943619721,
      "grad_norm": 6.0,
      "learning_rate": 1.8792400584570418e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4224786.0,
      "step": 9714
    },
    {
      "epoch": 0.7472502115221906,
      "grad_norm": 14.8125,
      "learning_rate": 1.879124682716714e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4225122.0,
      "step": 9715
    },
    {
      "epoch": 0.747327128682409,
      "grad_norm": 5.09375,
      "learning_rate": 1.8790093069763865e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4225446.0,
      "step": 9716
    },
    {
      "epoch": 0.7474040458426275,
      "grad_norm": 16.125,
      "learning_rate": 1.878893931236059e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 4225778.0,
      "step": 9717
    },
    {
      "epoch": 0.7474809630028459,
      "grad_norm": 45.5,
      "learning_rate": 1.8787785554957313e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4226102.0,
      "step": 9718
    },
    {
      "epoch": 0.7475578801630643,
      "grad_norm": 27.75,
      "learning_rate": 1.8786631797554033e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4226442.0,
      "step": 9719
    },
    {
      "epoch": 0.7476347973232829,
      "grad_norm": 22.5,
      "learning_rate": 1.8785478040150758e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 4227502.0,
      "step": 9720
    },
    {
      "epoch": 0.7477117144835013,
      "grad_norm": 7.3125,
      "learning_rate": 1.8784324282747483e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 4228070.0,
      "step": 9721
    },
    {
      "epoch": 0.7477886316437197,
      "grad_norm": 10.8125,
      "learning_rate": 1.8783170525344205e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4228614.0,
      "step": 9722
    },
    {
      "epoch": 0.7478655488039382,
      "grad_norm": 38.5,
      "learning_rate": 1.878201676794093e-06,
      "loss": 0.0441,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4228892.0,
      "step": 9723
    },
    {
      "epoch": 0.7479424659641566,
      "grad_norm": 63.0,
      "learning_rate": 1.878086301053765e-06,
      "loss": 0.0441,
      "mean_token_accuracy": 0.9809523820877075,
      "num_tokens": 4229208.0,
      "step": 9724
    },
    {
      "epoch": 0.7480193831243751,
      "grad_norm": 15.5625,
      "learning_rate": 1.8779709253134375e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 4229498.0,
      "step": 9725
    },
    {
      "epoch": 0.7480963002845935,
      "grad_norm": 16.125,
      "learning_rate": 1.8778555495731098e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 4230018.0,
      "step": 9726
    },
    {
      "epoch": 0.7481732174448119,
      "grad_norm": 6.625,
      "learning_rate": 1.8777401738327822e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 4230358.0,
      "step": 9727
    },
    {
      "epoch": 0.7482501346050304,
      "grad_norm": 16.625,
      "learning_rate": 1.8776247980924545e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9910233616828918,
      "num_tokens": 4230916.0,
      "step": 9728
    },
    {
      "epoch": 0.7483270517652488,
      "grad_norm": 22.5,
      "learning_rate": 1.877509422352127e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4231222.0,
      "step": 9729
    },
    {
      "epoch": 0.7484039689254672,
      "grad_norm": 6.34375,
      "learning_rate": 1.877394046611799e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4231522.0,
      "step": 9730
    },
    {
      "epoch": 0.7484808860856857,
      "grad_norm": 30.125,
      "learning_rate": 1.8772786708714715e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4231868.0,
      "step": 9731
    },
    {
      "epoch": 0.7485578032459042,
      "grad_norm": 2.78125,
      "learning_rate": 1.8771632951311437e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4232424.0,
      "step": 9732
    },
    {
      "epoch": 0.7486347204061226,
      "grad_norm": 10.3125,
      "learning_rate": 1.8770479193908162e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4232756.0,
      "step": 9733
    },
    {
      "epoch": 0.7487116375663411,
      "grad_norm": 37.25,
      "learning_rate": 1.8769325436504887e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4233010.0,
      "step": 9734
    },
    {
      "epoch": 0.7487885547265595,
      "grad_norm": 14.3125,
      "learning_rate": 1.8768171679101607e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 4233376.0,
      "step": 9735
    },
    {
      "epoch": 0.748865471886778,
      "grad_norm": 19.375,
      "learning_rate": 1.876701792169833e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9855907559394836,
      "num_tokens": 4233724.0,
      "step": 9736
    },
    {
      "epoch": 0.7489423890469964,
      "grad_norm": 27.125,
      "learning_rate": 1.8765864164295054e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9858356714248657,
      "num_tokens": 4234078.0,
      "step": 9737
    },
    {
      "epoch": 0.7490193062072148,
      "grad_norm": 7.90625,
      "learning_rate": 1.876471040689178e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 4234834.0,
      "step": 9738
    },
    {
      "epoch": 0.7490962233674333,
      "grad_norm": 5.25,
      "learning_rate": 1.8763556649488502e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 4235614.0,
      "step": 9739
    },
    {
      "epoch": 0.7491731405276517,
      "grad_norm": 11.5625,
      "learning_rate": 1.8762402892085222e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4236160.0,
      "step": 9740
    },
    {
      "epoch": 0.7492500576878701,
      "grad_norm": 23.125,
      "learning_rate": 1.8761249134681947e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4236484.0,
      "step": 9741
    },
    {
      "epoch": 0.7493269748480886,
      "grad_norm": 17.75,
      "learning_rate": 1.8760095377278672e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9931740760803223,
      "num_tokens": 4236778.0,
      "step": 9742
    },
    {
      "epoch": 0.749403892008307,
      "grad_norm": 2.875,
      "learning_rate": 1.8758941619875394e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4237296.0,
      "step": 9743
    },
    {
      "epoch": 0.7494808091685256,
      "grad_norm": 8.5,
      "learning_rate": 1.8757787862472119e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4237608.0,
      "step": 9744
    },
    {
      "epoch": 0.749557726328744,
      "grad_norm": 14.1875,
      "learning_rate": 1.8756634105068844e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 4237906.0,
      "step": 9745
    },
    {
      "epoch": 0.7496346434889624,
      "grad_norm": 7.46875,
      "learning_rate": 1.8755480347665564e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4238218.0,
      "step": 9746
    },
    {
      "epoch": 0.7497115606491809,
      "grad_norm": 16.375,
      "learning_rate": 1.8754326590262287e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 4238508.0,
      "step": 9747
    },
    {
      "epoch": 0.7497884778093993,
      "grad_norm": 16.25,
      "learning_rate": 1.8753172832859011e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9983079433441162,
      "num_tokens": 4239100.0,
      "step": 9748
    },
    {
      "epoch": 0.7498653949696177,
      "grad_norm": 17.125,
      "learning_rate": 1.8752019075455736e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 4239620.0,
      "step": 9749
    },
    {
      "epoch": 0.7499423121298362,
      "grad_norm": 15.9375,
      "learning_rate": 1.8750865318052459e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 4239940.0,
      "step": 9750
    },
    {
      "epoch": 0.7500192292900546,
      "grad_norm": 41.0,
      "learning_rate": 1.874971156064918e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9916666746139526,
      "num_tokens": 4240302.0,
      "step": 9751
    },
    {
      "epoch": 0.750096146450273,
      "grad_norm": 16.875,
      "learning_rate": 1.8748557803245904e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4240622.0,
      "step": 9752
    },
    {
      "epoch": 0.7501730636104915,
      "grad_norm": 24.25,
      "learning_rate": 1.8747404045842628e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4240962.0,
      "step": 9753
    },
    {
      "epoch": 0.7502499807707099,
      "grad_norm": 6.0625,
      "learning_rate": 1.874625028843935e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4241508.0,
      "step": 9754
    },
    {
      "epoch": 0.7503268979309284,
      "grad_norm": 18.0,
      "learning_rate": 1.8745096531036076e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 4241846.0,
      "step": 9755
    },
    {
      "epoch": 0.7504038150911468,
      "grad_norm": 10.125,
      "learning_rate": 1.87439427736328e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4242396.0,
      "step": 9756
    },
    {
      "epoch": 0.7504807322513652,
      "grad_norm": 7.0,
      "learning_rate": 1.874278901622952e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 4242722.0,
      "step": 9757
    },
    {
      "epoch": 0.7505576494115838,
      "grad_norm": 4.125,
      "learning_rate": 1.8741635258826243e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4243262.0,
      "step": 9758
    },
    {
      "epoch": 0.7506345665718022,
      "grad_norm": 18.375,
      "learning_rate": 1.8740481501422968e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 4243830.0,
      "step": 9759
    },
    {
      "epoch": 0.7507114837320206,
      "grad_norm": 14.875,
      "learning_rate": 1.8739327744019693e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9948006868362427,
      "num_tokens": 4244408.0,
      "step": 9760
    },
    {
      "epoch": 0.7507884008922391,
      "grad_norm": 7.5,
      "learning_rate": 1.8738173986616415e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 4244970.0,
      "step": 9761
    },
    {
      "epoch": 0.7508653180524575,
      "grad_norm": 46.75,
      "learning_rate": 1.8737020229213136e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4245286.0,
      "step": 9762
    },
    {
      "epoch": 0.750942235212676,
      "grad_norm": 19.875,
      "learning_rate": 1.873586647180986e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4245606.0,
      "step": 9763
    },
    {
      "epoch": 0.7510191523728944,
      "grad_norm": 9.0625,
      "learning_rate": 1.8734712714406585e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4246124.0,
      "step": 9764
    },
    {
      "epoch": 0.7510960695331128,
      "grad_norm": 10.625,
      "learning_rate": 1.8733558957003308e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 4246920.0,
      "step": 9765
    },
    {
      "epoch": 0.7511729866933313,
      "grad_norm": 17.5,
      "learning_rate": 1.8732405199600033e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4247248.0,
      "step": 9766
    },
    {
      "epoch": 0.7512499038535497,
      "grad_norm": 4.84375,
      "learning_rate": 1.8731251442196757e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4247796.0,
      "step": 9767
    },
    {
      "epoch": 0.7513268210137681,
      "grad_norm": 25.5,
      "learning_rate": 1.8730097684793478e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4248140.0,
      "step": 9768
    },
    {
      "epoch": 0.7514037381739866,
      "grad_norm": 32.25,
      "learning_rate": 1.87289439273902e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4248474.0,
      "step": 9769
    },
    {
      "epoch": 0.751480655334205,
      "grad_norm": 18.375,
      "learning_rate": 1.8727790169986925e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4248768.0,
      "step": 9770
    },
    {
      "epoch": 0.7515575724944235,
      "grad_norm": 2.765625,
      "learning_rate": 1.872663641258365e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 4249326.0,
      "step": 9771
    },
    {
      "epoch": 0.751634489654642,
      "grad_norm": 4.21875,
      "learning_rate": 1.8725482655180372e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4249868.0,
      "step": 9772
    },
    {
      "epoch": 0.7517114068148604,
      "grad_norm": 25.75,
      "learning_rate": 1.8724328897777093e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4250190.0,
      "step": 9773
    },
    {
      "epoch": 0.7517883239750789,
      "grad_norm": 7.375,
      "learning_rate": 1.8723175140373817e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.998740553855896,
      "num_tokens": 4250986.0,
      "step": 9774
    },
    {
      "epoch": 0.7518652411352973,
      "grad_norm": 72.0,
      "learning_rate": 1.8722021382970542e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4251314.0,
      "step": 9775
    },
    {
      "epoch": 0.7519421582955157,
      "grad_norm": 13.8125,
      "learning_rate": 1.8720867625567265e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4251860.0,
      "step": 9776
    },
    {
      "epoch": 0.7520190754557342,
      "grad_norm": 16.375,
      "learning_rate": 1.871971386816399e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 4252140.0,
      "step": 9777
    },
    {
      "epoch": 0.7520959926159526,
      "grad_norm": 4.8125,
      "learning_rate": 1.871856011076071e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4252658.0,
      "step": 9778
    },
    {
      "epoch": 0.752172909776171,
      "grad_norm": 14.875,
      "learning_rate": 1.8717406353357435e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4253208.0,
      "step": 9779
    },
    {
      "epoch": 0.7522498269363895,
      "grad_norm": 31.375,
      "learning_rate": 1.8716252595954157e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4253564.0,
      "step": 9780
    },
    {
      "epoch": 0.7523267440966079,
      "grad_norm": 32.75,
      "learning_rate": 1.8715098838550882e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 4253902.0,
      "step": 9781
    },
    {
      "epoch": 0.7524036612568265,
      "grad_norm": 7.96875,
      "learning_rate": 1.8713945081147604e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4254448.0,
      "step": 9782
    },
    {
      "epoch": 0.7524805784170449,
      "grad_norm": 51.25,
      "learning_rate": 1.871279132374433e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4254762.0,
      "step": 9783
    },
    {
      "epoch": 0.7525574955772633,
      "grad_norm": 18.5,
      "learning_rate": 1.871163756634105e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4255080.0,
      "step": 9784
    },
    {
      "epoch": 0.7526344127374818,
      "grad_norm": 51.25,
      "learning_rate": 1.8710483808937774e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4255388.0,
      "step": 9785
    },
    {
      "epoch": 0.7527113298977002,
      "grad_norm": 21.125,
      "learning_rate": 1.8709330051534497e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4255684.0,
      "step": 9786
    },
    {
      "epoch": 0.7527882470579186,
      "grad_norm": 21.75,
      "learning_rate": 1.8708176294131222e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4255996.0,
      "step": 9787
    },
    {
      "epoch": 0.7528651642181371,
      "grad_norm": 28.0,
      "learning_rate": 1.8707022536727946e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4256274.0,
      "step": 9788
    },
    {
      "epoch": 0.7529420813783555,
      "grad_norm": 20.75,
      "learning_rate": 1.8705868779324667e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4256620.0,
      "step": 9789
    },
    {
      "epoch": 0.7530189985385739,
      "grad_norm": 13.0625,
      "learning_rate": 1.870471502192139e-06,
      "loss": 0.0065,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4257424.0,
      "step": 9790
    },
    {
      "epoch": 0.7530959156987924,
      "grad_norm": 20.75,
      "learning_rate": 1.8703561264518114e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4257998.0,
      "step": 9791
    },
    {
      "epoch": 0.7531728328590108,
      "grad_norm": 25.625,
      "learning_rate": 1.8702407507114839e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 4258290.0,
      "step": 9792
    },
    {
      "epoch": 0.7532497500192293,
      "grad_norm": 16.75,
      "learning_rate": 1.8701253749711561e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4258614.0,
      "step": 9793
    },
    {
      "epoch": 0.7533266671794477,
      "grad_norm": 8.5625,
      "learning_rate": 1.8700099992308286e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 4259158.0,
      "step": 9794
    },
    {
      "epoch": 0.7534035843396661,
      "grad_norm": 21.25,
      "learning_rate": 1.8698946234905006e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4259456.0,
      "step": 9795
    },
    {
      "epoch": 0.7534805014998847,
      "grad_norm": 37.25,
      "learning_rate": 1.8697792477501731e-06,
      "loss": 0.0453,
      "mean_token_accuracy": 0.984375,
      "num_tokens": 4259780.0,
      "step": 9796
    },
    {
      "epoch": 0.7535574186601031,
      "grad_norm": 20.375,
      "learning_rate": 1.8696638720098454e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4260120.0,
      "step": 9797
    },
    {
      "epoch": 0.7536343358203215,
      "grad_norm": 31.5,
      "learning_rate": 1.8695484962695178e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4260476.0,
      "step": 9798
    },
    {
      "epoch": 0.75371125298054,
      "grad_norm": 16.0,
      "learning_rate": 1.8694331205291903e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 4260748.0,
      "step": 9799
    },
    {
      "epoch": 0.7537881701407584,
      "grad_norm": 7.03125,
      "learning_rate": 1.8693177447888624e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9961612224578857,
      "num_tokens": 4261270.0,
      "step": 9800
    },
    {
      "epoch": 0.7538650873009769,
      "grad_norm": 21.25,
      "learning_rate": 1.8692023690485346e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 4261540.0,
      "step": 9801
    },
    {
      "epoch": 0.7539420044611953,
      "grad_norm": 16.875,
      "learning_rate": 1.869086993308207e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 4261866.0,
      "step": 9802
    },
    {
      "epoch": 0.7540189216214137,
      "grad_norm": 28.0,
      "learning_rate": 1.8689716175678796e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 4262180.0,
      "step": 9803
    },
    {
      "epoch": 0.7540958387816322,
      "grad_norm": 4.375,
      "learning_rate": 1.8688562418275518e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4262750.0,
      "step": 9804
    },
    {
      "epoch": 0.7541727559418506,
      "grad_norm": 67.0,
      "learning_rate": 1.8687408660872239e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4263050.0,
      "step": 9805
    },
    {
      "epoch": 0.754249673102069,
      "grad_norm": 22.0,
      "learning_rate": 1.8686254903468963e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 4263836.0,
      "step": 9806
    },
    {
      "epoch": 0.7543265902622875,
      "grad_norm": 23.75,
      "learning_rate": 1.8685101146065688e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4264422.0,
      "step": 9807
    },
    {
      "epoch": 0.754403507422506,
      "grad_norm": 13.8125,
      "learning_rate": 1.868394738866241e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 4264692.0,
      "step": 9808
    },
    {
      "epoch": 0.7544804245827244,
      "grad_norm": 15.25,
      "learning_rate": 1.8682793631259135e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 4265218.0,
      "step": 9809
    },
    {
      "epoch": 0.7545573417429429,
      "grad_norm": 44.25,
      "learning_rate": 1.868163987385586e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4265548.0,
      "step": 9810
    },
    {
      "epoch": 0.7546342589031613,
      "grad_norm": 38.25,
      "learning_rate": 1.868048611645258e-06,
      "loss": 0.0351,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4265852.0,
      "step": 9811
    },
    {
      "epoch": 0.7547111760633798,
      "grad_norm": 48.0,
      "learning_rate": 1.8679332359049303e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4266160.0,
      "step": 9812
    },
    {
      "epoch": 0.7547880932235982,
      "grad_norm": 16.125,
      "learning_rate": 1.8678178601646028e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4266710.0,
      "step": 9813
    },
    {
      "epoch": 0.7548650103838166,
      "grad_norm": 17.75,
      "learning_rate": 1.8677024844242752e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 4267266.0,
      "step": 9814
    },
    {
      "epoch": 0.7549419275440351,
      "grad_norm": 7.9375,
      "learning_rate": 1.8675871086839475e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4267590.0,
      "step": 9815
    },
    {
      "epoch": 0.7550188447042535,
      "grad_norm": 7.84375,
      "learning_rate": 1.8674717329436195e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4267876.0,
      "step": 9816
    },
    {
      "epoch": 0.7550957618644719,
      "grad_norm": 17.0,
      "learning_rate": 1.867356357203292e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4268214.0,
      "step": 9817
    },
    {
      "epoch": 0.7551726790246904,
      "grad_norm": 33.25,
      "learning_rate": 1.8672409814629645e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4268470.0,
      "step": 9818
    },
    {
      "epoch": 0.7552495961849088,
      "grad_norm": 2.5,
      "learning_rate": 1.8671256057226367e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4269022.0,
      "step": 9819
    },
    {
      "epoch": 0.7553265133451273,
      "grad_norm": 9.375,
      "learning_rate": 1.8670102299823092e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 4269378.0,
      "step": 9820
    },
    {
      "epoch": 0.7554034305053458,
      "grad_norm": 15.5625,
      "learning_rate": 1.8668948542419817e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 4269950.0,
      "step": 9821
    },
    {
      "epoch": 0.7554803476655642,
      "grad_norm": 18.875,
      "learning_rate": 1.8667794785016537e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4270514.0,
      "step": 9822
    },
    {
      "epoch": 0.7555572648257827,
      "grad_norm": 27.375,
      "learning_rate": 1.866664102761326e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 4271072.0,
      "step": 9823
    },
    {
      "epoch": 0.7556341819860011,
      "grad_norm": 67.0,
      "learning_rate": 1.8665487270209985e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4271402.0,
      "step": 9824
    },
    {
      "epoch": 0.7557110991462195,
      "grad_norm": 49.75,
      "learning_rate": 1.866433351280671e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4271700.0,
      "step": 9825
    },
    {
      "epoch": 0.755788016306438,
      "grad_norm": 12.9375,
      "learning_rate": 1.8663179755403432e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4272274.0,
      "step": 9826
    },
    {
      "epoch": 0.7558649334666564,
      "grad_norm": 40.75,
      "learning_rate": 1.8662025998000152e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4272584.0,
      "step": 9827
    },
    {
      "epoch": 0.7559418506268748,
      "grad_norm": 9.875,
      "learning_rate": 1.8660872240596877e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4273130.0,
      "step": 9828
    },
    {
      "epoch": 0.7560187677870933,
      "grad_norm": 22.375,
      "learning_rate": 1.8659718483193602e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9888682961463928,
      "num_tokens": 4273670.0,
      "step": 9829
    },
    {
      "epoch": 0.7560956849473117,
      "grad_norm": 34.0,
      "learning_rate": 1.8658564725790324e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 4274246.0,
      "step": 9830
    },
    {
      "epoch": 0.7561726021075302,
      "grad_norm": 24.0,
      "learning_rate": 1.865741096838705e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.998171865940094,
      "num_tokens": 4274794.0,
      "step": 9831
    },
    {
      "epoch": 0.7562495192677486,
      "grad_norm": 46.5,
      "learning_rate": 1.8656257210983772e-06,
      "loss": 0.0478,
      "mean_token_accuracy": 0.9781249761581421,
      "num_tokens": 4275118.0,
      "step": 9832
    },
    {
      "epoch": 0.756326436427967,
      "grad_norm": 5.0,
      "learning_rate": 1.8655103453580494e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9982425570487976,
      "num_tokens": 4275688.0,
      "step": 9833
    },
    {
      "epoch": 0.7564033535881856,
      "grad_norm": 17.0,
      "learning_rate": 1.8653949696177217e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4275968.0,
      "step": 9834
    },
    {
      "epoch": 0.756480270748404,
      "grad_norm": 5.15625,
      "learning_rate": 1.8652795938773941e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4276516.0,
      "step": 9835
    },
    {
      "epoch": 0.7565571879086224,
      "grad_norm": 2.40625,
      "learning_rate": 1.8651642181370664e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4277078.0,
      "step": 9836
    },
    {
      "epoch": 0.7566341050688409,
      "grad_norm": 1.8359375,
      "learning_rate": 1.8650488423967389e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4277642.0,
      "step": 9837
    },
    {
      "epoch": 0.7567110222290593,
      "grad_norm": 18.0,
      "learning_rate": 1.864933466656411e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4277954.0,
      "step": 9838
    },
    {
      "epoch": 0.7567879393892778,
      "grad_norm": 37.25,
      "learning_rate": 1.8648180909160834e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4278298.0,
      "step": 9839
    },
    {
      "epoch": 0.7568648565494962,
      "grad_norm": 15.625,
      "learning_rate": 1.8647027151757556e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 4278818.0,
      "step": 9840
    },
    {
      "epoch": 0.7569417737097146,
      "grad_norm": 22.75,
      "learning_rate": 1.8645873394354281e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 4279106.0,
      "step": 9841
    },
    {
      "epoch": 0.7570186908699331,
      "grad_norm": 59.5,
      "learning_rate": 1.8644719636951006e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4279360.0,
      "step": 9842
    },
    {
      "epoch": 0.7570956080301515,
      "grad_norm": 16.375,
      "learning_rate": 1.8643565879547726e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9884169697761536,
      "num_tokens": 4279620.0,
      "step": 9843
    },
    {
      "epoch": 0.7571725251903699,
      "grad_norm": 7.46875,
      "learning_rate": 1.8642412122144449e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 4280188.0,
      "step": 9844
    },
    {
      "epoch": 0.7572494423505884,
      "grad_norm": 20.0,
      "learning_rate": 1.8641258364741174e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9844236969947815,
      "num_tokens": 4280510.0,
      "step": 9845
    },
    {
      "epoch": 0.7573263595108068,
      "grad_norm": 39.0,
      "learning_rate": 1.8640104607337898e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4280830.0,
      "step": 9846
    },
    {
      "epoch": 0.7574032766710254,
      "grad_norm": 15.5,
      "learning_rate": 1.863895084993462e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 4281394.0,
      "step": 9847
    },
    {
      "epoch": 0.7574801938312438,
      "grad_norm": 82.5,
      "learning_rate": 1.8637797092531346e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4281750.0,
      "step": 9848
    },
    {
      "epoch": 0.7575571109914622,
      "grad_norm": 13.8125,
      "learning_rate": 1.8636643335128066e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4282298.0,
      "step": 9849
    },
    {
      "epoch": 0.7576340281516807,
      "grad_norm": 24.25,
      "learning_rate": 1.863548957772479e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4282844.0,
      "step": 9850
    },
    {
      "epoch": 0.7577109453118991,
      "grad_norm": 31.375,
      "learning_rate": 1.8634335820321513e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4283400.0,
      "step": 9851
    },
    {
      "epoch": 0.7577878624721175,
      "grad_norm": 23.625,
      "learning_rate": 1.8633182062918238e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9925187230110168,
      "num_tokens": 4284204.0,
      "step": 9852
    },
    {
      "epoch": 0.757864779632336,
      "grad_norm": 8.6875,
      "learning_rate": 1.8632028305514963e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4284722.0,
      "step": 9853
    },
    {
      "epoch": 0.7579416967925544,
      "grad_norm": 50.25,
      "learning_rate": 1.8630874548111683e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4285046.0,
      "step": 9854
    },
    {
      "epoch": 0.7580186139527728,
      "grad_norm": 9.0,
      "learning_rate": 1.8629720790708406e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4285608.0,
      "step": 9855
    },
    {
      "epoch": 0.7580955311129913,
      "grad_norm": 11.625,
      "learning_rate": 1.862856703330513e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4285936.0,
      "step": 9856
    },
    {
      "epoch": 0.7581724482732097,
      "grad_norm": 2.015625,
      "learning_rate": 1.8627413275901855e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9974554777145386,
      "num_tokens": 4286724.0,
      "step": 9857
    },
    {
      "epoch": 0.7582493654334282,
      "grad_norm": 23.5,
      "learning_rate": 1.8626259518498578e-06,
      "loss": 0.0531,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 4287026.0,
      "step": 9858
    },
    {
      "epoch": 0.7583262825936467,
      "grad_norm": 63.5,
      "learning_rate": 1.8625105761095302e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 4287332.0,
      "step": 9859
    },
    {
      "epoch": 0.758403199753865,
      "grad_norm": 37.25,
      "learning_rate": 1.8623952003692023e-06,
      "loss": 0.0357,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 4287642.0,
      "step": 9860
    },
    {
      "epoch": 0.7584801169140836,
      "grad_norm": 13.75,
      "learning_rate": 1.8622798246288748e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 4288200.0,
      "step": 9861
    },
    {
      "epoch": 0.758557034074302,
      "grad_norm": 34.5,
      "learning_rate": 1.862164448888547e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4288546.0,
      "step": 9862
    },
    {
      "epoch": 0.7586339512345204,
      "grad_norm": 28.25,
      "learning_rate": 1.8620490731482195e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4288826.0,
      "step": 9863
    },
    {
      "epoch": 0.7587108683947389,
      "grad_norm": 18.25,
      "learning_rate": 1.861933697407892e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4289160.0,
      "step": 9864
    },
    {
      "epoch": 0.7587877855549573,
      "grad_norm": 42.25,
      "learning_rate": 1.861818321667564e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4289488.0,
      "step": 9865
    },
    {
      "epoch": 0.7588647027151758,
      "grad_norm": 26.5,
      "learning_rate": 1.8617029459272363e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 4290058.0,
      "step": 9866
    },
    {
      "epoch": 0.7589416198753942,
      "grad_norm": 57.5,
      "learning_rate": 1.8615875701869087e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 4290654.0,
      "step": 9867
    },
    {
      "epoch": 0.7590185370356126,
      "grad_norm": 32.5,
      "learning_rate": 1.8614721944465812e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 4291190.0,
      "step": 9868
    },
    {
      "epoch": 0.7590954541958311,
      "grad_norm": 25.625,
      "learning_rate": 1.8613568187062535e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4291762.0,
      "step": 9869
    },
    {
      "epoch": 0.7591723713560495,
      "grad_norm": 7.78125,
      "learning_rate": 1.861241442965926e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4292046.0,
      "step": 9870
    },
    {
      "epoch": 0.7592492885162679,
      "grad_norm": 38.0,
      "learning_rate": 1.861126067225598e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 4292842.0,
      "step": 9871
    },
    {
      "epoch": 0.7593262056764865,
      "grad_norm": 12.1875,
      "learning_rate": 1.8610106914852704e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9961758852005005,
      "num_tokens": 4293366.0,
      "step": 9872
    },
    {
      "epoch": 0.7594031228367049,
      "grad_norm": 9.4375,
      "learning_rate": 1.8608953157449427e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 4293628.0,
      "step": 9873
    },
    {
      "epoch": 0.7594800399969233,
      "grad_norm": 13.625,
      "learning_rate": 1.8607799400046152e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 4293918.0,
      "step": 9874
    },
    {
      "epoch": 0.7595569571571418,
      "grad_norm": 17.0,
      "learning_rate": 1.8606645642642876e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9921568632125854,
      "num_tokens": 4294174.0,
      "step": 9875
    },
    {
      "epoch": 0.7596338743173602,
      "grad_norm": 20.5,
      "learning_rate": 1.8605491885239597e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4294508.0,
      "step": 9876
    },
    {
      "epoch": 0.7597107914775787,
      "grad_norm": 14.3125,
      "learning_rate": 1.860433812783632e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4294788.0,
      "step": 9877
    },
    {
      "epoch": 0.7597877086377971,
      "grad_norm": 10.375,
      "learning_rate": 1.8603184370433044e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4295338.0,
      "step": 9878
    },
    {
      "epoch": 0.7598646257980155,
      "grad_norm": 16.125,
      "learning_rate": 1.8602030613029769e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4295648.0,
      "step": 9879
    },
    {
      "epoch": 0.759941542958234,
      "grad_norm": 11.0625,
      "learning_rate": 1.8600876855626491e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4296198.0,
      "step": 9880
    },
    {
      "epoch": 0.7600184601184524,
      "grad_norm": 17.75,
      "learning_rate": 1.8599723098223212e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 4296730.0,
      "step": 9881
    },
    {
      "epoch": 0.7600953772786708,
      "grad_norm": 18.25,
      "learning_rate": 1.8598569340819937e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 4297556.0,
      "step": 9882
    },
    {
      "epoch": 0.7601722944388893,
      "grad_norm": 15.375,
      "learning_rate": 1.8597415583416661e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4297890.0,
      "step": 9883
    },
    {
      "epoch": 0.7602492115991077,
      "grad_norm": 3.640625,
      "learning_rate": 1.8596261826013384e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 4298674.0,
      "step": 9884
    },
    {
      "epoch": 0.7603261287593263,
      "grad_norm": 30.75,
      "learning_rate": 1.8595108068610109e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4298980.0,
      "step": 9885
    },
    {
      "epoch": 0.7604030459195447,
      "grad_norm": 14.0,
      "learning_rate": 1.8593954311206831e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4299264.0,
      "step": 9886
    },
    {
      "epoch": 0.7604799630797631,
      "grad_norm": 14.3125,
      "learning_rate": 1.8592800553803554e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9964413046836853,
      "num_tokens": 4299546.0,
      "step": 9887
    },
    {
      "epoch": 0.7605568802399816,
      "grad_norm": 6.25,
      "learning_rate": 1.8591646796400276e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 4300350.0,
      "step": 9888
    },
    {
      "epoch": 0.7606337974002,
      "grad_norm": 8.0,
      "learning_rate": 1.8590493038997001e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9885057210922241,
      "num_tokens": 4300612.0,
      "step": 9889
    },
    {
      "epoch": 0.7607107145604184,
      "grad_norm": 25.375,
      "learning_rate": 1.8589339281593724e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4301198.0,
      "step": 9890
    },
    {
      "epoch": 0.7607876317206369,
      "grad_norm": 13.3125,
      "learning_rate": 1.8588185524190448e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 4302012.0,
      "step": 9891
    },
    {
      "epoch": 0.7608645488808553,
      "grad_norm": 78.5,
      "learning_rate": 1.8587031766787169e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4302350.0,
      "step": 9892
    },
    {
      "epoch": 0.7609414660410737,
      "grad_norm": 10.75,
      "learning_rate": 1.8585878009383894e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4302674.0,
      "step": 9893
    },
    {
      "epoch": 0.7610183832012922,
      "grad_norm": 21.875,
      "learning_rate": 1.8584724251980616e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4302980.0,
      "step": 9894
    },
    {
      "epoch": 0.7610953003615106,
      "grad_norm": 10.0625,
      "learning_rate": 1.858357049457734e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4303212.0,
      "step": 9895
    },
    {
      "epoch": 0.7611722175217291,
      "grad_norm": 4.90625,
      "learning_rate": 1.8582416737174065e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9961464405059814,
      "num_tokens": 4303732.0,
      "step": 9896
    },
    {
      "epoch": 0.7612491346819475,
      "grad_norm": 5.9375,
      "learning_rate": 1.8581262979770788e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.998171865940094,
      "num_tokens": 4304280.0,
      "step": 9897
    },
    {
      "epoch": 0.761326051842166,
      "grad_norm": 20.25,
      "learning_rate": 1.8580109222367509e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 4304550.0,
      "step": 9898
    },
    {
      "epoch": 0.7614029690023845,
      "grad_norm": 4.84375,
      "learning_rate": 1.8578955464964233e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4305134.0,
      "step": 9899
    },
    {
      "epoch": 0.7614798861626029,
      "grad_norm": 13.0625,
      "learning_rate": 1.8577801707560958e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4305456.0,
      "step": 9900
    },
    {
      "epoch": 0.7615568033228213,
      "grad_norm": 13.25,
      "learning_rate": 1.857664795015768e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 4306242.0,
      "step": 9901
    },
    {
      "epoch": 0.7616337204830398,
      "grad_norm": 41.0,
      "learning_rate": 1.8575494192754405e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4306520.0,
      "step": 9902
    },
    {
      "epoch": 0.7617106376432582,
      "grad_norm": 3.140625,
      "learning_rate": 1.8574340435351126e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4307038.0,
      "step": 9903
    },
    {
      "epoch": 0.7617875548034767,
      "grad_norm": 15.625,
      "learning_rate": 1.857318667794785e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 4307330.0,
      "step": 9904
    },
    {
      "epoch": 0.7618644719636951,
      "grad_norm": 7.375,
      "learning_rate": 1.8572032920544573e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4307876.0,
      "step": 9905
    },
    {
      "epoch": 0.7619413891239135,
      "grad_norm": 3.875,
      "learning_rate": 1.8570879163141298e-06,
      "loss": 0.01,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4308370.0,
      "step": 9906
    },
    {
      "epoch": 0.762018306284132,
      "grad_norm": 13.125,
      "learning_rate": 1.8569725405738022e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 4308704.0,
      "step": 9907
    },
    {
      "epoch": 0.7620952234443504,
      "grad_norm": 6.59375,
      "learning_rate": 1.8568571648334743e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4309274.0,
      "step": 9908
    },
    {
      "epoch": 0.7621721406045688,
      "grad_norm": 15.9375,
      "learning_rate": 1.8567417890931465e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 4309838.0,
      "step": 9909
    },
    {
      "epoch": 0.7622490577647874,
      "grad_norm": 5.5625,
      "learning_rate": 1.856626413352819e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4310390.0,
      "step": 9910
    },
    {
      "epoch": 0.7623259749250058,
      "grad_norm": 18.875,
      "learning_rate": 1.8565110376124915e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 4310686.0,
      "step": 9911
    },
    {
      "epoch": 0.7624028920852242,
      "grad_norm": 9.75,
      "learning_rate": 1.8563956618721637e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4311228.0,
      "step": 9912
    },
    {
      "epoch": 0.7624798092454427,
      "grad_norm": 10.0625,
      "learning_rate": 1.8562802861318362e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 4311800.0,
      "step": 9913
    },
    {
      "epoch": 0.7625567264056611,
      "grad_norm": 28.625,
      "learning_rate": 1.8561649103915083e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4312104.0,
      "step": 9914
    },
    {
      "epoch": 0.7626336435658796,
      "grad_norm": 20.5,
      "learning_rate": 1.8560495346511807e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4312412.0,
      "step": 9915
    },
    {
      "epoch": 0.762710560726098,
      "grad_norm": 14.5625,
      "learning_rate": 1.855934158910853e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 4312752.0,
      "step": 9916
    },
    {
      "epoch": 0.7627874778863164,
      "grad_norm": 25.75,
      "learning_rate": 1.8558187831705255e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 4313044.0,
      "step": 9917
    },
    {
      "epoch": 0.7628643950465349,
      "grad_norm": 9.6875,
      "learning_rate": 1.855703407430198e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4313618.0,
      "step": 9918
    },
    {
      "epoch": 0.7629413122067533,
      "grad_norm": 6.65625,
      "learning_rate": 1.85558803168987e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4314164.0,
      "step": 9919
    },
    {
      "epoch": 0.7630182293669717,
      "grad_norm": 34.0,
      "learning_rate": 1.8554726559495422e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 4314464.0,
      "step": 9920
    },
    {
      "epoch": 0.7630951465271902,
      "grad_norm": 11.6875,
      "learning_rate": 1.8553572802092147e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9980732202529907,
      "num_tokens": 4314984.0,
      "step": 9921
    },
    {
      "epoch": 0.7631720636874086,
      "grad_norm": 55.0,
      "learning_rate": 1.8552419044688872e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4315306.0,
      "step": 9922
    },
    {
      "epoch": 0.7632489808476272,
      "grad_norm": 71.5,
      "learning_rate": 1.8551265287285594e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4315662.0,
      "step": 9923
    },
    {
      "epoch": 0.7633258980078456,
      "grad_norm": 12.6875,
      "learning_rate": 1.8550111529882319e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4316002.0,
      "step": 9924
    },
    {
      "epoch": 0.763402815168064,
      "grad_norm": 5.0,
      "learning_rate": 1.854895777247904e-06,
      "loss": 0.0068,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 4316788.0,
      "step": 9925
    },
    {
      "epoch": 0.7634797323282825,
      "grad_norm": 24.125,
      "learning_rate": 1.8547804015075764e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4317042.0,
      "step": 9926
    },
    {
      "epoch": 0.7635566494885009,
      "grad_norm": 28.375,
      "learning_rate": 1.8546650257672487e-06,
      "loss": 0.0444,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4317344.0,
      "step": 9927
    },
    {
      "epoch": 0.7636335666487193,
      "grad_norm": 11.25,
      "learning_rate": 1.8545496500269211e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4317914.0,
      "step": 9928
    },
    {
      "epoch": 0.7637104838089378,
      "grad_norm": 19.75,
      "learning_rate": 1.8544342742865936e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.991482138633728,
      "num_tokens": 4318502.0,
      "step": 9929
    },
    {
      "epoch": 0.7637874009691562,
      "grad_norm": 43.0,
      "learning_rate": 1.8543188985462657e-06,
      "loss": 0.0467,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4318810.0,
      "step": 9930
    },
    {
      "epoch": 0.7638643181293746,
      "grad_norm": 5.8125,
      "learning_rate": 1.854203522805938e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 4319634.0,
      "step": 9931
    },
    {
      "epoch": 0.7639412352895931,
      "grad_norm": 4.625,
      "learning_rate": 1.8540881470656104e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4320176.0,
      "step": 9932
    },
    {
      "epoch": 0.7640181524498115,
      "grad_norm": 35.75,
      "learning_rate": 1.8539727713252829e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4320482.0,
      "step": 9933
    },
    {
      "epoch": 0.76409506961003,
      "grad_norm": 32.0,
      "learning_rate": 1.8538573955849551e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9862068891525269,
      "num_tokens": 4320774.0,
      "step": 9934
    },
    {
      "epoch": 0.7641719867702484,
      "grad_norm": 31.375,
      "learning_rate": 1.8537420198446276e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 4321106.0,
      "step": 9935
    },
    {
      "epoch": 0.7642489039304668,
      "grad_norm": 31.25,
      "learning_rate": 1.8536266441042996e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4321434.0,
      "step": 9936
    },
    {
      "epoch": 0.7643258210906854,
      "grad_norm": 10.0,
      "learning_rate": 1.853511268363972e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 4321768.0,
      "step": 9937
    },
    {
      "epoch": 0.7644027382509038,
      "grad_norm": 35.25,
      "learning_rate": 1.8533958926236444e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4322108.0,
      "step": 9938
    },
    {
      "epoch": 0.7644796554111222,
      "grad_norm": 45.25,
      "learning_rate": 1.8532805168833168e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4322386.0,
      "step": 9939
    },
    {
      "epoch": 0.7645565725713407,
      "grad_norm": 25.625,
      "learning_rate": 1.853165141142989e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4322700.0,
      "step": 9940
    },
    {
      "epoch": 0.7646334897315591,
      "grad_norm": 14.375,
      "learning_rate": 1.8530497654026613e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4323264.0,
      "step": 9941
    },
    {
      "epoch": 0.7647104068917776,
      "grad_norm": 19.75,
      "learning_rate": 1.8529343896623336e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 4323568.0,
      "step": 9942
    },
    {
      "epoch": 0.764787324051996,
      "grad_norm": 24.125,
      "learning_rate": 1.852819013922006e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 4324388.0,
      "step": 9943
    },
    {
      "epoch": 0.7648642412122144,
      "grad_norm": 26.375,
      "learning_rate": 1.8527036381816783e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4324710.0,
      "step": 9944
    },
    {
      "epoch": 0.7649411583724329,
      "grad_norm": 21.375,
      "learning_rate": 1.8525882624413508e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9851190447807312,
      "num_tokens": 4325048.0,
      "step": 9945
    },
    {
      "epoch": 0.7650180755326513,
      "grad_norm": 8.8125,
      "learning_rate": 1.8524728867010228e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4325392.0,
      "step": 9946
    },
    {
      "epoch": 0.7650949926928697,
      "grad_norm": 42.75,
      "learning_rate": 1.8523575109606953e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4325704.0,
      "step": 9947
    },
    {
      "epoch": 0.7651719098530883,
      "grad_norm": 9.8125,
      "learning_rate": 1.8522421352203676e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4326246.0,
      "step": 9948
    },
    {
      "epoch": 0.7652488270133067,
      "grad_norm": 20.5,
      "learning_rate": 1.85212675948004e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 4326578.0,
      "step": 9949
    },
    {
      "epoch": 0.7653257441735252,
      "grad_norm": 8.5,
      "learning_rate": 1.8520113837397125e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4327140.0,
      "step": 9950
    },
    {
      "epoch": 0.7654026613337436,
      "grad_norm": 11.0625,
      "learning_rate": 1.8518960079993848e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4327696.0,
      "step": 9951
    },
    {
      "epoch": 0.765479578493962,
      "grad_norm": 42.5,
      "learning_rate": 1.8517806322590568e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4328048.0,
      "step": 9952
    },
    {
      "epoch": 0.7655564956541805,
      "grad_norm": 12.0625,
      "learning_rate": 1.8516652565187293e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 4328572.0,
      "step": 9953
    },
    {
      "epoch": 0.7656334128143989,
      "grad_norm": 5.875,
      "learning_rate": 1.8515498807784018e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 4329126.0,
      "step": 9954
    },
    {
      "epoch": 0.7657103299746173,
      "grad_norm": 51.25,
      "learning_rate": 1.851434505038074e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9968253970146179,
      "num_tokens": 4329442.0,
      "step": 9955
    },
    {
      "epoch": 0.7657872471348358,
      "grad_norm": 19.25,
      "learning_rate": 1.8513191292977465e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4329752.0,
      "step": 9956
    },
    {
      "epoch": 0.7658641642950542,
      "grad_norm": 34.5,
      "learning_rate": 1.8512037535574185e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 4330320.0,
      "step": 9957
    },
    {
      "epoch": 0.7659410814552726,
      "grad_norm": 5.6875,
      "learning_rate": 1.851088377817091e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4330894.0,
      "step": 9958
    },
    {
      "epoch": 0.7660179986154911,
      "grad_norm": 35.0,
      "learning_rate": 1.8509730020767633e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 4331196.0,
      "step": 9959
    },
    {
      "epoch": 0.7660949157757095,
      "grad_norm": 9.3125,
      "learning_rate": 1.8508576263364357e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4331742.0,
      "step": 9960
    },
    {
      "epoch": 0.766171832935928,
      "grad_norm": 4.375,
      "learning_rate": 1.8507422505961082e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4332316.0,
      "step": 9961
    },
    {
      "epoch": 0.7662487500961465,
      "grad_norm": 4.84375,
      "learning_rate": 1.8506268748557805e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 4332810.0,
      "step": 9962
    },
    {
      "epoch": 0.7663256672563649,
      "grad_norm": 54.5,
      "learning_rate": 1.8505114991154525e-06,
      "loss": 0.0369,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 4333086.0,
      "step": 9963
    },
    {
      "epoch": 0.7664025844165834,
      "grad_norm": 25.875,
      "learning_rate": 1.850396123375125e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 4333602.0,
      "step": 9964
    },
    {
      "epoch": 0.7664795015768018,
      "grad_norm": 7.65625,
      "learning_rate": 1.8502807476347974e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 4333936.0,
      "step": 9965
    },
    {
      "epoch": 0.7665564187370202,
      "grad_norm": 16.5,
      "learning_rate": 1.8501653718944697e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4334192.0,
      "step": 9966
    },
    {
      "epoch": 0.7666333358972387,
      "grad_norm": 4.46875,
      "learning_rate": 1.8500499961541422e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4334734.0,
      "step": 9967
    },
    {
      "epoch": 0.7667102530574571,
      "grad_norm": 10.8125,
      "learning_rate": 1.8499346204138142e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 4335520.0,
      "step": 9968
    },
    {
      "epoch": 0.7667871702176756,
      "grad_norm": 64.5,
      "learning_rate": 1.8498192446734867e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4336094.0,
      "step": 9969
    },
    {
      "epoch": 0.766864087377894,
      "grad_norm": 17.875,
      "learning_rate": 1.849703868933159e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4336388.0,
      "step": 9970
    },
    {
      "epoch": 0.7669410045381124,
      "grad_norm": 20.25,
      "learning_rate": 1.8495884931928314e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4336672.0,
      "step": 9971
    },
    {
      "epoch": 0.7670179216983309,
      "grad_norm": 40.0,
      "learning_rate": 1.8494731174525039e-06,
      "loss": 0.0469,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 4336998.0,
      "step": 9972
    },
    {
      "epoch": 0.7670948388585493,
      "grad_norm": 36.5,
      "learning_rate": 1.8493577417121761e-06,
      "loss": 0.0419,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4337298.0,
      "step": 9973
    },
    {
      "epoch": 0.7671717560187677,
      "grad_norm": 29.25,
      "learning_rate": 1.8492423659718482e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4337844.0,
      "step": 9974
    },
    {
      "epoch": 0.7672486731789863,
      "grad_norm": 9.0,
      "learning_rate": 1.8491269902315207e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4338152.0,
      "step": 9975
    },
    {
      "epoch": 0.7673255903392047,
      "grad_norm": 20.875,
      "learning_rate": 1.8490116144911931e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4338700.0,
      "step": 9976
    },
    {
      "epoch": 0.7674025074994231,
      "grad_norm": 9.5,
      "learning_rate": 1.8488962387508654e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 4339486.0,
      "step": 9977
    },
    {
      "epoch": 0.7674794246596416,
      "grad_norm": 5.34375,
      "learning_rate": 1.8487808630105379e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4340028.0,
      "step": 9978
    },
    {
      "epoch": 0.76755634181986,
      "grad_norm": 4.34375,
      "learning_rate": 1.84866548727021e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4340574.0,
      "step": 9979
    },
    {
      "epoch": 0.7676332589800785,
      "grad_norm": 15.5,
      "learning_rate": 1.8485501115298824e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4340920.0,
      "step": 9980
    },
    {
      "epoch": 0.7677101761402969,
      "grad_norm": 32.25,
      "learning_rate": 1.8484347357895546e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 4341264.0,
      "step": 9981
    },
    {
      "epoch": 0.7677870933005153,
      "grad_norm": 9.0625,
      "learning_rate": 1.848319360049227e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4341822.0,
      "step": 9982
    },
    {
      "epoch": 0.7678640104607338,
      "grad_norm": 8.8125,
      "learning_rate": 1.8482039843088996e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 4342184.0,
      "step": 9983
    },
    {
      "epoch": 0.7679409276209522,
      "grad_norm": 9.1875,
      "learning_rate": 1.8480886085685716e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 4342970.0,
      "step": 9984
    },
    {
      "epoch": 0.7680178447811706,
      "grad_norm": 10.3125,
      "learning_rate": 1.8479732328282439e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4343544.0,
      "step": 9985
    },
    {
      "epoch": 0.7680947619413891,
      "grad_norm": 30.375,
      "learning_rate": 1.8478578570879163e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4344066.0,
      "step": 9986
    },
    {
      "epoch": 0.7681716791016076,
      "grad_norm": 26.25,
      "learning_rate": 1.8477424813475888e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4344588.0,
      "step": 9987
    },
    {
      "epoch": 0.7682485962618261,
      "grad_norm": 11.75,
      "learning_rate": 1.847627105607261e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4345112.0,
      "step": 9988
    },
    {
      "epoch": 0.7683255134220445,
      "grad_norm": 17.125,
      "learning_rate": 1.8475117298669335e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4345634.0,
      "step": 9989
    },
    {
      "epoch": 0.7684024305822629,
      "grad_norm": 39.0,
      "learning_rate": 1.8473963541266056e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4345940.0,
      "step": 9990
    },
    {
      "epoch": 0.7684793477424814,
      "grad_norm": 16.0,
      "learning_rate": 1.847280978386278e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 4346252.0,
      "step": 9991
    },
    {
      "epoch": 0.7685562649026998,
      "grad_norm": 10.0625,
      "learning_rate": 1.8471656026459503e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4346598.0,
      "step": 9992
    },
    {
      "epoch": 0.7686331820629182,
      "grad_norm": 81.0,
      "learning_rate": 1.8470502269056228e-06,
      "loss": 0.0338,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4346882.0,
      "step": 9993
    },
    {
      "epoch": 0.7687100992231367,
      "grad_norm": 37.75,
      "learning_rate": 1.846934851165295e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4347222.0,
      "step": 9994
    },
    {
      "epoch": 0.7687870163833551,
      "grad_norm": 34.25,
      "learning_rate": 1.8468194754249673e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9975369572639465,
      "num_tokens": 4348036.0,
      "step": 9995
    },
    {
      "epoch": 0.7688639335435735,
      "grad_norm": 46.0,
      "learning_rate": 1.8467040996846396e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9967532753944397,
      "num_tokens": 4348346.0,
      "step": 9996
    },
    {
      "epoch": 0.768940850703792,
      "grad_norm": 23.25,
      "learning_rate": 1.846588723944312e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 4348926.0,
      "step": 9997
    },
    {
      "epoch": 0.7690177678640104,
      "grad_norm": 74.0,
      "learning_rate": 1.8464733482039843e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 4349276.0,
      "step": 9998
    },
    {
      "epoch": 0.769094685024229,
      "grad_norm": 34.0,
      "learning_rate": 1.8463579724636568e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 4350062.0,
      "step": 9999
    },
    {
      "epoch": 0.7691716021844474,
      "grad_norm": 38.75,
      "learning_rate": 1.8462425967233292e-06,
      "loss": 0.0215,
      "step": 10000
    },
    {
      "epoch": 0.7691716021844474,
      "eval_loss": 0.02166532725095749,
      "eval_mean_token_accuracy": 0.9926573014259339,
      "eval_num_tokens": 4350406.0,
      "eval_runtime": 21.0089,
      "eval_samples_per_second": 9.52,
      "eval_steps_per_second": 1.19,
      "step": 10000
    },
    {
      "epoch": 0.7692485193446658,
      "grad_norm": 41.5,
      "learning_rate": 1.8461272209830013e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9920102059841156,
      "num_tokens": 4350960.0,
      "step": 10001
    },
    {
      "epoch": 0.7693254365048843,
      "grad_norm": 51.75,
      "learning_rate": 1.8460118452426735e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 4351256.0,
      "step": 10002
    },
    {
      "epoch": 0.7694023536651027,
      "grad_norm": 45.75,
      "learning_rate": 1.845896469502346e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 4351838.0,
      "step": 10003
    },
    {
      "epoch": 0.7694792708253211,
      "grad_norm": 6.75,
      "learning_rate": 1.8457810937620185e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4352386.0,
      "step": 10004
    },
    {
      "epoch": 0.7695561879855396,
      "grad_norm": 22.25,
      "learning_rate": 1.8456657180216907e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 4352910.0,
      "step": 10005
    },
    {
      "epoch": 0.769633105145758,
      "grad_norm": 14.6875,
      "learning_rate": 1.8455503422813628e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4353232.0,
      "step": 10006
    },
    {
      "epoch": 0.7697100223059765,
      "grad_norm": 34.0,
      "learning_rate": 1.8454349665410352e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4353532.0,
      "step": 10007
    },
    {
      "epoch": 0.7697869394661949,
      "grad_norm": 74.5,
      "learning_rate": 1.8453195908007077e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4353816.0,
      "step": 10008
    },
    {
      "epoch": 0.7698638566264133,
      "grad_norm": 60.75,
      "learning_rate": 1.84520421506038e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4354110.0,
      "step": 10009
    },
    {
      "epoch": 0.7699407737866318,
      "grad_norm": 12.3125,
      "learning_rate": 1.8450888393200524e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4354432.0,
      "step": 10010
    },
    {
      "epoch": 0.7700176909468502,
      "grad_norm": 7.09375,
      "learning_rate": 1.8449734635797245e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4354994.0,
      "step": 10011
    },
    {
      "epoch": 0.7700946081070686,
      "grad_norm": 6.03125,
      "learning_rate": 1.844858087839397e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4355512.0,
      "step": 10012
    },
    {
      "epoch": 0.7701715252672872,
      "grad_norm": 25.0,
      "learning_rate": 1.8447427120990692e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 4355850.0,
      "step": 10013
    },
    {
      "epoch": 0.7702484424275056,
      "grad_norm": 7.40625,
      "learning_rate": 1.8446273363587417e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 4356440.0,
      "step": 10014
    },
    {
      "epoch": 0.770325359587724,
      "grad_norm": 9.125,
      "learning_rate": 1.8445119606184142e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 4356752.0,
      "step": 10015
    },
    {
      "epoch": 0.7704022767479425,
      "grad_norm": 10.8125,
      "learning_rate": 1.8443965848780864e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9969696998596191,
      "num_tokens": 4357084.0,
      "step": 10016
    },
    {
      "epoch": 0.7704791939081609,
      "grad_norm": 32.0,
      "learning_rate": 1.8442812091377585e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9968253970146179,
      "num_tokens": 4357400.0,
      "step": 10017
    },
    {
      "epoch": 0.7705561110683794,
      "grad_norm": 22.75,
      "learning_rate": 1.844165833397431e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4357744.0,
      "step": 10018
    },
    {
      "epoch": 0.7706330282285978,
      "grad_norm": 9.6875,
      "learning_rate": 1.8440504576571034e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4358030.0,
      "step": 10019
    },
    {
      "epoch": 0.7707099453888162,
      "grad_norm": 96.0,
      "learning_rate": 1.8439350819167757e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4358340.0,
      "step": 10020
    },
    {
      "epoch": 0.7707868625490347,
      "grad_norm": 30.625,
      "learning_rate": 1.8438197061764481e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4358624.0,
      "step": 10021
    },
    {
      "epoch": 0.7708637797092531,
      "grad_norm": 27.625,
      "learning_rate": 1.8437043304361202e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4359208.0,
      "step": 10022
    },
    {
      "epoch": 0.7709406968694715,
      "grad_norm": 52.25,
      "learning_rate": 1.8435889546957926e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 4359558.0,
      "step": 10023
    },
    {
      "epoch": 0.77101761402969,
      "grad_norm": 24.125,
      "learning_rate": 1.843473578955465e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4359842.0,
      "step": 10024
    },
    {
      "epoch": 0.7710945311899084,
      "grad_norm": 24.5,
      "learning_rate": 1.8433582032151374e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 4360398.0,
      "step": 10025
    },
    {
      "epoch": 0.771171448350127,
      "grad_norm": 17.375,
      "learning_rate": 1.8432428274748098e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4360958.0,
      "step": 10026
    },
    {
      "epoch": 0.7712483655103454,
      "grad_norm": 21.875,
      "learning_rate": 1.843127451734482e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4361506.0,
      "step": 10027
    },
    {
      "epoch": 0.7713252826705638,
      "grad_norm": 12.375,
      "learning_rate": 1.8430120759941541e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4362064.0,
      "step": 10028
    },
    {
      "epoch": 0.7714021998307823,
      "grad_norm": 91.0,
      "learning_rate": 1.8428967002538266e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9892086386680603,
      "num_tokens": 4362344.0,
      "step": 10029
    },
    {
      "epoch": 0.7714791169910007,
      "grad_norm": 14.9375,
      "learning_rate": 1.842781324513499e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9939758777618408,
      "num_tokens": 4362678.0,
      "step": 10030
    },
    {
      "epoch": 0.7715560341512191,
      "grad_norm": 2.703125,
      "learning_rate": 1.8426659487731713e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 4363474.0,
      "step": 10031
    },
    {
      "epoch": 0.7716329513114376,
      "grad_norm": 13.625,
      "learning_rate": 1.8425505730328438e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 4364038.0,
      "step": 10032
    },
    {
      "epoch": 0.771709868471656,
      "grad_norm": 9.625,
      "learning_rate": 1.8424351972925159e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 4364536.0,
      "step": 10033
    },
    {
      "epoch": 0.7717867856318744,
      "grad_norm": 14.8125,
      "learning_rate": 1.8423198215521883e-06,
      "loss": 0.0064,
      "mean_token_accuracy": 0.9976019263267517,
      "num_tokens": 4365372.0,
      "step": 10034
    },
    {
      "epoch": 0.7718637027920929,
      "grad_norm": 13.5,
      "learning_rate": 1.8422044458118606e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 4365912.0,
      "step": 10035
    },
    {
      "epoch": 0.7719406199523113,
      "grad_norm": 22.5,
      "learning_rate": 1.842089070071533e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4366256.0,
      "step": 10036
    },
    {
      "epoch": 0.7720175371125299,
      "grad_norm": 15.375,
      "learning_rate": 1.8419736943312055e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4366566.0,
      "step": 10037
    },
    {
      "epoch": 0.7720944542727483,
      "grad_norm": 9.625,
      "learning_rate": 1.8418583185908778e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4367088.0,
      "step": 10038
    },
    {
      "epoch": 0.7721713714329667,
      "grad_norm": 16.5,
      "learning_rate": 1.8417429428505498e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4367428.0,
      "step": 10039
    },
    {
      "epoch": 0.7722482885931852,
      "grad_norm": 10.0,
      "learning_rate": 1.8416275671102223e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9968847632408142,
      "num_tokens": 4367750.0,
      "step": 10040
    },
    {
      "epoch": 0.7723252057534036,
      "grad_norm": 33.0,
      "learning_rate": 1.8415121913698948e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4368082.0,
      "step": 10041
    },
    {
      "epoch": 0.772402122913622,
      "grad_norm": 7.59375,
      "learning_rate": 1.841396815629567e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9971346855163574,
      "num_tokens": 4368432.0,
      "step": 10042
    },
    {
      "epoch": 0.7724790400738405,
      "grad_norm": 6.5,
      "learning_rate": 1.8412814398892395e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9972066879272461,
      "num_tokens": 4368792.0,
      "step": 10043
    },
    {
      "epoch": 0.7725559572340589,
      "grad_norm": 26.875,
      "learning_rate": 1.8411660641489115e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4369102.0,
      "step": 10044
    },
    {
      "epoch": 0.7726328743942774,
      "grad_norm": 14.875,
      "learning_rate": 1.841050688408584e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.996221661567688,
      "num_tokens": 4369898.0,
      "step": 10045
    },
    {
      "epoch": 0.7727097915544958,
      "grad_norm": 43.25,
      "learning_rate": 1.8409353126682563e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4370178.0,
      "step": 10046
    },
    {
      "epoch": 0.7727867087147142,
      "grad_norm": 20.75,
      "learning_rate": 1.8408199369279287e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4370434.0,
      "step": 10047
    },
    {
      "epoch": 0.7728636258749327,
      "grad_norm": 5.8125,
      "learning_rate": 1.8407045611876012e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 4370978.0,
      "step": 10048
    },
    {
      "epoch": 0.7729405430351511,
      "grad_norm": 43.25,
      "learning_rate": 1.8405891854472733e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4371278.0,
      "step": 10049
    },
    {
      "epoch": 0.7730174601953695,
      "grad_norm": 46.0,
      "learning_rate": 1.8404738097069455e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 4372066.0,
      "step": 10050
    },
    {
      "epoch": 0.7730943773555881,
      "grad_norm": 8.625,
      "learning_rate": 1.840358433966618e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4372372.0,
      "step": 10051
    },
    {
      "epoch": 0.7731712945158065,
      "grad_norm": 18.625,
      "learning_rate": 1.8402430582262905e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4372678.0,
      "step": 10052
    },
    {
      "epoch": 0.7732482116760249,
      "grad_norm": 29.0,
      "learning_rate": 1.8401276824859627e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 4372998.0,
      "step": 10053
    },
    {
      "epoch": 0.7733251288362434,
      "grad_norm": 5.46875,
      "learning_rate": 1.8400123067456352e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4373544.0,
      "step": 10054
    },
    {
      "epoch": 0.7734020459964618,
      "grad_norm": 39.75,
      "learning_rate": 1.8398969310053072e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 4373874.0,
      "step": 10055
    },
    {
      "epoch": 0.7734789631566803,
      "grad_norm": 99.0,
      "learning_rate": 1.8397815552649797e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 4374176.0,
      "step": 10056
    },
    {
      "epoch": 0.7735558803168987,
      "grad_norm": 28.625,
      "learning_rate": 1.839666179524652e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 4374716.0,
      "step": 10057
    },
    {
      "epoch": 0.7736327974771171,
      "grad_norm": 17.375,
      "learning_rate": 1.8395508037843244e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 4375038.0,
      "step": 10058
    },
    {
      "epoch": 0.7737097146373356,
      "grad_norm": 55.25,
      "learning_rate": 1.8394354280439967e-06,
      "loss": 0.0386,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4375316.0,
      "step": 10059
    },
    {
      "epoch": 0.773786631797554,
      "grad_norm": 26.5,
      "learning_rate": 1.839320052303669e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 4375656.0,
      "step": 10060
    },
    {
      "epoch": 0.7738635489577724,
      "grad_norm": 12.5625,
      "learning_rate": 1.8392046765633412e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4376204.0,
      "step": 10061
    },
    {
      "epoch": 0.7739404661179909,
      "grad_norm": 24.125,
      "learning_rate": 1.8390893008230137e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 4376552.0,
      "step": 10062
    },
    {
      "epoch": 0.7740173832782093,
      "grad_norm": 9.125,
      "learning_rate": 1.838973925082686e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4376860.0,
      "step": 10063
    },
    {
      "epoch": 0.7740943004384279,
      "grad_norm": 56.25,
      "learning_rate": 1.8388585493423584e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 4377152.0,
      "step": 10064
    },
    {
      "epoch": 0.7741712175986463,
      "grad_norm": 33.25,
      "learning_rate": 1.8387431736020309e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4377498.0,
      "step": 10065
    },
    {
      "epoch": 0.7742481347588647,
      "grad_norm": 5.34375,
      "learning_rate": 1.838627797861703e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4378084.0,
      "step": 10066
    },
    {
      "epoch": 0.7743250519190832,
      "grad_norm": 14.375,
      "learning_rate": 1.8385124221213752e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4378420.0,
      "step": 10067
    },
    {
      "epoch": 0.7744019690793016,
      "grad_norm": 19.0,
      "learning_rate": 1.8383970463810476e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4378738.0,
      "step": 10068
    },
    {
      "epoch": 0.77447888623952,
      "grad_norm": 16.25,
      "learning_rate": 1.8382816706407201e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4379044.0,
      "step": 10069
    },
    {
      "epoch": 0.7745558033997385,
      "grad_norm": 51.25,
      "learning_rate": 1.8381662949003924e-06,
      "loss": 0.0383,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4379300.0,
      "step": 10070
    },
    {
      "epoch": 0.7746327205599569,
      "grad_norm": 39.5,
      "learning_rate": 1.8380509191600644e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 4380126.0,
      "step": 10071
    },
    {
      "epoch": 0.7747096377201754,
      "grad_norm": 45.25,
      "learning_rate": 1.8379355434197369e-06,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9784172773361206,
      "num_tokens": 4380406.0,
      "step": 10072
    },
    {
      "epoch": 0.7747865548803938,
      "grad_norm": 18.0,
      "learning_rate": 1.8378201676794094e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 4381214.0,
      "step": 10073
    },
    {
      "epoch": 0.7748634720406122,
      "grad_norm": 54.75,
      "learning_rate": 1.8377047919390816e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4381560.0,
      "step": 10074
    },
    {
      "epoch": 0.7749403892008307,
      "grad_norm": 21.125,
      "learning_rate": 1.837589416198754e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4382130.0,
      "step": 10075
    },
    {
      "epoch": 0.7750173063610492,
      "grad_norm": 39.25,
      "learning_rate": 1.8374740404584266e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9971510171890259,
      "num_tokens": 4382482.0,
      "step": 10076
    },
    {
      "epoch": 0.7750942235212676,
      "grad_norm": 10.875,
      "learning_rate": 1.8373586647180986e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 4382798.0,
      "step": 10077
    },
    {
      "epoch": 0.7751711406814861,
      "grad_norm": 81.0,
      "learning_rate": 1.8372432889777709e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 4383056.0,
      "step": 10078
    },
    {
      "epoch": 0.7752480578417045,
      "grad_norm": 17.375,
      "learning_rate": 1.8371279132374433e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4383376.0,
      "step": 10079
    },
    {
      "epoch": 0.7753249750019229,
      "grad_norm": 27.625,
      "learning_rate": 1.8370125374971158e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4383722.0,
      "step": 10080
    },
    {
      "epoch": 0.7754018921621414,
      "grad_norm": 5.1875,
      "learning_rate": 1.836897161756788e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4384240.0,
      "step": 10081
    },
    {
      "epoch": 0.7754788093223598,
      "grad_norm": 19.75,
      "learning_rate": 1.8367817860164601e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 4385284.0,
      "step": 10082
    },
    {
      "epoch": 0.7755557264825783,
      "grad_norm": 4.71875,
      "learning_rate": 1.8366664102761326e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4385842.0,
      "step": 10083
    },
    {
      "epoch": 0.7756326436427967,
      "grad_norm": 14.8125,
      "learning_rate": 1.836551034535805e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 4386358.0,
      "step": 10084
    },
    {
      "epoch": 0.7757095608030151,
      "grad_norm": 23.875,
      "learning_rate": 1.8364356587954773e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 4386708.0,
      "step": 10085
    },
    {
      "epoch": 0.7757864779632336,
      "grad_norm": 35.25,
      "learning_rate": 1.8363202830551498e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 4387072.0,
      "step": 10086
    },
    {
      "epoch": 0.775863395123452,
      "grad_norm": 12.8125,
      "learning_rate": 1.8362049073148218e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9968253970146179,
      "num_tokens": 4387390.0,
      "step": 10087
    },
    {
      "epoch": 0.7759403122836704,
      "grad_norm": 20.25,
      "learning_rate": 1.8360895315744943e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4387712.0,
      "step": 10088
    },
    {
      "epoch": 0.776017229443889,
      "grad_norm": 66.5,
      "learning_rate": 1.8359741558341665e-06,
      "loss": 0.043,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 4388014.0,
      "step": 10089
    },
    {
      "epoch": 0.7760941466041074,
      "grad_norm": 18.75,
      "learning_rate": 1.835858780093839e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 4388296.0,
      "step": 10090
    },
    {
      "epoch": 0.7761710637643259,
      "grad_norm": 3.203125,
      "learning_rate": 1.8357434043535115e-06,
      "loss": 0.0067,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 4389100.0,
      "step": 10091
    },
    {
      "epoch": 0.7762479809245443,
      "grad_norm": 21.125,
      "learning_rate": 1.8356280286131837e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4389356.0,
      "step": 10092
    },
    {
      "epoch": 0.7763248980847627,
      "grad_norm": 6.96875,
      "learning_rate": 1.8355126528728558e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 4389680.0,
      "step": 10093
    },
    {
      "epoch": 0.7764018152449812,
      "grad_norm": 53.25,
      "learning_rate": 1.8353972771325283e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4390024.0,
      "step": 10094
    },
    {
      "epoch": 0.7764787324051996,
      "grad_norm": 33.5,
      "learning_rate": 1.8352819013922007e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4390542.0,
      "step": 10095
    },
    {
      "epoch": 0.776555649565418,
      "grad_norm": 49.0,
      "learning_rate": 1.835166525651873e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9853479862213135,
      "num_tokens": 4390816.0,
      "step": 10096
    },
    {
      "epoch": 0.7766325667256365,
      "grad_norm": 10.75,
      "learning_rate": 1.8350511499115455e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4391356.0,
      "step": 10097
    },
    {
      "epoch": 0.7767094838858549,
      "grad_norm": 27.0,
      "learning_rate": 1.8349357741712175e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 4392406.0,
      "step": 10098
    },
    {
      "epoch": 0.7767864010460733,
      "grad_norm": 12.3125,
      "learning_rate": 1.83482039843089e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 4392946.0,
      "step": 10099
    },
    {
      "epoch": 0.7768633182062918,
      "grad_norm": 13.3125,
      "learning_rate": 1.8347050226905622e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 4393508.0,
      "step": 10100
    },
    {
      "epoch": 0.7769402353665102,
      "grad_norm": 28.125,
      "learning_rate": 1.8345896469502347e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9903661012649536,
      "num_tokens": 4394028.0,
      "step": 10101
    },
    {
      "epoch": 0.7770171525267288,
      "grad_norm": 4.125,
      "learning_rate": 1.8344742712099072e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4394546.0,
      "step": 10102
    },
    {
      "epoch": 0.7770940696869472,
      "grad_norm": 29.0,
      "learning_rate": 1.8343588954695794e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4394882.0,
      "step": 10103
    },
    {
      "epoch": 0.7771709868471656,
      "grad_norm": 31.375,
      "learning_rate": 1.8342435197292515e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 4395670.0,
      "step": 10104
    },
    {
      "epoch": 0.7772479040073841,
      "grad_norm": 41.75,
      "learning_rate": 1.834128143988924e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4395924.0,
      "step": 10105
    },
    {
      "epoch": 0.7773248211676025,
      "grad_norm": 4.1875,
      "learning_rate": 1.8340127682485964e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4396510.0,
      "step": 10106
    },
    {
      "epoch": 0.7774017383278209,
      "grad_norm": 43.0,
      "learning_rate": 1.8338973925082687e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 4396850.0,
      "step": 10107
    },
    {
      "epoch": 0.7774786554880394,
      "grad_norm": 78.0,
      "learning_rate": 1.8337820167679411e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.9774919748306274,
      "num_tokens": 4397164.0,
      "step": 10108
    },
    {
      "epoch": 0.7775555726482578,
      "grad_norm": 14.3125,
      "learning_rate": 1.8336666410276132e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4397460.0,
      "step": 10109
    },
    {
      "epoch": 0.7776324898084763,
      "grad_norm": 39.5,
      "learning_rate": 1.8335512652872857e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 4397724.0,
      "step": 10110
    },
    {
      "epoch": 0.7777094069686947,
      "grad_norm": 10.375,
      "learning_rate": 1.833435889546958e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4398028.0,
      "step": 10111
    },
    {
      "epoch": 0.7777863241289131,
      "grad_norm": 27.0,
      "learning_rate": 1.8333205138066304e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4398334.0,
      "step": 10112
    },
    {
      "epoch": 0.7778632412891316,
      "grad_norm": 9.125,
      "learning_rate": 1.8332051380663026e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9962616562843323,
      "num_tokens": 4398870.0,
      "step": 10113
    },
    {
      "epoch": 0.77794015844935,
      "grad_norm": 26.5,
      "learning_rate": 1.833089762325975e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4399192.0,
      "step": 10114
    },
    {
      "epoch": 0.7780170756095685,
      "grad_norm": 3.078125,
      "learning_rate": 1.8329743865856472e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4399710.0,
      "step": 10115
    },
    {
      "epoch": 0.778093992769787,
      "grad_norm": 9.8125,
      "learning_rate": 1.8328590108453196e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9950900077819824,
      "num_tokens": 4400322.0,
      "step": 10116
    },
    {
      "epoch": 0.7781709099300054,
      "grad_norm": 14.0625,
      "learning_rate": 1.8327436351049919e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4400896.0,
      "step": 10117
    },
    {
      "epoch": 0.7782478270902238,
      "grad_norm": 90.5,
      "learning_rate": 1.8326282593646644e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 4401256.0,
      "step": 10118
    },
    {
      "epoch": 0.7783247442504423,
      "grad_norm": 33.0,
      "learning_rate": 1.8325128836243368e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9950494766235352,
      "num_tokens": 4402066.0,
      "step": 10119
    },
    {
      "epoch": 0.7784016614106607,
      "grad_norm": 11.125,
      "learning_rate": 1.8323975078840089e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 4402432.0,
      "step": 10120
    },
    {
      "epoch": 0.7784785785708792,
      "grad_norm": 7.8125,
      "learning_rate": 1.8322821321436811e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4402988.0,
      "step": 10121
    },
    {
      "epoch": 0.7785554957310976,
      "grad_norm": 27.75,
      "learning_rate": 1.8321667564033536e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 4403274.0,
      "step": 10122
    },
    {
      "epoch": 0.778632412891316,
      "grad_norm": 4.71875,
      "learning_rate": 1.832051380663026e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 4403874.0,
      "step": 10123
    },
    {
      "epoch": 0.7787093300515345,
      "grad_norm": 6.71875,
      "learning_rate": 1.8319360049226983e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 4404236.0,
      "step": 10124
    },
    {
      "epoch": 0.7787862472117529,
      "grad_norm": 37.5,
      "learning_rate": 1.8318206291823704e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9893428087234497,
      "num_tokens": 4404800.0,
      "step": 10125
    },
    {
      "epoch": 0.7788631643719713,
      "grad_norm": 11.6875,
      "learning_rate": 1.8317052534420429e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4405350.0,
      "step": 10126
    },
    {
      "epoch": 0.7789400815321899,
      "grad_norm": 22.375,
      "learning_rate": 1.8315898777017153e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 4405620.0,
      "step": 10127
    },
    {
      "epoch": 0.7790169986924083,
      "grad_norm": 10.8125,
      "learning_rate": 1.8314745019613876e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4406200.0,
      "step": 10128
    },
    {
      "epoch": 0.7790939158526268,
      "grad_norm": 10.875,
      "learning_rate": 1.83135912622106e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 4406544.0,
      "step": 10129
    },
    {
      "epoch": 0.7791708330128452,
      "grad_norm": 9.375,
      "learning_rate": 1.8312437504807325e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 4407118.0,
      "step": 10130
    },
    {
      "epoch": 0.7792477501730636,
      "grad_norm": 46.75,
      "learning_rate": 1.8311283747404046e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4407434.0,
      "step": 10131
    },
    {
      "epoch": 0.7793246673332821,
      "grad_norm": 23.125,
      "learning_rate": 1.8310129990000768e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 4408248.0,
      "step": 10132
    },
    {
      "epoch": 0.7794015844935005,
      "grad_norm": 8.8125,
      "learning_rate": 1.8308976232597493e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 4408832.0,
      "step": 10133
    },
    {
      "epoch": 0.7794785016537189,
      "grad_norm": 10.875,
      "learning_rate": 1.8307822475194218e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4409382.0,
      "step": 10134
    },
    {
      "epoch": 0.7795554188139374,
      "grad_norm": 45.25,
      "learning_rate": 1.830666871779094e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4409946.0,
      "step": 10135
    },
    {
      "epoch": 0.7796323359741558,
      "grad_norm": 36.25,
      "learning_rate": 1.830551496038766e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 4410282.0,
      "step": 10136
    },
    {
      "epoch": 0.7797092531343742,
      "grad_norm": 36.25,
      "learning_rate": 1.8304361202984385e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4411090.0,
      "step": 10137
    },
    {
      "epoch": 0.7797861702945927,
      "grad_norm": 13.8125,
      "learning_rate": 1.830320744558111e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4411636.0,
      "step": 10138
    },
    {
      "epoch": 0.7798630874548111,
      "grad_norm": 30.625,
      "learning_rate": 1.8302053688177833e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 4411972.0,
      "step": 10139
    },
    {
      "epoch": 0.7799400046150297,
      "grad_norm": 8.5625,
      "learning_rate": 1.8300899930774557e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9972899556159973,
      "num_tokens": 4412344.0,
      "step": 10140
    },
    {
      "epoch": 0.7800169217752481,
      "grad_norm": 24.75,
      "learning_rate": 1.8299746173371282e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 4412648.0,
      "step": 10141
    },
    {
      "epoch": 0.7800938389354665,
      "grad_norm": 12.3125,
      "learning_rate": 1.8298592415968003e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4412934.0,
      "step": 10142
    },
    {
      "epoch": 0.780170756095685,
      "grad_norm": 12.5625,
      "learning_rate": 1.8297438658564725e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 4413760.0,
      "step": 10143
    },
    {
      "epoch": 0.7802476732559034,
      "grad_norm": 50.25,
      "learning_rate": 1.829628490116145e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4414046.0,
      "step": 10144
    },
    {
      "epoch": 0.7803245904161218,
      "grad_norm": 18.75,
      "learning_rate": 1.8295131143758174e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4414610.0,
      "step": 10145
    },
    {
      "epoch": 0.7804015075763403,
      "grad_norm": 29.125,
      "learning_rate": 1.8293977386354897e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 4414910.0,
      "step": 10146
    },
    {
      "epoch": 0.7804784247365587,
      "grad_norm": 39.25,
      "learning_rate": 1.8292823628951618e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9915013909339905,
      "num_tokens": 4415264.0,
      "step": 10147
    },
    {
      "epoch": 0.7805553418967772,
      "grad_norm": 6.875,
      "learning_rate": 1.8291669871548342e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 4415588.0,
      "step": 10148
    },
    {
      "epoch": 0.7806322590569956,
      "grad_norm": 25.25,
      "learning_rate": 1.8290516114145067e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 4416178.0,
      "step": 10149
    },
    {
      "epoch": 0.780709176217214,
      "grad_norm": 151.0,
      "learning_rate": 1.828936235674179e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 4416502.0,
      "step": 10150
    },
    {
      "epoch": 0.7807860933774325,
      "grad_norm": 3.796875,
      "learning_rate": 1.8288208599338514e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4417050.0,
      "step": 10151
    },
    {
      "epoch": 0.780863010537651,
      "grad_norm": 10.5,
      "learning_rate": 1.8287054841935235e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 4417354.0,
      "step": 10152
    },
    {
      "epoch": 0.7809399276978694,
      "grad_norm": 19.75,
      "learning_rate": 1.828590108453196e-06,
      "loss": 0.0355,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4417630.0,
      "step": 10153
    },
    {
      "epoch": 0.7810168448580879,
      "grad_norm": 36.75,
      "learning_rate": 1.8284747327128682e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4417976.0,
      "step": 10154
    },
    {
      "epoch": 0.7810937620183063,
      "grad_norm": 58.0,
      "learning_rate": 1.8283593569725407e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.993127167224884,
      "num_tokens": 4418268.0,
      "step": 10155
    },
    {
      "epoch": 0.7811706791785247,
      "grad_norm": 33.25,
      "learning_rate": 1.8282439812322131e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 4418588.0,
      "step": 10156
    },
    {
      "epoch": 0.7812475963387432,
      "grad_norm": 21.875,
      "learning_rate": 1.8281286054918854e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4419160.0,
      "step": 10157
    },
    {
      "epoch": 0.7813245134989616,
      "grad_norm": 6.46875,
      "learning_rate": 1.8280132297515574e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 4419946.0,
      "step": 10158
    },
    {
      "epoch": 0.7814014306591801,
      "grad_norm": 3.28125,
      "learning_rate": 1.82789785401123e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 4420510.0,
      "step": 10159
    },
    {
      "epoch": 0.7814783478193985,
      "grad_norm": 12.9375,
      "learning_rate": 1.8277824782709024e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4421062.0,
      "step": 10160
    },
    {
      "epoch": 0.7815552649796169,
      "grad_norm": 17.625,
      "learning_rate": 1.8276671025305746e-06,
      "loss": 0.0421,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4421340.0,
      "step": 10161
    },
    {
      "epoch": 0.7816321821398354,
      "grad_norm": 47.5,
      "learning_rate": 1.8275517267902471e-06,
      "loss": 0.0425,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 4421574.0,
      "step": 10162
    },
    {
      "epoch": 0.7817090993000538,
      "grad_norm": 18.5,
      "learning_rate": 1.8274363510499192e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4422122.0,
      "step": 10163
    },
    {
      "epoch": 0.7817860164602722,
      "grad_norm": 48.0,
      "learning_rate": 1.8273209753095916e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 4422724.0,
      "step": 10164
    },
    {
      "epoch": 0.7818629336204908,
      "grad_norm": 3.46875,
      "learning_rate": 1.8272055995692639e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.9974293112754822,
      "num_tokens": 4423504.0,
      "step": 10165
    },
    {
      "epoch": 0.7819398507807092,
      "grad_norm": 10.3125,
      "learning_rate": 1.8270902238289364e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 4424076.0,
      "step": 10166
    },
    {
      "epoch": 0.7820167679409277,
      "grad_norm": 15.9375,
      "learning_rate": 1.8269748480886086e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4424662.0,
      "step": 10167
    },
    {
      "epoch": 0.7820936851011461,
      "grad_norm": 19.0,
      "learning_rate": 1.826859472348281e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4425246.0,
      "step": 10168
    },
    {
      "epoch": 0.7821706022613645,
      "grad_norm": 7.6875,
      "learning_rate": 1.8267440966079531e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9982547760009766,
      "num_tokens": 4425820.0,
      "step": 10169
    },
    {
      "epoch": 0.782247519421583,
      "grad_norm": 72.5,
      "learning_rate": 1.8266287208676256e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4426130.0,
      "step": 10170
    },
    {
      "epoch": 0.7823244365818014,
      "grad_norm": 17.5,
      "learning_rate": 1.8265133451272979e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4426648.0,
      "step": 10171
    },
    {
      "epoch": 0.7824013537420198,
      "grad_norm": 14.5,
      "learning_rate": 1.8263979693869703e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9974424839019775,
      "num_tokens": 4427432.0,
      "step": 10172
    },
    {
      "epoch": 0.7824782709022383,
      "grad_norm": 3.125,
      "learning_rate": 1.8262825936466428e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4427978.0,
      "step": 10173
    },
    {
      "epoch": 0.7825551880624567,
      "grad_norm": 4.75,
      "learning_rate": 1.8261672179063148e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4428562.0,
      "step": 10174
    },
    {
      "epoch": 0.7826321052226752,
      "grad_norm": 19.75,
      "learning_rate": 1.826051842165987e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4428856.0,
      "step": 10175
    },
    {
      "epoch": 0.7827090223828936,
      "grad_norm": 13.125,
      "learning_rate": 1.8259364664256596e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4429156.0,
      "step": 10176
    },
    {
      "epoch": 0.782785939543112,
      "grad_norm": 4.9375,
      "learning_rate": 1.825821090685332e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4429714.0,
      "step": 10177
    },
    {
      "epoch": 0.7828628567033306,
      "grad_norm": 8.75,
      "learning_rate": 1.8257057149450043e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4430036.0,
      "step": 10178
    },
    {
      "epoch": 0.782939773863549,
      "grad_norm": 3.921875,
      "learning_rate": 1.8255903392046768e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4430604.0,
      "step": 10179
    },
    {
      "epoch": 0.7830166910237674,
      "grad_norm": 23.125,
      "learning_rate": 1.8254749634643488e-06,
      "loss": 0.0373,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4430882.0,
      "step": 10180
    },
    {
      "epoch": 0.7830936081839859,
      "grad_norm": 80.0,
      "learning_rate": 1.8253595877240213e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4431178.0,
      "step": 10181
    },
    {
      "epoch": 0.7831705253442043,
      "grad_norm": 46.75,
      "learning_rate": 1.8252442119836935e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 4431494.0,
      "step": 10182
    },
    {
      "epoch": 0.7832474425044227,
      "grad_norm": 32.75,
      "learning_rate": 1.825128836243366e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4431826.0,
      "step": 10183
    },
    {
      "epoch": 0.7833243596646412,
      "grad_norm": 14.0625,
      "learning_rate": 1.8250134605030385e-06,
      "loss": 0.0067,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 4432876.0,
      "step": 10184
    },
    {
      "epoch": 0.7834012768248596,
      "grad_norm": 9.25,
      "learning_rate": 1.8248980847627105e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4433460.0,
      "step": 10185
    },
    {
      "epoch": 0.7834781939850781,
      "grad_norm": 11.0,
      "learning_rate": 1.8247827090223828e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 4434028.0,
      "step": 10186
    },
    {
      "epoch": 0.7835551111452965,
      "grad_norm": 42.0,
      "learning_rate": 1.8246673332820553e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4434322.0,
      "step": 10187
    },
    {
      "epoch": 0.7836320283055149,
      "grad_norm": 12.3125,
      "learning_rate": 1.8245519575417277e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 4434612.0,
      "step": 10188
    },
    {
      "epoch": 0.7837089454657334,
      "grad_norm": 9.5625,
      "learning_rate": 1.8244365818014e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4434946.0,
      "step": 10189
    },
    {
      "epoch": 0.7837858626259518,
      "grad_norm": 8.8125,
      "learning_rate": 1.824321206061072e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4435492.0,
      "step": 10190
    },
    {
      "epoch": 0.7838627797861702,
      "grad_norm": 30.875,
      "learning_rate": 1.8242058303207445e-06,
      "loss": 0.0361,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4435770.0,
      "step": 10191
    },
    {
      "epoch": 0.7839396969463888,
      "grad_norm": 7.84375,
      "learning_rate": 1.824090454580417e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4436354.0,
      "step": 10192
    },
    {
      "epoch": 0.7840166141066072,
      "grad_norm": 4.21875,
      "learning_rate": 1.8239750788400892e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9942857027053833,
      "num_tokens": 4436880.0,
      "step": 10193
    },
    {
      "epoch": 0.7840935312668257,
      "grad_norm": 11.25,
      "learning_rate": 1.8238597030997617e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4437400.0,
      "step": 10194
    },
    {
      "epoch": 0.7841704484270441,
      "grad_norm": 43.5,
      "learning_rate": 1.8237443273594342e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4437716.0,
      "step": 10195
    },
    {
      "epoch": 0.7842473655872625,
      "grad_norm": 24.25,
      "learning_rate": 1.8236289516191062e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 4437996.0,
      "step": 10196
    },
    {
      "epoch": 0.784324282747481,
      "grad_norm": 18.125,
      "learning_rate": 1.8235135758787785e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4438514.0,
      "step": 10197
    },
    {
      "epoch": 0.7844011999076994,
      "grad_norm": 109.0,
      "learning_rate": 1.823398200138451e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4438878.0,
      "step": 10198
    },
    {
      "epoch": 0.7844781170679178,
      "grad_norm": 32.5,
      "learning_rate": 1.8232828243981234e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 4439166.0,
      "step": 10199
    },
    {
      "epoch": 0.7845550342281363,
      "grad_norm": 20.125,
      "learning_rate": 1.8231674486577957e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 4439922.0,
      "step": 10200
    },
    {
      "epoch": 0.7846319513883547,
      "grad_norm": 43.0,
      "learning_rate": 1.8230520729174677e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4440440.0,
      "step": 10201
    },
    {
      "epoch": 0.7847088685485731,
      "grad_norm": 23.875,
      "learning_rate": 1.8229366971771402e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9961904883384705,
      "num_tokens": 4440966.0,
      "step": 10202
    },
    {
      "epoch": 0.7847857857087916,
      "grad_norm": 4.84375,
      "learning_rate": 1.8228213214368127e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4441506.0,
      "step": 10203
    },
    {
      "epoch": 0.78486270286901,
      "grad_norm": 10.75,
      "learning_rate": 1.822705945696485e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4441818.0,
      "step": 10204
    },
    {
      "epoch": 0.7849396200292286,
      "grad_norm": 15.8125,
      "learning_rate": 1.8225905699561574e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4442128.0,
      "step": 10205
    },
    {
      "epoch": 0.785016537189447,
      "grad_norm": 11.9375,
      "learning_rate": 1.8224751942158299e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4442458.0,
      "step": 10206
    },
    {
      "epoch": 0.7850934543496654,
      "grad_norm": 27.125,
      "learning_rate": 1.822359818475502e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4442788.0,
      "step": 10207
    },
    {
      "epoch": 0.7851703715098839,
      "grad_norm": 16.5,
      "learning_rate": 1.8222444427351742e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.990439772605896,
      "num_tokens": 4443312.0,
      "step": 10208
    },
    {
      "epoch": 0.7852472886701023,
      "grad_norm": 20.25,
      "learning_rate": 1.8221290669948466e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.982758641242981,
      "num_tokens": 4443548.0,
      "step": 10209
    },
    {
      "epoch": 0.7853242058303207,
      "grad_norm": 44.0,
      "learning_rate": 1.822013691254519e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4443892.0,
      "step": 10210
    },
    {
      "epoch": 0.7854011229905392,
      "grad_norm": 19.75,
      "learning_rate": 1.8218983155141914e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 4444424.0,
      "step": 10211
    },
    {
      "epoch": 0.7854780401507576,
      "grad_norm": 51.25,
      "learning_rate": 1.8217829397738634e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 4444740.0,
      "step": 10212
    },
    {
      "epoch": 0.7855549573109761,
      "grad_norm": 25.125,
      "learning_rate": 1.8216675640335359e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4445288.0,
      "step": 10213
    },
    {
      "epoch": 0.7856318744711945,
      "grad_norm": 72.0,
      "learning_rate": 1.8215521882932083e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4445616.0,
      "step": 10214
    },
    {
      "epoch": 0.7857087916314129,
      "grad_norm": 22.125,
      "learning_rate": 1.8214368125528806e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 4445926.0,
      "step": 10215
    },
    {
      "epoch": 0.7857857087916315,
      "grad_norm": 13.8125,
      "learning_rate": 1.821321436812553e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4446232.0,
      "step": 10216
    },
    {
      "epoch": 0.7858626259518499,
      "grad_norm": 30.375,
      "learning_rate": 1.8212060610722253e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 4446554.0,
      "step": 10217
    },
    {
      "epoch": 0.7859395431120683,
      "grad_norm": 10.5625,
      "learning_rate": 1.8210906853318976e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4446836.0,
      "step": 10218
    },
    {
      "epoch": 0.7860164602722868,
      "grad_norm": 6.15625,
      "learning_rate": 1.8209753095915698e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 4447662.0,
      "step": 10219
    },
    {
      "epoch": 0.7860933774325052,
      "grad_norm": 8.875,
      "learning_rate": 1.8208599338512423e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4448180.0,
      "step": 10220
    },
    {
      "epoch": 0.7861702945927236,
      "grad_norm": 54.0,
      "learning_rate": 1.8207445581109146e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4448458.0,
      "step": 10221
    },
    {
      "epoch": 0.7862472117529421,
      "grad_norm": 34.5,
      "learning_rate": 1.820629182370587e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4448774.0,
      "step": 10222
    },
    {
      "epoch": 0.7863241289131605,
      "grad_norm": 8.8125,
      "learning_rate": 1.820513806630259e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.997183084487915,
      "num_tokens": 4449130.0,
      "step": 10223
    },
    {
      "epoch": 0.786401046073379,
      "grad_norm": 22.75,
      "learning_rate": 1.8203984308899316e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4449446.0,
      "step": 10224
    },
    {
      "epoch": 0.7864779632335974,
      "grad_norm": 15.5,
      "learning_rate": 1.8202830551496038e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 4450014.0,
      "step": 10225
    },
    {
      "epoch": 0.7865548803938158,
      "grad_norm": 44.5,
      "learning_rate": 1.8201676794092763e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 4450314.0,
      "step": 10226
    },
    {
      "epoch": 0.7866317975540343,
      "grad_norm": 4.21875,
      "learning_rate": 1.8200523036689488e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4450872.0,
      "step": 10227
    },
    {
      "epoch": 0.7867087147142527,
      "grad_norm": 23.875,
      "learning_rate": 1.8199369279286208e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 4451228.0,
      "step": 10228
    },
    {
      "epoch": 0.7867856318744711,
      "grad_norm": 15.875,
      "learning_rate": 1.819821552188293e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4451560.0,
      "step": 10229
    },
    {
      "epoch": 0.7868625490346897,
      "grad_norm": 38.75,
      "learning_rate": 1.8197061764479655e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9966158866882324,
      "num_tokens": 4452152.0,
      "step": 10230
    },
    {
      "epoch": 0.7869394661949081,
      "grad_norm": 14.125,
      "learning_rate": 1.819590800707638e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 4452704.0,
      "step": 10231
    },
    {
      "epoch": 0.7870163833551266,
      "grad_norm": 55.0,
      "learning_rate": 1.8194754249673103e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 4452992.0,
      "step": 10232
    },
    {
      "epoch": 0.787093300515345,
      "grad_norm": 51.5,
      "learning_rate": 1.8193600492269827e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4453308.0,
      "step": 10233
    },
    {
      "epoch": 0.7871702176755634,
      "grad_norm": 94.0,
      "learning_rate": 1.8192446734866548e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4453620.0,
      "step": 10234
    },
    {
      "epoch": 0.7872471348357819,
      "grad_norm": 84.0,
      "learning_rate": 1.8191292977463272e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4453976.0,
      "step": 10235
    },
    {
      "epoch": 0.7873240519960003,
      "grad_norm": 8.125,
      "learning_rate": 1.8190139220059995e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9982425570487976,
      "num_tokens": 4454546.0,
      "step": 10236
    },
    {
      "epoch": 0.7874009691562187,
      "grad_norm": 30.875,
      "learning_rate": 1.818898546265672e-06,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4454802.0,
      "step": 10237
    },
    {
      "epoch": 0.7874778863164372,
      "grad_norm": 24.25,
      "learning_rate": 1.8187831705253444e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4455112.0,
      "step": 10238
    },
    {
      "epoch": 0.7875548034766556,
      "grad_norm": 4.0,
      "learning_rate": 1.8186677947850165e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 4455646.0,
      "step": 10239
    },
    {
      "epoch": 0.787631720636874,
      "grad_norm": 30.625,
      "learning_rate": 1.8185524190446887e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4455958.0,
      "step": 10240
    },
    {
      "epoch": 0.7877086377970925,
      "grad_norm": 15.75,
      "learning_rate": 1.8184370433043612e-06,
      "loss": 0.0349,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 4456216.0,
      "step": 10241
    },
    {
      "epoch": 0.787785554957311,
      "grad_norm": 44.5,
      "learning_rate": 1.8183216675640337e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4456556.0,
      "step": 10242
    },
    {
      "epoch": 0.7878624721175295,
      "grad_norm": 26.875,
      "learning_rate": 1.818206291823706e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 4456862.0,
      "step": 10243
    },
    {
      "epoch": 0.7879393892777479,
      "grad_norm": 13.75,
      "learning_rate": 1.8180909160833784e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4457172.0,
      "step": 10244
    },
    {
      "epoch": 0.7880163064379663,
      "grad_norm": 12.1875,
      "learning_rate": 1.8179755403430505e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4457722.0,
      "step": 10245
    },
    {
      "epoch": 0.7880932235981848,
      "grad_norm": 13.0,
      "learning_rate": 1.817860164602723e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 4458280.0,
      "step": 10246
    },
    {
      "epoch": 0.7881701407584032,
      "grad_norm": 13.125,
      "learning_rate": 1.8177447888623952e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9925925731658936,
      "num_tokens": 4459092.0,
      "step": 10247
    },
    {
      "epoch": 0.7882470579186216,
      "grad_norm": 15.125,
      "learning_rate": 1.8176294131220677e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4459638.0,
      "step": 10248
    },
    {
      "epoch": 0.7883239750788401,
      "grad_norm": 7.625,
      "learning_rate": 1.8175140373817401e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4459978.0,
      "step": 10249
    },
    {
      "epoch": 0.7884008922390585,
      "grad_norm": 20.625,
      "learning_rate": 1.8173986616414122e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4460316.0,
      "step": 10250
    },
    {
      "epoch": 0.788477809399277,
      "grad_norm": 9.1875,
      "learning_rate": 1.8172832859010844e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.995110034942627,
      "num_tokens": 4461136.0,
      "step": 10251
    },
    {
      "epoch": 0.7885547265594954,
      "grad_norm": 51.5,
      "learning_rate": 1.817167910160757e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4461422.0,
      "step": 10252
    },
    {
      "epoch": 0.7886316437197138,
      "grad_norm": 12.8125,
      "learning_rate": 1.8170525344204294e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 4461942.0,
      "step": 10253
    },
    {
      "epoch": 0.7887085608799324,
      "grad_norm": 50.5,
      "learning_rate": 1.8169371586801016e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.980327844619751,
      "num_tokens": 4462248.0,
      "step": 10254
    },
    {
      "epoch": 0.7887854780401508,
      "grad_norm": 4.25,
      "learning_rate": 1.8168217829397737e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4462548.0,
      "step": 10255
    },
    {
      "epoch": 0.7888623952003692,
      "grad_norm": 6.71875,
      "learning_rate": 1.8167064071994461e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4463066.0,
      "step": 10256
    },
    {
      "epoch": 0.7889393123605877,
      "grad_norm": 6.09375,
      "learning_rate": 1.8165910314591186e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4463628.0,
      "step": 10257
    },
    {
      "epoch": 0.7890162295208061,
      "grad_norm": 48.75,
      "learning_rate": 1.8164756557187909e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4463972.0,
      "step": 10258
    },
    {
      "epoch": 0.7890931466810245,
      "grad_norm": 58.25,
      "learning_rate": 1.8163602799784633e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4464276.0,
      "step": 10259
    },
    {
      "epoch": 0.789170063841243,
      "grad_norm": 18.0,
      "learning_rate": 1.8162449042381358e-06,
      "loss": 0.0295,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4464560.0,
      "step": 10260
    },
    {
      "epoch": 0.7892469810014614,
      "grad_norm": 4.0625,
      "learning_rate": 1.8161295284978079e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 4465122.0,
      "step": 10261
    },
    {
      "epoch": 0.7893238981616799,
      "grad_norm": 15.8125,
      "learning_rate": 1.8160141527574801e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 4465442.0,
      "step": 10262
    },
    {
      "epoch": 0.7894008153218983,
      "grad_norm": 80.0,
      "learning_rate": 1.8158987770171526e-06,
      "loss": 0.0451,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 4465742.0,
      "step": 10263
    },
    {
      "epoch": 0.7894777324821167,
      "grad_norm": 5.46875,
      "learning_rate": 1.815783401276825e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 4466562.0,
      "step": 10264
    },
    {
      "epoch": 0.7895546496423352,
      "grad_norm": 6.6875,
      "learning_rate": 1.8156680255364973e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 4466874.0,
      "step": 10265
    },
    {
      "epoch": 0.7896315668025536,
      "grad_norm": 12.4375,
      "learning_rate": 1.8155526497961694e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.997183084487915,
      "num_tokens": 4467230.0,
      "step": 10266
    },
    {
      "epoch": 0.789708483962772,
      "grad_norm": 29.125,
      "learning_rate": 1.8154372740558418e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 4467510.0,
      "step": 10267
    },
    {
      "epoch": 0.7897854011229906,
      "grad_norm": 13.4375,
      "learning_rate": 1.8153218983155143e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4468096.0,
      "step": 10268
    },
    {
      "epoch": 0.789862318283209,
      "grad_norm": 11.0625,
      "learning_rate": 1.8152065225751866e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4468428.0,
      "step": 10269
    },
    {
      "epoch": 0.7899392354434275,
      "grad_norm": 12.8125,
      "learning_rate": 1.815091146834859e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4469002.0,
      "step": 10270
    },
    {
      "epoch": 0.7900161526036459,
      "grad_norm": 14.0,
      "learning_rate": 1.8149757710945313e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 4469314.0,
      "step": 10271
    },
    {
      "epoch": 0.7900930697638643,
      "grad_norm": 11.0625,
      "learning_rate": 1.8148603953542035e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 4469616.0,
      "step": 10272
    },
    {
      "epoch": 0.7901699869240828,
      "grad_norm": 15.1875,
      "learning_rate": 1.8147450196138758e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 4470176.0,
      "step": 10273
    },
    {
      "epoch": 0.7902469040843012,
      "grad_norm": 40.0,
      "learning_rate": 1.8146296438735483e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4470494.0,
      "step": 10274
    },
    {
      "epoch": 0.7903238212445196,
      "grad_norm": 43.5,
      "learning_rate": 1.8145142681332205e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 4471078.0,
      "step": 10275
    },
    {
      "epoch": 0.7904007384047381,
      "grad_norm": 30.625,
      "learning_rate": 1.814398892392893e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4471648.0,
      "step": 10276
    },
    {
      "epoch": 0.7904776555649565,
      "grad_norm": 60.25,
      "learning_rate": 1.814283516652565e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 4472180.0,
      "step": 10277
    },
    {
      "epoch": 0.7905545727251749,
      "grad_norm": 54.75,
      "learning_rate": 1.8141681409122375e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9971510171890259,
      "num_tokens": 4472532.0,
      "step": 10278
    },
    {
      "epoch": 0.7906314898853934,
      "grad_norm": 7.71875,
      "learning_rate": 1.8140527651719098e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4473340.0,
      "step": 10279
    },
    {
      "epoch": 0.7907084070456118,
      "grad_norm": 10.0,
      "learning_rate": 1.8139373894315822e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4473904.0,
      "step": 10280
    },
    {
      "epoch": 0.7907853242058304,
      "grad_norm": 20.625,
      "learning_rate": 1.8138220136912547e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4474224.0,
      "step": 10281
    },
    {
      "epoch": 0.7908622413660488,
      "grad_norm": 6.25,
      "learning_rate": 1.813706637950927e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4474564.0,
      "step": 10282
    },
    {
      "epoch": 0.7909391585262672,
      "grad_norm": 16.25,
      "learning_rate": 1.813591262210599e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 4475368.0,
      "step": 10283
    },
    {
      "epoch": 0.7910160756864857,
      "grad_norm": 4.0,
      "learning_rate": 1.8134758864702715e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4475926.0,
      "step": 10284
    },
    {
      "epoch": 0.7910929928467041,
      "grad_norm": 3.578125,
      "learning_rate": 1.813360510729944e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 4476460.0,
      "step": 10285
    },
    {
      "epoch": 0.7911699100069225,
      "grad_norm": 9.9375,
      "learning_rate": 1.8132451349896162e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 4477216.0,
      "step": 10286
    },
    {
      "epoch": 0.791246827167141,
      "grad_norm": 13.25,
      "learning_rate": 1.8131297592492887e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4477776.0,
      "step": 10287
    },
    {
      "epoch": 0.7913237443273594,
      "grad_norm": 8.6875,
      "learning_rate": 1.8130143835089607e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 4478356.0,
      "step": 10288
    },
    {
      "epoch": 0.7914006614875779,
      "grad_norm": 10.75,
      "learning_rate": 1.8128990077686332e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 4478952.0,
      "step": 10289
    },
    {
      "epoch": 0.7914775786477963,
      "grad_norm": 6.6875,
      "learning_rate": 1.8127836320283055e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 4479486.0,
      "step": 10290
    },
    {
      "epoch": 0.7915544958080147,
      "grad_norm": 56.25,
      "learning_rate": 1.812668256287978e-06,
      "loss": 0.0434,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4479764.0,
      "step": 10291
    },
    {
      "epoch": 0.7916314129682333,
      "grad_norm": 57.75,
      "learning_rate": 1.8125528805476504e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 4480078.0,
      "step": 10292
    },
    {
      "epoch": 0.7917083301284517,
      "grad_norm": 5.15625,
      "learning_rate": 1.8124375048073224e-06,
      "loss": 0.0072,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 4480858.0,
      "step": 10293
    },
    {
      "epoch": 0.7917852472886701,
      "grad_norm": 91.0,
      "learning_rate": 1.8123221290669947e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4481140.0,
      "step": 10294
    },
    {
      "epoch": 0.7918621644488886,
      "grad_norm": 31.5,
      "learning_rate": 1.8122067533266672e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 4481424.0,
      "step": 10295
    },
    {
      "epoch": 0.791939081609107,
      "grad_norm": 20.0,
      "learning_rate": 1.8120913775863396e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 4481986.0,
      "step": 10296
    },
    {
      "epoch": 0.7920159987693255,
      "grad_norm": 84.5,
      "learning_rate": 1.811976001846012e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4482284.0,
      "step": 10297
    },
    {
      "epoch": 0.7920929159295439,
      "grad_norm": 8.4375,
      "learning_rate": 1.8118606261056844e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9903846383094788,
      "num_tokens": 4482598.0,
      "step": 10298
    },
    {
      "epoch": 0.7921698330897623,
      "grad_norm": 54.0,
      "learning_rate": 1.8117452503653564e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9826839566230774,
      "num_tokens": 4482832.0,
      "step": 10299
    },
    {
      "epoch": 0.7922467502499808,
      "grad_norm": 38.75,
      "learning_rate": 1.8116298746250289e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4483396.0,
      "step": 10300
    },
    {
      "epoch": 0.7923236674101992,
      "grad_norm": 8.125,
      "learning_rate": 1.8115144988847011e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4483948.0,
      "step": 10301
    },
    {
      "epoch": 0.7924005845704176,
      "grad_norm": 10.0,
      "learning_rate": 1.8113991231443736e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4484284.0,
      "step": 10302
    },
    {
      "epoch": 0.7924775017306361,
      "grad_norm": 40.0,
      "learning_rate": 1.811283747404046e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 4484546.0,
      "step": 10303
    },
    {
      "epoch": 0.7925544188908545,
      "grad_norm": 46.0,
      "learning_rate": 1.8111683716637181e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4484866.0,
      "step": 10304
    },
    {
      "epoch": 0.7926313360510729,
      "grad_norm": 11.125,
      "learning_rate": 1.8110529959233904e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4485412.0,
      "step": 10305
    },
    {
      "epoch": 0.7927082532112915,
      "grad_norm": 23.875,
      "learning_rate": 1.8109376201830629e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 4485904.0,
      "step": 10306
    },
    {
      "epoch": 0.7927851703715099,
      "grad_norm": 38.75,
      "learning_rate": 1.8108222444427353e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4486450.0,
      "step": 10307
    },
    {
      "epoch": 0.7928620875317284,
      "grad_norm": 24.5,
      "learning_rate": 1.8107068687024076e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 4486818.0,
      "step": 10308
    },
    {
      "epoch": 0.7929390046919468,
      "grad_norm": 4.3125,
      "learning_rate": 1.81059149296208e-06,
      "loss": 0.0059,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4487578.0,
      "step": 10309
    },
    {
      "epoch": 0.7930159218521652,
      "grad_norm": 13.625,
      "learning_rate": 1.8104761172217521e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4487832.0,
      "step": 10310
    },
    {
      "epoch": 0.7930928390123837,
      "grad_norm": 13.6875,
      "learning_rate": 1.8103607414814246e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4488162.0,
      "step": 10311
    },
    {
      "epoch": 0.7931697561726021,
      "grad_norm": 3.921875,
      "learning_rate": 1.8102453657410968e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 4488948.0,
      "step": 10312
    },
    {
      "epoch": 0.7932466733328205,
      "grad_norm": 8.125,
      "learning_rate": 1.8101299900007693e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4489288.0,
      "step": 10313
    },
    {
      "epoch": 0.793323590493039,
      "grad_norm": 12.3125,
      "learning_rate": 1.8100146142604418e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4489600.0,
      "step": 10314
    },
    {
      "epoch": 0.7934005076532574,
      "grad_norm": 5.84375,
      "learning_rate": 1.8098992385201138e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 4490202.0,
      "step": 10315
    },
    {
      "epoch": 0.7934774248134759,
      "grad_norm": 40.0,
      "learning_rate": 1.809783862779786e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4490500.0,
      "step": 10316
    },
    {
      "epoch": 0.7935543419736943,
      "grad_norm": 34.75,
      "learning_rate": 1.8096684870394585e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4490816.0,
      "step": 10317
    },
    {
      "epoch": 0.7936312591339127,
      "grad_norm": 7.6875,
      "learning_rate": 1.809553111299131e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9963503479957581,
      "num_tokens": 4491640.0,
      "step": 10318
    },
    {
      "epoch": 0.7937081762941313,
      "grad_norm": 36.5,
      "learning_rate": 1.8094377355588033e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4491962.0,
      "step": 10319
    },
    {
      "epoch": 0.7937850934543497,
      "grad_norm": 6.78125,
      "learning_rate": 1.8093223598184757e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 4492770.0,
      "step": 10320
    },
    {
      "epoch": 0.7938620106145681,
      "grad_norm": 8.8125,
      "learning_rate": 1.8092069840781478e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4493326.0,
      "step": 10321
    },
    {
      "epoch": 0.7939389277747866,
      "grad_norm": 17.25,
      "learning_rate": 1.8090916083378203e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4493608.0,
      "step": 10322
    },
    {
      "epoch": 0.794015844935005,
      "grad_norm": 6.90625,
      "learning_rate": 1.8089762325974925e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 4494176.0,
      "step": 10323
    },
    {
      "epoch": 0.7940927620952234,
      "grad_norm": 45.75,
      "learning_rate": 1.808860856857165e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4494516.0,
      "step": 10324
    },
    {
      "epoch": 0.7941696792554419,
      "grad_norm": 22.0,
      "learning_rate": 1.8087454811168372e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4494844.0,
      "step": 10325
    },
    {
      "epoch": 0.7942465964156603,
      "grad_norm": 13.5,
      "learning_rate": 1.8086301053765095e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4495394.0,
      "step": 10326
    },
    {
      "epoch": 0.7943235135758788,
      "grad_norm": 44.0,
      "learning_rate": 1.8085147296361818e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4495712.0,
      "step": 10327
    },
    {
      "epoch": 0.7944004307360972,
      "grad_norm": 80.5,
      "learning_rate": 1.8083993538958542e-06,
      "loss": 0.0549,
      "mean_token_accuracy": 0.9764309525489807,
      "num_tokens": 4496012.0,
      "step": 10328
    },
    {
      "epoch": 0.7944773478963156,
      "grad_norm": 53.0,
      "learning_rate": 1.8082839781555265e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 4496598.0,
      "step": 10329
    },
    {
      "epoch": 0.7945542650565341,
      "grad_norm": 59.0,
      "learning_rate": 1.808168602415199e-06,
      "loss": 0.0361,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 4496870.0,
      "step": 10330
    },
    {
      "epoch": 0.7946311822167526,
      "grad_norm": 30.75,
      "learning_rate": 1.808053226674871e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4497216.0,
      "step": 10331
    },
    {
      "epoch": 0.794708099376971,
      "grad_norm": 46.0,
      "learning_rate": 1.8079378509345435e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4497530.0,
      "step": 10332
    },
    {
      "epoch": 0.7947850165371895,
      "grad_norm": 14.0625,
      "learning_rate": 1.8078224751942157e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4497860.0,
      "step": 10333
    },
    {
      "epoch": 0.7948619336974079,
      "grad_norm": 27.5,
      "learning_rate": 1.8077070994538882e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 4498674.0,
      "step": 10334
    },
    {
      "epoch": 0.7949388508576264,
      "grad_norm": 27.5,
      "learning_rate": 1.8075917237135607e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4499020.0,
      "step": 10335
    },
    {
      "epoch": 0.7950157680178448,
      "grad_norm": 11.1875,
      "learning_rate": 1.807476347973233e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4499802.0,
      "step": 10336
    },
    {
      "epoch": 0.7950926851780632,
      "grad_norm": 7.875,
      "learning_rate": 1.807360972232905e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4500584.0,
      "step": 10337
    },
    {
      "epoch": 0.7951696023382817,
      "grad_norm": 11.6875,
      "learning_rate": 1.8072455964925775e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 4501144.0,
      "step": 10338
    },
    {
      "epoch": 0.7952465194985001,
      "grad_norm": 13.25,
      "learning_rate": 1.80713022075225e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 4501688.0,
      "step": 10339
    },
    {
      "epoch": 0.7953234366587185,
      "grad_norm": 66.0,
      "learning_rate": 1.8070148450119222e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4501970.0,
      "step": 10340
    },
    {
      "epoch": 0.795400353818937,
      "grad_norm": 55.25,
      "learning_rate": 1.8068994692715946e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4502314.0,
      "step": 10341
    },
    {
      "epoch": 0.7954772709791554,
      "grad_norm": 5.09375,
      "learning_rate": 1.8067840935312667e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4502900.0,
      "step": 10342
    },
    {
      "epoch": 0.7955541881393738,
      "grad_norm": 24.0,
      "learning_rate": 1.8066687177909392e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.991482138633728,
      "num_tokens": 4503488.0,
      "step": 10343
    },
    {
      "epoch": 0.7956311052995924,
      "grad_norm": 47.5,
      "learning_rate": 1.8065533420506114e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4503768.0,
      "step": 10344
    },
    {
      "epoch": 0.7957080224598108,
      "grad_norm": 30.75,
      "learning_rate": 1.8064379663102839e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4504078.0,
      "step": 10345
    },
    {
      "epoch": 0.7957849396200293,
      "grad_norm": 41.0,
      "learning_rate": 1.8063225905699564e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4504360.0,
      "step": 10346
    },
    {
      "epoch": 0.7958618567802477,
      "grad_norm": 18.125,
      "learning_rate": 1.8062072148296286e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4504646.0,
      "step": 10347
    },
    {
      "epoch": 0.7959387739404661,
      "grad_norm": 5.84375,
      "learning_rate": 1.8060918390893007e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 4505218.0,
      "step": 10348
    },
    {
      "epoch": 0.7960156911006846,
      "grad_norm": 12.9375,
      "learning_rate": 1.8059764633489731e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4505514.0,
      "step": 10349
    },
    {
      "epoch": 0.796092608260903,
      "grad_norm": 11.0,
      "learning_rate": 1.8058610876086456e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4506060.0,
      "step": 10350
    },
    {
      "epoch": 0.7961695254211214,
      "grad_norm": 22.0,
      "learning_rate": 1.8057457118683179e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 4506392.0,
      "step": 10351
    },
    {
      "epoch": 0.7962464425813399,
      "grad_norm": 25.75,
      "learning_rate": 1.8056303361279903e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 4506700.0,
      "step": 10352
    },
    {
      "epoch": 0.7963233597415583,
      "grad_norm": 32.0,
      "learning_rate": 1.8055149603876624e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4506978.0,
      "step": 10353
    },
    {
      "epoch": 0.7964002769017768,
      "grad_norm": 13.1875,
      "learning_rate": 1.8053995846473349e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 4507562.0,
      "step": 10354
    },
    {
      "epoch": 0.7964771940619952,
      "grad_norm": 53.25,
      "learning_rate": 1.8052842089070071e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4507818.0,
      "step": 10355
    },
    {
      "epoch": 0.7965541112222136,
      "grad_norm": 23.75,
      "learning_rate": 1.8051688331666796e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4508360.0,
      "step": 10356
    },
    {
      "epoch": 0.7966310283824322,
      "grad_norm": 10.25,
      "learning_rate": 1.805053457426352e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9904458522796631,
      "num_tokens": 4508676.0,
      "step": 10357
    },
    {
      "epoch": 0.7967079455426506,
      "grad_norm": 10.0625,
      "learning_rate": 1.804938081686024e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 4508980.0,
      "step": 10358
    },
    {
      "epoch": 0.796784862702869,
      "grad_norm": 59.5,
      "learning_rate": 1.8048227059456964e-06,
      "loss": 0.0372,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4509280.0,
      "step": 10359
    },
    {
      "epoch": 0.7968617798630875,
      "grad_norm": 13.1875,
      "learning_rate": 1.8047073302053688e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4509612.0,
      "step": 10360
    },
    {
      "epoch": 0.7969386970233059,
      "grad_norm": 21.125,
      "learning_rate": 1.8045919544650413e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4509928.0,
      "step": 10361
    },
    {
      "epoch": 0.7970156141835243,
      "grad_norm": 9.375,
      "learning_rate": 1.8044765787247136e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 4510264.0,
      "step": 10362
    },
    {
      "epoch": 0.7970925313437428,
      "grad_norm": 47.75,
      "learning_rate": 1.804361202984386e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 4510554.0,
      "step": 10363
    },
    {
      "epoch": 0.7971694485039612,
      "grad_norm": 60.75,
      "learning_rate": 1.804245827244058e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 4510926.0,
      "step": 10364
    },
    {
      "epoch": 0.7972463656641797,
      "grad_norm": 13.25,
      "learning_rate": 1.8041304515037305e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 4511214.0,
      "step": 10365
    },
    {
      "epoch": 0.7973232828243981,
      "grad_norm": 23.0,
      "learning_rate": 1.8040150757634028e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4511516.0,
      "step": 10366
    },
    {
      "epoch": 0.7974001999846165,
      "grad_norm": 23.0,
      "learning_rate": 1.8038997000230753e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 4512088.0,
      "step": 10367
    },
    {
      "epoch": 0.797477117144835,
      "grad_norm": 22.625,
      "learning_rate": 1.8037843242827477e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4512368.0,
      "step": 10368
    },
    {
      "epoch": 0.7975540343050534,
      "grad_norm": 7.03125,
      "learning_rate": 1.8036689485424198e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9948979616165161,
      "num_tokens": 4513154.0,
      "step": 10369
    },
    {
      "epoch": 0.7976309514652719,
      "grad_norm": 5.65625,
      "learning_rate": 1.803553572802092e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4513672.0,
      "step": 10370
    },
    {
      "epoch": 0.7977078686254904,
      "grad_norm": 24.75,
      "learning_rate": 1.8034381970617645e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4514010.0,
      "step": 10371
    },
    {
      "epoch": 0.7977847857857088,
      "grad_norm": 4.5625,
      "learning_rate": 1.803322821321437e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4514310.0,
      "step": 10372
    },
    {
      "epoch": 0.7978617029459273,
      "grad_norm": 14.5,
      "learning_rate": 1.8032074455811092e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4514628.0,
      "step": 10373
    },
    {
      "epoch": 0.7979386201061457,
      "grad_norm": 46.25,
      "learning_rate": 1.8030920698407817e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 4514892.0,
      "step": 10374
    },
    {
      "epoch": 0.7980155372663641,
      "grad_norm": 23.0,
      "learning_rate": 1.8029766941004538e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 4515230.0,
      "step": 10375
    },
    {
      "epoch": 0.7980924544265826,
      "grad_norm": 20.625,
      "learning_rate": 1.8028613183601262e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.997032642364502,
      "num_tokens": 4515568.0,
      "step": 10376
    },
    {
      "epoch": 0.798169371586801,
      "grad_norm": 19.0,
      "learning_rate": 1.8027459426197985e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4516090.0,
      "step": 10377
    },
    {
      "epoch": 0.7982462887470194,
      "grad_norm": 33.5,
      "learning_rate": 1.802630566879471e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 4516350.0,
      "step": 10378
    },
    {
      "epoch": 0.7983232059072379,
      "grad_norm": 12.0,
      "learning_rate": 1.8025151911391432e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4516868.0,
      "step": 10379
    },
    {
      "epoch": 0.7984001230674563,
      "grad_norm": 43.75,
      "learning_rate": 1.8023998153988155e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4517150.0,
      "step": 10380
    },
    {
      "epoch": 0.7984770402276747,
      "grad_norm": 8.375,
      "learning_rate": 1.8022844396584877e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9950860142707825,
      "num_tokens": 4517966.0,
      "step": 10381
    },
    {
      "epoch": 0.7985539573878933,
      "grad_norm": 9.125,
      "learning_rate": 1.8021690639181602e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4518540.0,
      "step": 10382
    },
    {
      "epoch": 0.7986308745481117,
      "grad_norm": 5.78125,
      "learning_rate": 1.8020536881778325e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 4519092.0,
      "step": 10383
    },
    {
      "epoch": 0.7987077917083302,
      "grad_norm": 21.125,
      "learning_rate": 1.801938312437505e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 4519680.0,
      "step": 10384
    },
    {
      "epoch": 0.7987847088685486,
      "grad_norm": 11.0,
      "learning_rate": 1.8018229366971774e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9933221936225891,
      "num_tokens": 4520280.0,
      "step": 10385
    },
    {
      "epoch": 0.798861626028767,
      "grad_norm": 19.375,
      "learning_rate": 1.8017075609568494e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4520632.0,
      "step": 10386
    },
    {
      "epoch": 0.7989385431889855,
      "grad_norm": 24.625,
      "learning_rate": 1.8015921852165217e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9883268475532532,
      "num_tokens": 4520890.0,
      "step": 10387
    },
    {
      "epoch": 0.7990154603492039,
      "grad_norm": 9.0625,
      "learning_rate": 1.8014768094761942e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4521212.0,
      "step": 10388
    },
    {
      "epoch": 0.7990923775094223,
      "grad_norm": 10.1875,
      "learning_rate": 1.8013614337358666e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 4521514.0,
      "step": 10389
    },
    {
      "epoch": 0.7991692946696408,
      "grad_norm": 8.4375,
      "learning_rate": 1.801246057995539e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 4522334.0,
      "step": 10390
    },
    {
      "epoch": 0.7992462118298592,
      "grad_norm": 19.25,
      "learning_rate": 1.801130682255211e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 4522694.0,
      "step": 10391
    },
    {
      "epoch": 0.7993231289900777,
      "grad_norm": 21.625,
      "learning_rate": 1.8010153065148834e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 4523038.0,
      "step": 10392
    },
    {
      "epoch": 0.7994000461502961,
      "grad_norm": 11.4375,
      "learning_rate": 1.8008999307745559e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 4523640.0,
      "step": 10393
    },
    {
      "epoch": 0.7994769633105145,
      "grad_norm": 14.1875,
      "learning_rate": 1.8007845550342281e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4523962.0,
      "step": 10394
    },
    {
      "epoch": 0.7995538804707331,
      "grad_norm": 6.875,
      "learning_rate": 1.8006691792939006e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4524302.0,
      "step": 10395
    },
    {
      "epoch": 0.7996307976309515,
      "grad_norm": 3.03125,
      "learning_rate": 1.8005538035535727e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9962311387062073,
      "num_tokens": 4525100.0,
      "step": 10396
    },
    {
      "epoch": 0.7997077147911699,
      "grad_norm": 7.53125,
      "learning_rate": 1.8004384278132451e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4525684.0,
      "step": 10397
    },
    {
      "epoch": 0.7997846319513884,
      "grad_norm": 4.6875,
      "learning_rate": 1.8003230520729174e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 4526498.0,
      "step": 10398
    },
    {
      "epoch": 0.7998615491116068,
      "grad_norm": 20.0,
      "learning_rate": 1.8002076763325899e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 4526806.0,
      "step": 10399
    },
    {
      "epoch": 0.7999384662718253,
      "grad_norm": 9.8125,
      "learning_rate": 1.8000923005922623e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4527116.0,
      "step": 10400
    },
    {
      "epoch": 0.8000153834320437,
      "grad_norm": 8.6875,
      "learning_rate": 1.7999769248519346e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9970414042472839,
      "num_tokens": 4527456.0,
      "step": 10401
    },
    {
      "epoch": 0.8000923005922621,
      "grad_norm": 19.0,
      "learning_rate": 1.7998615491116066e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4527978.0,
      "step": 10402
    },
    {
      "epoch": 0.8001692177524806,
      "grad_norm": 8.5,
      "learning_rate": 1.799746173371279e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4528548.0,
      "step": 10403
    },
    {
      "epoch": 0.800246134912699,
      "grad_norm": 75.0,
      "learning_rate": 1.7996307976309516e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4528854.0,
      "step": 10404
    },
    {
      "epoch": 0.8003230520729174,
      "grad_norm": 46.25,
      "learning_rate": 1.7995154218906238e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 4529168.0,
      "step": 10405
    },
    {
      "epoch": 0.8003999692331359,
      "grad_norm": 34.75,
      "learning_rate": 1.7994000461502963e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4529490.0,
      "step": 10406
    },
    {
      "epoch": 0.8004768863933543,
      "grad_norm": 57.75,
      "learning_rate": 1.7992846704099683e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 4529846.0,
      "step": 10407
    },
    {
      "epoch": 0.8005538035535728,
      "grad_norm": 12.0,
      "learning_rate": 1.7991692946696408e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4530364.0,
      "step": 10408
    },
    {
      "epoch": 0.8006307207137913,
      "grad_norm": 17.5,
      "learning_rate": 1.799053918929313e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 4530700.0,
      "step": 10409
    },
    {
      "epoch": 0.8007076378740097,
      "grad_norm": 5.75,
      "learning_rate": 1.7989385431889855e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4531244.0,
      "step": 10410
    },
    {
      "epoch": 0.8007845550342282,
      "grad_norm": 6.3125,
      "learning_rate": 1.798823167448658e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 4531816.0,
      "step": 10411
    },
    {
      "epoch": 0.8008614721944466,
      "grad_norm": 6.96875,
      "learning_rate": 1.7987077917083303e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4532100.0,
      "step": 10412
    },
    {
      "epoch": 0.800938389354665,
      "grad_norm": 14.6875,
      "learning_rate": 1.7985924159680023e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4532386.0,
      "step": 10413
    },
    {
      "epoch": 0.8010153065148835,
      "grad_norm": 8.9375,
      "learning_rate": 1.7984770402276748e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4532694.0,
      "step": 10414
    },
    {
      "epoch": 0.8010922236751019,
      "grad_norm": 14.25,
      "learning_rate": 1.7983616644873473e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9965338110923767,
      "num_tokens": 4533272.0,
      "step": 10415
    },
    {
      "epoch": 0.8011691408353203,
      "grad_norm": 19.125,
      "learning_rate": 1.7982462887470195e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 4533766.0,
      "step": 10416
    },
    {
      "epoch": 0.8012460579955388,
      "grad_norm": 40.25,
      "learning_rate": 1.798130913006692e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 4534062.0,
      "step": 10417
    },
    {
      "epoch": 0.8013229751557572,
      "grad_norm": 65.5,
      "learning_rate": 1.798015537266364e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 4534394.0,
      "step": 10418
    },
    {
      "epoch": 0.8013998923159757,
      "grad_norm": 31.375,
      "learning_rate": 1.7979001615260365e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4534704.0,
      "step": 10419
    },
    {
      "epoch": 0.8014768094761942,
      "grad_norm": 17.0,
      "learning_rate": 1.7977847857857088e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4535020.0,
      "step": 10420
    },
    {
      "epoch": 0.8015537266364126,
      "grad_norm": 38.0,
      "learning_rate": 1.7976694100453812e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 4535354.0,
      "step": 10421
    },
    {
      "epoch": 0.8016306437966311,
      "grad_norm": 26.5,
      "learning_rate": 1.7975540343050537e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4535670.0,
      "step": 10422
    },
    {
      "epoch": 0.8017075609568495,
      "grad_norm": 33.75,
      "learning_rate": 1.797438658564726e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 4536006.0,
      "step": 10423
    },
    {
      "epoch": 0.8017844781170679,
      "grad_norm": 60.75,
      "learning_rate": 1.797323282824398e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 4536332.0,
      "step": 10424
    },
    {
      "epoch": 0.8018613952772864,
      "grad_norm": 35.75,
      "learning_rate": 1.7972079070840705e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4536664.0,
      "step": 10425
    },
    {
      "epoch": 0.8019383124375048,
      "grad_norm": 20.0,
      "learning_rate": 1.797092531343743e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 4537204.0,
      "step": 10426
    },
    {
      "epoch": 0.8020152295977232,
      "grad_norm": 9.6875,
      "learning_rate": 1.7969771556034152e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4537486.0,
      "step": 10427
    },
    {
      "epoch": 0.8020921467579417,
      "grad_norm": 9.8125,
      "learning_rate": 1.7968617798630877e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4538032.0,
      "step": 10428
    },
    {
      "epoch": 0.8021690639181601,
      "grad_norm": 56.25,
      "learning_rate": 1.7967464041227597e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4538292.0,
      "step": 10429
    },
    {
      "epoch": 0.8022459810783786,
      "grad_norm": 9.5625,
      "learning_rate": 1.7966310283824322e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4538600.0,
      "step": 10430
    },
    {
      "epoch": 0.802322898238597,
      "grad_norm": 20.0,
      "learning_rate": 1.7965156526421044e-06,
      "loss": 0.0072,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 4539426.0,
      "step": 10431
    },
    {
      "epoch": 0.8023998153988154,
      "grad_norm": 34.25,
      "learning_rate": 1.796400276901777e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 4539760.0,
      "step": 10432
    },
    {
      "epoch": 0.802476732559034,
      "grad_norm": 32.0,
      "learning_rate": 1.7962849011614494e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4540036.0,
      "step": 10433
    },
    {
      "epoch": 0.8025536497192524,
      "grad_norm": 24.75,
      "learning_rate": 1.7961695254211214e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4540608.0,
      "step": 10434
    },
    {
      "epoch": 0.8026305668794708,
      "grad_norm": 29.25,
      "learning_rate": 1.7960541496807937e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4540930.0,
      "step": 10435
    },
    {
      "epoch": 0.8027074840396893,
      "grad_norm": 15.6875,
      "learning_rate": 1.7959387739404662e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4541446.0,
      "step": 10436
    },
    {
      "epoch": 0.8027844011999077,
      "grad_norm": 33.25,
      "learning_rate": 1.7958233982001386e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 4542018.0,
      "step": 10437
    },
    {
      "epoch": 0.8028613183601262,
      "grad_norm": 35.0,
      "learning_rate": 1.7957080224598109e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4542330.0,
      "step": 10438
    },
    {
      "epoch": 0.8029382355203446,
      "grad_norm": 50.25,
      "learning_rate": 1.7955926467194834e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4542614.0,
      "step": 10439
    },
    {
      "epoch": 0.803015152680563,
      "grad_norm": 12.8125,
      "learning_rate": 1.7954772709791554e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4543170.0,
      "step": 10440
    },
    {
      "epoch": 0.8030920698407815,
      "grad_norm": 10.8125,
      "learning_rate": 1.7953618952388279e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 4543456.0,
      "step": 10441
    },
    {
      "epoch": 0.8031689870009999,
      "grad_norm": 16.625,
      "learning_rate": 1.7952465194985001e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4543756.0,
      "step": 10442
    },
    {
      "epoch": 0.8032459041612183,
      "grad_norm": 24.0,
      "learning_rate": 1.7951311437581726e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9811912178993225,
      "num_tokens": 4544076.0,
      "step": 10443
    },
    {
      "epoch": 0.8033228213214368,
      "grad_norm": 13.8125,
      "learning_rate": 1.7950157680178449e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4544628.0,
      "step": 10444
    },
    {
      "epoch": 0.8033997384816552,
      "grad_norm": 10.0625,
      "learning_rate": 1.7949003922775171e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9967948794364929,
      "num_tokens": 4544942.0,
      "step": 10445
    },
    {
      "epoch": 0.8034766556418736,
      "grad_norm": 19.375,
      "learning_rate": 1.7947850165371894e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4545488.0,
      "step": 10446
    },
    {
      "epoch": 0.8035535728020922,
      "grad_norm": 34.25,
      "learning_rate": 1.7946696407968618e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4545806.0,
      "step": 10447
    },
    {
      "epoch": 0.8036304899623106,
      "grad_norm": 17.125,
      "learning_rate": 1.794554265056534e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4546380.0,
      "step": 10448
    },
    {
      "epoch": 0.8037074071225291,
      "grad_norm": 12.125,
      "learning_rate": 1.7944388893162066e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4546926.0,
      "step": 10449
    },
    {
      "epoch": 0.8037843242827475,
      "grad_norm": 37.25,
      "learning_rate": 1.794323513575879e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4547208.0,
      "step": 10450
    },
    {
      "epoch": 0.8038612414429659,
      "grad_norm": 7.0625,
      "learning_rate": 1.794208137835551e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9964349269866943,
      "num_tokens": 4547770.0,
      "step": 10451
    },
    {
      "epoch": 0.8039381586031844,
      "grad_norm": 41.5,
      "learning_rate": 1.7940927620952233e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4548070.0,
      "step": 10452
    },
    {
      "epoch": 0.8040150757634028,
      "grad_norm": 51.0,
      "learning_rate": 1.7939773863548958e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4548410.0,
      "step": 10453
    },
    {
      "epoch": 0.8040919929236212,
      "grad_norm": 5.4375,
      "learning_rate": 1.7938620106145683e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4548958.0,
      "step": 10454
    },
    {
      "epoch": 0.8041689100838397,
      "grad_norm": 7.5625,
      "learning_rate": 1.7937466348742405e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4549502.0,
      "step": 10455
    },
    {
      "epoch": 0.8042458272440581,
      "grad_norm": 9.4375,
      "learning_rate": 1.7936312591339126e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 4550306.0,
      "step": 10456
    },
    {
      "epoch": 0.8043227444042766,
      "grad_norm": 9.4375,
      "learning_rate": 1.793515883393585e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 4550658.0,
      "step": 10457
    },
    {
      "epoch": 0.804399661564495,
      "grad_norm": 34.75,
      "learning_rate": 1.7934005076532575e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4550962.0,
      "step": 10458
    },
    {
      "epoch": 0.8044765787247135,
      "grad_norm": 17.125,
      "learning_rate": 1.7932851319129298e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 4551274.0,
      "step": 10459
    },
    {
      "epoch": 0.804553495884932,
      "grad_norm": 44.5,
      "learning_rate": 1.7931697561726023e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4551578.0,
      "step": 10460
    },
    {
      "epoch": 0.8046304130451504,
      "grad_norm": 18.125,
      "learning_rate": 1.7930543804322743e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4552100.0,
      "step": 10461
    },
    {
      "epoch": 0.8047073302053688,
      "grad_norm": 25.0,
      "learning_rate": 1.7929390046919468e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 4552926.0,
      "step": 10462
    },
    {
      "epoch": 0.8047842473655873,
      "grad_norm": 12.5625,
      "learning_rate": 1.792823628951619e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4553448.0,
      "step": 10463
    },
    {
      "epoch": 0.8048611645258057,
      "grad_norm": 5.8125,
      "learning_rate": 1.7927082532112915e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4553966.0,
      "step": 10464
    },
    {
      "epoch": 0.8049380816860241,
      "grad_norm": 33.75,
      "learning_rate": 1.792592877470964e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4554248.0,
      "step": 10465
    },
    {
      "epoch": 0.8050149988462426,
      "grad_norm": 9.375,
      "learning_rate": 1.7924775017306362e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 4554578.0,
      "step": 10466
    },
    {
      "epoch": 0.805091916006461,
      "grad_norm": 18.0,
      "learning_rate": 1.7923621259903083e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4554872.0,
      "step": 10467
    },
    {
      "epoch": 0.8051688331666795,
      "grad_norm": 5.4375,
      "learning_rate": 1.7922467502499807e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4555414.0,
      "step": 10468
    },
    {
      "epoch": 0.8052457503268979,
      "grad_norm": 3.421875,
      "learning_rate": 1.7921313745096532e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4556196.0,
      "step": 10469
    },
    {
      "epoch": 0.8053226674871163,
      "grad_norm": 15.25,
      "learning_rate": 1.7920159987693255e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4556490.0,
      "step": 10470
    },
    {
      "epoch": 0.8053995846473349,
      "grad_norm": 9.9375,
      "learning_rate": 1.791900623028998e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4557006.0,
      "step": 10471
    },
    {
      "epoch": 0.8054765018075533,
      "grad_norm": 86.5,
      "learning_rate": 1.79178524728867e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 4557316.0,
      "step": 10472
    },
    {
      "epoch": 0.8055534189677717,
      "grad_norm": 23.125,
      "learning_rate": 1.7916698715483425e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4557672.0,
      "step": 10473
    },
    {
      "epoch": 0.8056303361279902,
      "grad_norm": 58.0,
      "learning_rate": 1.7915544958080147e-06,
      "loss": 0.0347,
      "mean_token_accuracy": 0.9774919748306274,
      "num_tokens": 4557986.0,
      "step": 10474
    },
    {
      "epoch": 0.8057072532882086,
      "grad_norm": 21.625,
      "learning_rate": 1.7914391200676872e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4558504.0,
      "step": 10475
    },
    {
      "epoch": 0.8057841704484271,
      "grad_norm": 23.0,
      "learning_rate": 1.7913237443273597e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4559052.0,
      "step": 10476
    },
    {
      "epoch": 0.8058610876086455,
      "grad_norm": 25.0,
      "learning_rate": 1.791208368587032e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 4559402.0,
      "step": 10477
    },
    {
      "epoch": 0.8059380047688639,
      "grad_norm": 24.875,
      "learning_rate": 1.791092992846704e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4559976.0,
      "step": 10478
    },
    {
      "epoch": 0.8060149219290824,
      "grad_norm": 31.75,
      "learning_rate": 1.7909776171063764e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 4560326.0,
      "step": 10479
    },
    {
      "epoch": 0.8060918390893008,
      "grad_norm": 13.9375,
      "learning_rate": 1.790862241366049e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4560846.0,
      "step": 10480
    },
    {
      "epoch": 0.8061687562495192,
      "grad_norm": 26.25,
      "learning_rate": 1.7907468656257212e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4561164.0,
      "step": 10481
    },
    {
      "epoch": 0.8062456734097377,
      "grad_norm": 26.25,
      "learning_rate": 1.7906314898853936e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4561484.0,
      "step": 10482
    },
    {
      "epoch": 0.8063225905699561,
      "grad_norm": 4.875,
      "learning_rate": 1.7905161141450657e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 4562240.0,
      "step": 10483
    },
    {
      "epoch": 0.8063995077301745,
      "grad_norm": 14.5625,
      "learning_rate": 1.7904007384047381e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4562560.0,
      "step": 10484
    },
    {
      "epoch": 0.8064764248903931,
      "grad_norm": 24.125,
      "learning_rate": 1.7902853626644104e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4562844.0,
      "step": 10485
    },
    {
      "epoch": 0.8065533420506115,
      "grad_norm": 21.375,
      "learning_rate": 1.7901699869240829e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 4563136.0,
      "step": 10486
    },
    {
      "epoch": 0.80663025921083,
      "grad_norm": 22.5,
      "learning_rate": 1.7900546111837553e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 4563462.0,
      "step": 10487
    },
    {
      "epoch": 0.8067071763710484,
      "grad_norm": 3.1875,
      "learning_rate": 1.7899392354434276e-06,
      "loss": 0.0057,
      "mean_token_accuracy": 0.9970559477806091,
      "num_tokens": 4564484.0,
      "step": 10488
    },
    {
      "epoch": 0.8067840935312668,
      "grad_norm": 48.5,
      "learning_rate": 1.7898238597030996e-06,
      "loss": 0.0442,
      "mean_token_accuracy": 0.981249988079071,
      "num_tokens": 4564808.0,
      "step": 10489
    },
    {
      "epoch": 0.8068610106914853,
      "grad_norm": 28.875,
      "learning_rate": 1.7897084839627721e-06,
      "loss": 0.0299,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4565148.0,
      "step": 10490
    },
    {
      "epoch": 0.8069379278517037,
      "grad_norm": 13.5625,
      "learning_rate": 1.7895931082224446e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4565442.0,
      "step": 10491
    },
    {
      "epoch": 0.8070148450119221,
      "grad_norm": 26.0,
      "learning_rate": 1.7894777324821168e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4565724.0,
      "step": 10492
    },
    {
      "epoch": 0.8070917621721406,
      "grad_norm": 8.875,
      "learning_rate": 1.7893623567417893e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4566056.0,
      "step": 10493
    },
    {
      "epoch": 0.807168679332359,
      "grad_norm": 24.25,
      "learning_rate": 1.7892469810014614e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4566362.0,
      "step": 10494
    },
    {
      "epoch": 0.8072455964925775,
      "grad_norm": 47.25,
      "learning_rate": 1.7891316052611338e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4566942.0,
      "step": 10495
    },
    {
      "epoch": 0.807322513652796,
      "grad_norm": 9.125,
      "learning_rate": 1.789016229520806e-06,
      "loss": 0.0067,
      "mean_token_accuracy": 0.9962157011032104,
      "num_tokens": 4568002.0,
      "step": 10496
    },
    {
      "epoch": 0.8073994308130144,
      "grad_norm": 7.96875,
      "learning_rate": 1.7889008537804786e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4568548.0,
      "step": 10497
    },
    {
      "epoch": 0.8074763479732329,
      "grad_norm": 31.75,
      "learning_rate": 1.7887854780401508e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 4568846.0,
      "step": 10498
    },
    {
      "epoch": 0.8075532651334513,
      "grad_norm": 24.25,
      "learning_rate": 1.788670102299823e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 4569174.0,
      "step": 10499
    },
    {
      "epoch": 0.8076301822936697,
      "grad_norm": 31.0,
      "learning_rate": 1.7885547265594953e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4569484.0,
      "step": 10500
    },
    {
      "epoch": 0.8077070994538882,
      "grad_norm": 23.125,
      "learning_rate": 1.7884393508191678e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 4569786.0,
      "step": 10501
    },
    {
      "epoch": 0.8077840166141066,
      "grad_norm": 11.8125,
      "learning_rate": 1.78832397507884e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4570304.0,
      "step": 10502
    },
    {
      "epoch": 0.8078609337743251,
      "grad_norm": 33.75,
      "learning_rate": 1.7882085993385125e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4570618.0,
      "step": 10503
    },
    {
      "epoch": 0.8079378509345435,
      "grad_norm": 17.125,
      "learning_rate": 1.788093223598185e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 4570978.0,
      "step": 10504
    },
    {
      "epoch": 0.8080147680947619,
      "grad_norm": 12.3125,
      "learning_rate": 1.787977847857857e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4571296.0,
      "step": 10505
    },
    {
      "epoch": 0.8080916852549804,
      "grad_norm": 10.125,
      "learning_rate": 1.7878624721175293e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9930192232131958,
      "num_tokens": 4571870.0,
      "step": 10506
    },
    {
      "epoch": 0.8081686024151988,
      "grad_norm": 35.25,
      "learning_rate": 1.7877470963772018e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9922178983688354,
      "num_tokens": 4572128.0,
      "step": 10507
    },
    {
      "epoch": 0.8082455195754172,
      "grad_norm": 14.5,
      "learning_rate": 1.7876317206368742e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4572690.0,
      "step": 10508
    },
    {
      "epoch": 0.8083224367356358,
      "grad_norm": 25.0,
      "learning_rate": 1.7875163448965465e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4572972.0,
      "step": 10509
    },
    {
      "epoch": 0.8083993538958542,
      "grad_norm": 8.0625,
      "learning_rate": 1.7874009691562185e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9918533563613892,
      "num_tokens": 4573464.0,
      "step": 10510
    },
    {
      "epoch": 0.8084762710560726,
      "grad_norm": 19.125,
      "learning_rate": 1.787285593415891e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 4573820.0,
      "step": 10511
    },
    {
      "epoch": 0.8085531882162911,
      "grad_norm": 9.5,
      "learning_rate": 1.7871702176755635e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4574142.0,
      "step": 10512
    },
    {
      "epoch": 0.8086301053765095,
      "grad_norm": 7.6875,
      "learning_rate": 1.7870548419352357e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4574716.0,
      "step": 10513
    },
    {
      "epoch": 0.808707022536728,
      "grad_norm": 9.8125,
      "learning_rate": 1.7869394661949082e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 4575496.0,
      "step": 10514
    },
    {
      "epoch": 0.8087839396969464,
      "grad_norm": 23.0,
      "learning_rate": 1.7868240904545807e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4575802.0,
      "step": 10515
    },
    {
      "epoch": 0.8088608568571648,
      "grad_norm": 7.375,
      "learning_rate": 1.7867087147142527e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4576134.0,
      "step": 10516
    },
    {
      "epoch": 0.8089377740173833,
      "grad_norm": 10.6875,
      "learning_rate": 1.786593338973925e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4576442.0,
      "step": 10517
    },
    {
      "epoch": 0.8090146911776017,
      "grad_norm": 15.4375,
      "learning_rate": 1.7864779632335975e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4576728.0,
      "step": 10518
    },
    {
      "epoch": 0.8090916083378201,
      "grad_norm": 8.375,
      "learning_rate": 1.78636258749327e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4577274.0,
      "step": 10519
    },
    {
      "epoch": 0.8091685254980386,
      "grad_norm": 39.5,
      "learning_rate": 1.7862472117529422e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4577836.0,
      "step": 10520
    },
    {
      "epoch": 0.809245442658257,
      "grad_norm": 36.75,
      "learning_rate": 1.7861318360126142e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 4578624.0,
      "step": 10521
    },
    {
      "epoch": 0.8093223598184756,
      "grad_norm": 10.6875,
      "learning_rate": 1.7860164602722867e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 4578956.0,
      "step": 10522
    },
    {
      "epoch": 0.809399276978694,
      "grad_norm": 47.75,
      "learning_rate": 1.7859010845319592e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 4579266.0,
      "step": 10523
    },
    {
      "epoch": 0.8094761941389124,
      "grad_norm": 4.25,
      "learning_rate": 1.7857857087916314e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4580074.0,
      "step": 10524
    },
    {
      "epoch": 0.8095531112991309,
      "grad_norm": 26.5,
      "learning_rate": 1.785670333051304e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 4580384.0,
      "step": 10525
    },
    {
      "epoch": 0.8096300284593493,
      "grad_norm": 4.59375,
      "learning_rate": 1.7855549573109764e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 4580944.0,
      "step": 10526
    },
    {
      "epoch": 0.8097069456195677,
      "grad_norm": 13.3125,
      "learning_rate": 1.7854395815706484e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4581466.0,
      "step": 10527
    },
    {
      "epoch": 0.8097838627797862,
      "grad_norm": 11.75,
      "learning_rate": 1.7853242058303207e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 4582036.0,
      "step": 10528
    },
    {
      "epoch": 0.8098607799400046,
      "grad_norm": 16.25,
      "learning_rate": 1.7852088300899931e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4582580.0,
      "step": 10529
    },
    {
      "epoch": 0.809937697100223,
      "grad_norm": 13.0,
      "learning_rate": 1.7850934543496656e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 4582898.0,
      "step": 10530
    },
    {
      "epoch": 0.8100146142604415,
      "grad_norm": 9.125,
      "learning_rate": 1.7849780786093379e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4583198.0,
      "step": 10531
    },
    {
      "epoch": 0.8100915314206599,
      "grad_norm": 6.6875,
      "learning_rate": 1.78486270286901e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.996370255947113,
      "num_tokens": 4583750.0,
      "step": 10532
    },
    {
      "epoch": 0.8101684485808784,
      "grad_norm": 37.0,
      "learning_rate": 1.7847473271286824e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4584056.0,
      "step": 10533
    },
    {
      "epoch": 0.8102453657410968,
      "grad_norm": 17.125,
      "learning_rate": 1.7846319513883549e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.996610164642334,
      "num_tokens": 4584352.0,
      "step": 10534
    },
    {
      "epoch": 0.8103222829013152,
      "grad_norm": 22.0,
      "learning_rate": 1.7845165756480271e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9982905983924866,
      "num_tokens": 4584938.0,
      "step": 10535
    },
    {
      "epoch": 0.8103992000615338,
      "grad_norm": 13.125,
      "learning_rate": 1.7844011999076996e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4585454.0,
      "step": 10536
    },
    {
      "epoch": 0.8104761172217522,
      "grad_norm": 14.8125,
      "learning_rate": 1.7842858241673716e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 4585798.0,
      "step": 10537
    },
    {
      "epoch": 0.8105530343819706,
      "grad_norm": 180.0,
      "learning_rate": 1.7841704484270441e-06,
      "loss": 0.0435,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 4586148.0,
      "step": 10538
    },
    {
      "epoch": 0.8106299515421891,
      "grad_norm": 26.875,
      "learning_rate": 1.7840550726867164e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4586456.0,
      "step": 10539
    },
    {
      "epoch": 0.8107068687024075,
      "grad_norm": 33.5,
      "learning_rate": 1.7839396969463888e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 4586786.0,
      "step": 10540
    },
    {
      "epoch": 0.810783785862626,
      "grad_norm": 23.75,
      "learning_rate": 1.7838243212060613e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4587064.0,
      "step": 10541
    },
    {
      "epoch": 0.8108607030228444,
      "grad_norm": 86.5,
      "learning_rate": 1.7837089454657336e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9887955188751221,
      "num_tokens": 4587422.0,
      "step": 10542
    },
    {
      "epoch": 0.8109376201830628,
      "grad_norm": 38.75,
      "learning_rate": 1.7835935697254056e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4587982.0,
      "step": 10543
    },
    {
      "epoch": 0.8110145373432813,
      "grad_norm": 88.0,
      "learning_rate": 1.783478193985078e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4588282.0,
      "step": 10544
    },
    {
      "epoch": 0.8110914545034997,
      "grad_norm": 35.5,
      "learning_rate": 1.7833628182447505e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 4588818.0,
      "step": 10545
    },
    {
      "epoch": 0.8111683716637181,
      "grad_norm": 40.0,
      "learning_rate": 1.7832474425044228e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4589120.0,
      "step": 10546
    },
    {
      "epoch": 0.8112452888239366,
      "grad_norm": 22.75,
      "learning_rate": 1.7831320667640953e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4589700.0,
      "step": 10547
    },
    {
      "epoch": 0.811322205984155,
      "grad_norm": 25.75,
      "learning_rate": 1.7830166910237673e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 4590034.0,
      "step": 10548
    },
    {
      "epoch": 0.8113991231443735,
      "grad_norm": 13.0,
      "learning_rate": 1.7829013152834398e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 4590356.0,
      "step": 10549
    },
    {
      "epoch": 0.811476040304592,
      "grad_norm": 42.75,
      "learning_rate": 1.782785939543112e-06,
      "loss": 0.0369,
      "mean_token_accuracy": 0.9832776188850403,
      "num_tokens": 4590658.0,
      "step": 10550
    },
    {
      "epoch": 0.8115529574648104,
      "grad_norm": 9.0625,
      "learning_rate": 1.7826705638027845e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4591204.0,
      "step": 10551
    },
    {
      "epoch": 0.8116298746250289,
      "grad_norm": 19.0,
      "learning_rate": 1.7825551880624568e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4591522.0,
      "step": 10552
    },
    {
      "epoch": 0.8117067917852473,
      "grad_norm": 6.78125,
      "learning_rate": 1.7824398123221292e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 4592302.0,
      "step": 10553
    },
    {
      "epoch": 0.8117837089454657,
      "grad_norm": 33.0,
      "learning_rate": 1.7823244365818013e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4592636.0,
      "step": 10554
    },
    {
      "epoch": 0.8118606261056842,
      "grad_norm": 23.25,
      "learning_rate": 1.7822090608414738e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 4592916.0,
      "step": 10555
    },
    {
      "epoch": 0.8119375432659026,
      "grad_norm": 5.875,
      "learning_rate": 1.782093685101146e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4593262.0,
      "step": 10556
    },
    {
      "epoch": 0.812014460426121,
      "grad_norm": 21.25,
      "learning_rate": 1.7819783093608185e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 4593592.0,
      "step": 10557
    },
    {
      "epoch": 0.8120913775863395,
      "grad_norm": 18.0,
      "learning_rate": 1.781862933620491e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 4593846.0,
      "step": 10558
    },
    {
      "epoch": 0.8121682947465579,
      "grad_norm": 20.25,
      "learning_rate": 1.781747557880163e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4594420.0,
      "step": 10559
    },
    {
      "epoch": 0.8122452119067765,
      "grad_norm": 54.5,
      "learning_rate": 1.7816321821398353e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4595000.0,
      "step": 10560
    },
    {
      "epoch": 0.8123221290669949,
      "grad_norm": 19.25,
      "learning_rate": 1.7815168063995077e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 4595496.0,
      "step": 10561
    },
    {
      "epoch": 0.8123990462272133,
      "grad_norm": 52.25,
      "learning_rate": 1.7814014306591802e-06,
      "loss": 0.0286,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 4596084.0,
      "step": 10562
    },
    {
      "epoch": 0.8124759633874318,
      "grad_norm": 39.75,
      "learning_rate": 1.7812860549188525e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4596642.0,
      "step": 10563
    },
    {
      "epoch": 0.8125528805476502,
      "grad_norm": 32.75,
      "learning_rate": 1.7811706791785245e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4596966.0,
      "step": 10564
    },
    {
      "epoch": 0.8126297977078686,
      "grad_norm": 18.375,
      "learning_rate": 1.781055303438197e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4597518.0,
      "step": 10565
    },
    {
      "epoch": 0.8127067148680871,
      "grad_norm": 40.5,
      "learning_rate": 1.7809399276978695e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4597852.0,
      "step": 10566
    },
    {
      "epoch": 0.8127836320283055,
      "grad_norm": 32.75,
      "learning_rate": 1.7808245519575417e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4598424.0,
      "step": 10567
    },
    {
      "epoch": 0.8128605491885239,
      "grad_norm": 15.8125,
      "learning_rate": 1.7807091762172142e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 4599008.0,
      "step": 10568
    },
    {
      "epoch": 0.8129374663487424,
      "grad_norm": 4.03125,
      "learning_rate": 1.7805938004768866e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 4599804.0,
      "step": 10569
    },
    {
      "epoch": 0.8130143835089608,
      "grad_norm": 5.78125,
      "learning_rate": 1.7804784247365587e-06,
      "loss": 0.0067,
      "mean_token_accuracy": 0.9975247383117676,
      "num_tokens": 4600614.0,
      "step": 10570
    },
    {
      "epoch": 0.8130913006691793,
      "grad_norm": 48.75,
      "learning_rate": 1.780363048996231e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4600922.0,
      "step": 10571
    },
    {
      "epoch": 0.8131682178293977,
      "grad_norm": 30.875,
      "learning_rate": 1.7802476732559034e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4601228.0,
      "step": 10572
    },
    {
      "epoch": 0.8132451349896161,
      "grad_norm": 26.625,
      "learning_rate": 1.7801322975155759e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4601792.0,
      "step": 10573
    },
    {
      "epoch": 0.8133220521498347,
      "grad_norm": 12.25,
      "learning_rate": 1.7800169217752482e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 4602136.0,
      "step": 10574
    },
    {
      "epoch": 0.8133989693100531,
      "grad_norm": 5.625,
      "learning_rate": 1.7799015460349202e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4602710.0,
      "step": 10575
    },
    {
      "epoch": 0.8134758864702715,
      "grad_norm": 9.0625,
      "learning_rate": 1.7797861702945927e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4603256.0,
      "step": 10576
    },
    {
      "epoch": 0.81355280363049,
      "grad_norm": 6.4375,
      "learning_rate": 1.7796707945542651e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9987179636955261,
      "num_tokens": 4604038.0,
      "step": 10577
    },
    {
      "epoch": 0.8136297207907084,
      "grad_norm": 18.875,
      "learning_rate": 1.7795554188139374e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4604294.0,
      "step": 10578
    },
    {
      "epoch": 0.8137066379509269,
      "grad_norm": 14.125,
      "learning_rate": 1.7794400430736099e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4604812.0,
      "step": 10579
    },
    {
      "epoch": 0.8137835551111453,
      "grad_norm": 32.25,
      "learning_rate": 1.7793246673332823e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9818840622901917,
      "num_tokens": 4605090.0,
      "step": 10580
    },
    {
      "epoch": 0.8138604722713637,
      "grad_norm": 17.375,
      "learning_rate": 1.7792092915929544e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 4605876.0,
      "step": 10581
    },
    {
      "epoch": 0.8139373894315822,
      "grad_norm": 59.0,
      "learning_rate": 1.7790939158526266e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4606172.0,
      "step": 10582
    },
    {
      "epoch": 0.8140143065918006,
      "grad_norm": 64.5,
      "learning_rate": 1.7789785401122991e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4606472.0,
      "step": 10583
    },
    {
      "epoch": 0.814091223752019,
      "grad_norm": 108.5,
      "learning_rate": 1.7788631643719716e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4606802.0,
      "step": 10584
    },
    {
      "epoch": 0.8141681409122375,
      "grad_norm": 13.125,
      "learning_rate": 1.7787477886316438e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 4607374.0,
      "step": 10585
    },
    {
      "epoch": 0.814245058072456,
      "grad_norm": 33.25,
      "learning_rate": 1.7786324128913159e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4607958.0,
      "step": 10586
    },
    {
      "epoch": 0.8143219752326744,
      "grad_norm": 23.5,
      "learning_rate": 1.7785170371509884e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9982174634933472,
      "num_tokens": 4608520.0,
      "step": 10587
    },
    {
      "epoch": 0.8143988923928929,
      "grad_norm": 20.75,
      "learning_rate": 1.7784016614106608e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4609104.0,
      "step": 10588
    },
    {
      "epoch": 0.8144758095531113,
      "grad_norm": 5.53125,
      "learning_rate": 1.778286285670333e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9942857027053833,
      "num_tokens": 4609630.0,
      "step": 10589
    },
    {
      "epoch": 0.8145527267133298,
      "grad_norm": 26.75,
      "learning_rate": 1.7781709099300056e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4609974.0,
      "step": 10590
    },
    {
      "epoch": 0.8146296438735482,
      "grad_norm": 8.5,
      "learning_rate": 1.778055534189678e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4610516.0,
      "step": 10591
    },
    {
      "epoch": 0.8147065610337666,
      "grad_norm": 9.5625,
      "learning_rate": 1.77794015844935e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4611038.0,
      "step": 10592
    },
    {
      "epoch": 0.8147834781939851,
      "grad_norm": 16.25,
      "learning_rate": 1.7778247827090223e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4611334.0,
      "step": 10593
    },
    {
      "epoch": 0.8148603953542035,
      "grad_norm": 4.0,
      "learning_rate": 1.7777094069686948e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4611918.0,
      "step": 10594
    },
    {
      "epoch": 0.8149373125144219,
      "grad_norm": 28.625,
      "learning_rate": 1.7775940312283673e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 4612206.0,
      "step": 10595
    },
    {
      "epoch": 0.8150142296746404,
      "grad_norm": 18.25,
      "learning_rate": 1.7774786554880395e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4612514.0,
      "step": 10596
    },
    {
      "epoch": 0.8150911468348588,
      "grad_norm": 19.125,
      "learning_rate": 1.7773632797477116e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9855907559394836,
      "num_tokens": 4612862.0,
      "step": 10597
    },
    {
      "epoch": 0.8151680639950774,
      "grad_norm": 10.9375,
      "learning_rate": 1.777247904007384e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4613430.0,
      "step": 10598
    },
    {
      "epoch": 0.8152449811552958,
      "grad_norm": 10.5,
      "learning_rate": 1.7771325282670565e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4613986.0,
      "step": 10599
    },
    {
      "epoch": 0.8153218983155142,
      "grad_norm": 12.8125,
      "learning_rate": 1.7770171525267288e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4614570.0,
      "step": 10600
    },
    {
      "epoch": 0.8153988154757327,
      "grad_norm": 16.5,
      "learning_rate": 1.7769017767864012e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4615132.0,
      "step": 10601
    },
    {
      "epoch": 0.8154757326359511,
      "grad_norm": 33.75,
      "learning_rate": 1.7767864010460733e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9858155846595764,
      "num_tokens": 4615416.0,
      "step": 10602
    },
    {
      "epoch": 0.8155526497961695,
      "grad_norm": 29.125,
      "learning_rate": 1.7766710253057458e-06,
      "loss": 0.0291,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4615726.0,
      "step": 10603
    },
    {
      "epoch": 0.815629566956388,
      "grad_norm": 29.625,
      "learning_rate": 1.776555649565418e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4616010.0,
      "step": 10604
    },
    {
      "epoch": 0.8157064841166064,
      "grad_norm": 43.75,
      "learning_rate": 1.7764402738250905e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4616350.0,
      "step": 10605
    },
    {
      "epoch": 0.8157834012768248,
      "grad_norm": 21.375,
      "learning_rate": 1.7763248980847627e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 4616914.0,
      "step": 10606
    },
    {
      "epoch": 0.8158603184370433,
      "grad_norm": 34.5,
      "learning_rate": 1.7762095223444352e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4617260.0,
      "step": 10607
    },
    {
      "epoch": 0.8159372355972617,
      "grad_norm": 9.5,
      "learning_rate": 1.7760941466041073e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9931856989860535,
      "num_tokens": 4617848.0,
      "step": 10608
    },
    {
      "epoch": 0.8160141527574802,
      "grad_norm": 11.75,
      "learning_rate": 1.7759787708637797e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4618392.0,
      "step": 10609
    },
    {
      "epoch": 0.8160910699176986,
      "grad_norm": 4.0625,
      "learning_rate": 1.775863395123452e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9973474740982056,
      "num_tokens": 4619148.0,
      "step": 10610
    },
    {
      "epoch": 0.816167987077917,
      "grad_norm": 26.375,
      "learning_rate": 1.7757480193831245e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4619446.0,
      "step": 10611
    },
    {
      "epoch": 0.8162449042381356,
      "grad_norm": 12.875,
      "learning_rate": 1.775632643642797e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4619964.0,
      "step": 10612
    },
    {
      "epoch": 0.816321821398354,
      "grad_norm": 20.75,
      "learning_rate": 1.775517267902469e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4620258.0,
      "step": 10613
    },
    {
      "epoch": 0.8163987385585724,
      "grad_norm": 8.25,
      "learning_rate": 1.7754018921621412e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9987684488296509,
      "num_tokens": 4621072.0,
      "step": 10614
    },
    {
      "epoch": 0.8164756557187909,
      "grad_norm": 17.75,
      "learning_rate": 1.7752865164218137e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4621418.0,
      "step": 10615
    },
    {
      "epoch": 0.8165525728790093,
      "grad_norm": 32.5,
      "learning_rate": 1.7751711406814862e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4621722.0,
      "step": 10616
    },
    {
      "epoch": 0.8166294900392278,
      "grad_norm": 11.5625,
      "learning_rate": 1.7750557649411584e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4622262.0,
      "step": 10617
    },
    {
      "epoch": 0.8167064071994462,
      "grad_norm": 11.0,
      "learning_rate": 1.774940389200831e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4622556.0,
      "step": 10618
    },
    {
      "epoch": 0.8167833243596646,
      "grad_norm": 73.5,
      "learning_rate": 1.774825013460503e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4622878.0,
      "step": 10619
    },
    {
      "epoch": 0.8168602415198831,
      "grad_norm": 39.0,
      "learning_rate": 1.7747096377201754e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9931623935699463,
      "num_tokens": 4623464.0,
      "step": 10620
    },
    {
      "epoch": 0.8169371586801015,
      "grad_norm": 28.25,
      "learning_rate": 1.7745942619798477e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9971181750297546,
      "num_tokens": 4623812.0,
      "step": 10621
    },
    {
      "epoch": 0.8170140758403199,
      "grad_norm": 26.625,
      "learning_rate": 1.7744788862395201e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 4624370.0,
      "step": 10622
    },
    {
      "epoch": 0.8170909930005384,
      "grad_norm": 41.5,
      "learning_rate": 1.7743635104991926e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4624698.0,
      "step": 10623
    },
    {
      "epoch": 0.8171679101607568,
      "grad_norm": 53.5,
      "learning_rate": 1.7742481347588647e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 4625310.0,
      "step": 10624
    },
    {
      "epoch": 0.8172448273209754,
      "grad_norm": 15.0,
      "learning_rate": 1.774132759018537e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 4625640.0,
      "step": 10625
    },
    {
      "epoch": 0.8173217444811938,
      "grad_norm": 8.9375,
      "learning_rate": 1.7740173832782094e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4625924.0,
      "step": 10626
    },
    {
      "epoch": 0.8173986616414122,
      "grad_norm": 22.0,
      "learning_rate": 1.7739020075378819e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4626476.0,
      "step": 10627
    },
    {
      "epoch": 0.8174755788016307,
      "grad_norm": 27.125,
      "learning_rate": 1.7737866317975541e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4626798.0,
      "step": 10628
    },
    {
      "epoch": 0.8175524959618491,
      "grad_norm": 26.0,
      "learning_rate": 1.7736712560572266e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 4627078.0,
      "step": 10629
    },
    {
      "epoch": 0.8176294131220675,
      "grad_norm": 9.3125,
      "learning_rate": 1.7735558803168986e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9949238300323486,
      "num_tokens": 4627670.0,
      "step": 10630
    },
    {
      "epoch": 0.817706330282286,
      "grad_norm": 62.75,
      "learning_rate": 1.773440504576571e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9789789915084839,
      "num_tokens": 4628006.0,
      "step": 10631
    },
    {
      "epoch": 0.8177832474425044,
      "grad_norm": 37.5,
      "learning_rate": 1.7733251288362434e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 4628264.0,
      "step": 10632
    },
    {
      "epoch": 0.8178601646027228,
      "grad_norm": 4.03125,
      "learning_rate": 1.7732097530959158e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9974489808082581,
      "num_tokens": 4629050.0,
      "step": 10633
    },
    {
      "epoch": 0.8179370817629413,
      "grad_norm": 8.5,
      "learning_rate": 1.7730943773555883e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 4629566.0,
      "step": 10634
    },
    {
      "epoch": 0.8180139989231597,
      "grad_norm": 28.125,
      "learning_rate": 1.7729790016152603e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 4629910.0,
      "step": 10635
    },
    {
      "epoch": 0.8180909160833783,
      "grad_norm": 31.5,
      "learning_rate": 1.7728636258749326e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4630214.0,
      "step": 10636
    },
    {
      "epoch": 0.8181678332435967,
      "grad_norm": 15.9375,
      "learning_rate": 1.772748250134605e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4630756.0,
      "step": 10637
    },
    {
      "epoch": 0.8182447504038151,
      "grad_norm": 23.5,
      "learning_rate": 1.7726328743942775e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4631096.0,
      "step": 10638
    },
    {
      "epoch": 0.8183216675640336,
      "grad_norm": 12.0,
      "learning_rate": 1.7725174986539498e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4631404.0,
      "step": 10639
    },
    {
      "epoch": 0.818398584724252,
      "grad_norm": 23.625,
      "learning_rate": 1.7724021229136218e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9888888597488403,
      "num_tokens": 4631766.0,
      "step": 10640
    },
    {
      "epoch": 0.8184755018844704,
      "grad_norm": 27.125,
      "learning_rate": 1.7722867471732943e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 4632104.0,
      "step": 10641
    },
    {
      "epoch": 0.8185524190446889,
      "grad_norm": 6.0625,
      "learning_rate": 1.7721713714329668e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 4632638.0,
      "step": 10642
    },
    {
      "epoch": 0.8186293362049073,
      "grad_norm": 22.625,
      "learning_rate": 1.772055995692639e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4632974.0,
      "step": 10643
    },
    {
      "epoch": 0.8187062533651258,
      "grad_norm": 14.1875,
      "learning_rate": 1.7719406199523115e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9967741966247559,
      "num_tokens": 4633286.0,
      "step": 10644
    },
    {
      "epoch": 0.8187831705253442,
      "grad_norm": 31.0,
      "learning_rate": 1.771825244211984e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 4633612.0,
      "step": 10645
    },
    {
      "epoch": 0.8188600876855626,
      "grad_norm": 24.25,
      "learning_rate": 1.771709868471656e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 4634146.0,
      "step": 10646
    },
    {
      "epoch": 0.8189370048457811,
      "grad_norm": 55.0,
      "learning_rate": 1.7715944927313283e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4634422.0,
      "step": 10647
    },
    {
      "epoch": 0.8190139220059995,
      "grad_norm": 6.1875,
      "learning_rate": 1.7714791169910008e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4635204.0,
      "step": 10648
    },
    {
      "epoch": 0.8190908391662179,
      "grad_norm": 7.40625,
      "learning_rate": 1.7713637412506732e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4635488.0,
      "step": 10649
    },
    {
      "epoch": 0.8191677563264365,
      "grad_norm": 21.75,
      "learning_rate": 1.7712483655103455e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4635826.0,
      "step": 10650
    },
    {
      "epoch": 0.8192446734866549,
      "grad_norm": 73.5,
      "learning_rate": 1.7711329897700175e-06,
      "loss": 0.0507,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 4636058.0,
      "step": 10651
    },
    {
      "epoch": 0.8193215906468733,
      "grad_norm": 16.625,
      "learning_rate": 1.77101761402969e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 4636402.0,
      "step": 10652
    },
    {
      "epoch": 0.8193985078070918,
      "grad_norm": 25.875,
      "learning_rate": 1.7709022382893625e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4636748.0,
      "step": 10653
    },
    {
      "epoch": 0.8194754249673102,
      "grad_norm": 30.625,
      "learning_rate": 1.7707868625490347e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 4637096.0,
      "step": 10654
    },
    {
      "epoch": 0.8195523421275287,
      "grad_norm": 7.9375,
      "learning_rate": 1.7706714868087072e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 4637904.0,
      "step": 10655
    },
    {
      "epoch": 0.8196292592877471,
      "grad_norm": 27.75,
      "learning_rate": 1.7705561110683795e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 4638242.0,
      "step": 10656
    },
    {
      "epoch": 0.8197061764479655,
      "grad_norm": 13.5,
      "learning_rate": 1.7704407353280517e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4638562.0,
      "step": 10657
    },
    {
      "epoch": 0.819783093608184,
      "grad_norm": 23.875,
      "learning_rate": 1.770325359587724e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 4638820.0,
      "step": 10658
    },
    {
      "epoch": 0.8198600107684024,
      "grad_norm": 17.75,
      "learning_rate": 1.7702099838473964e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4639362.0,
      "step": 10659
    },
    {
      "epoch": 0.8199369279286208,
      "grad_norm": 12.25,
      "learning_rate": 1.7700946081070687e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4639660.0,
      "step": 10660
    },
    {
      "epoch": 0.8200138450888393,
      "grad_norm": 10.6875,
      "learning_rate": 1.7699792323667412e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4640206.0,
      "step": 10661
    },
    {
      "epoch": 0.8200907622490577,
      "grad_norm": 34.75,
      "learning_rate": 1.7698638566264132e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4640488.0,
      "step": 10662
    },
    {
      "epoch": 0.8201676794092763,
      "grad_norm": 19.25,
      "learning_rate": 1.7697484808860857e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 4640818.0,
      "step": 10663
    },
    {
      "epoch": 0.8202445965694947,
      "grad_norm": 10.0,
      "learning_rate": 1.769633105145758e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4641366.0,
      "step": 10664
    },
    {
      "epoch": 0.8203215137297131,
      "grad_norm": 8.625,
      "learning_rate": 1.7695177294054304e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 4641938.0,
      "step": 10665
    },
    {
      "epoch": 0.8203984308899316,
      "grad_norm": 7.71875,
      "learning_rate": 1.7694023536651029e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9980732202529907,
      "num_tokens": 4642458.0,
      "step": 10666
    },
    {
      "epoch": 0.82047534805015,
      "grad_norm": 6.59375,
      "learning_rate": 1.7692869779247751e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 4642978.0,
      "step": 10667
    },
    {
      "epoch": 0.8205522652103684,
      "grad_norm": 36.0,
      "learning_rate": 1.7691716021844472e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 4643268.0,
      "step": 10668
    },
    {
      "epoch": 0.8206291823705869,
      "grad_norm": 27.875,
      "learning_rate": 1.7690562264441197e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4643544.0,
      "step": 10669
    },
    {
      "epoch": 0.8207060995308053,
      "grad_norm": 10.1875,
      "learning_rate": 1.7689408507037921e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 4644358.0,
      "step": 10670
    },
    {
      "epoch": 0.8207830166910237,
      "grad_norm": 21.375,
      "learning_rate": 1.7688254749634644e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 4644610.0,
      "step": 10671
    },
    {
      "epoch": 0.8208599338512422,
      "grad_norm": 30.0,
      "learning_rate": 1.7687100992231369e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4644928.0,
      "step": 10672
    },
    {
      "epoch": 0.8209368510114606,
      "grad_norm": 19.125,
      "learning_rate": 1.768594723482809e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9935483932495117,
      "num_tokens": 4645240.0,
      "step": 10673
    },
    {
      "epoch": 0.8210137681716791,
      "grad_norm": 22.75,
      "learning_rate": 1.7684793477424814e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4645786.0,
      "step": 10674
    },
    {
      "epoch": 0.8210906853318976,
      "grad_norm": 18.125,
      "learning_rate": 1.7683639720021536e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9892473220825195,
      "num_tokens": 4646066.0,
      "step": 10675
    },
    {
      "epoch": 0.821167602492116,
      "grad_norm": 36.0,
      "learning_rate": 1.768248596261826e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4646376.0,
      "step": 10676
    },
    {
      "epoch": 0.8212445196523345,
      "grad_norm": 19.0,
      "learning_rate": 1.7681332205214986e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 4646720.0,
      "step": 10677
    },
    {
      "epoch": 0.8213214368125529,
      "grad_norm": 16.375,
      "learning_rate": 1.7680178447811706e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4647502.0,
      "step": 10678
    },
    {
      "epoch": 0.8213983539727713,
      "grad_norm": 25.5,
      "learning_rate": 1.7679024690408429e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 4647820.0,
      "step": 10679
    },
    {
      "epoch": 0.8214752711329898,
      "grad_norm": 8.8125,
      "learning_rate": 1.7677870933005153e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4648132.0,
      "step": 10680
    },
    {
      "epoch": 0.8215521882932082,
      "grad_norm": 23.375,
      "learning_rate": 1.7676717175601878e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9925187230110168,
      "num_tokens": 4648936.0,
      "step": 10681
    },
    {
      "epoch": 0.8216291054534267,
      "grad_norm": 27.875,
      "learning_rate": 1.76755634181986e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 4649246.0,
      "step": 10682
    },
    {
      "epoch": 0.8217060226136451,
      "grad_norm": 10.0,
      "learning_rate": 1.7674409660795325e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 4650050.0,
      "step": 10683
    },
    {
      "epoch": 0.8217829397738635,
      "grad_norm": 17.5,
      "learning_rate": 1.7673255903392046e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4650370.0,
      "step": 10684
    },
    {
      "epoch": 0.821859856934082,
      "grad_norm": 6.78125,
      "learning_rate": 1.767210214598877e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 4650910.0,
      "step": 10685
    },
    {
      "epoch": 0.8219367740943004,
      "grad_norm": 7.78125,
      "learning_rate": 1.7670948388585493e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 4651444.0,
      "step": 10686
    },
    {
      "epoch": 0.8220136912545188,
      "grad_norm": 42.75,
      "learning_rate": 1.7669794631182218e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 4651816.0,
      "step": 10687
    },
    {
      "epoch": 0.8220906084147374,
      "grad_norm": 5.75,
      "learning_rate": 1.7668640873778943e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9961612224578857,
      "num_tokens": 4652338.0,
      "step": 10688
    },
    {
      "epoch": 0.8221675255749558,
      "grad_norm": 39.25,
      "learning_rate": 1.7667487116375663e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9965217113494873,
      "num_tokens": 4652914.0,
      "step": 10689
    },
    {
      "epoch": 0.8222444427351742,
      "grad_norm": 49.75,
      "learning_rate": 1.7666333358972386e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4653242.0,
      "step": 10690
    },
    {
      "epoch": 0.8223213598953927,
      "grad_norm": 14.25,
      "learning_rate": 1.766517960156911e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4653822.0,
      "step": 10691
    },
    {
      "epoch": 0.8223982770556111,
      "grad_norm": 8.3125,
      "learning_rate": 1.7664025844165835e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9981784820556641,
      "num_tokens": 4654372.0,
      "step": 10692
    },
    {
      "epoch": 0.8224751942158296,
      "grad_norm": 19.625,
      "learning_rate": 1.7662872086762558e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4654696.0,
      "step": 10693
    },
    {
      "epoch": 0.822552111376048,
      "grad_norm": 40.75,
      "learning_rate": 1.7661718329359282e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4654994.0,
      "step": 10694
    },
    {
      "epoch": 0.8226290285362664,
      "grad_norm": 12.0625,
      "learning_rate": 1.7660564571956003e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.998171865940094,
      "num_tokens": 4655542.0,
      "step": 10695
    },
    {
      "epoch": 0.8227059456964849,
      "grad_norm": 43.25,
      "learning_rate": 1.7659410814552727e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9802371263504028,
      "num_tokens": 4655796.0,
      "step": 10696
    },
    {
      "epoch": 0.8227828628567033,
      "grad_norm": 7.5625,
      "learning_rate": 1.765825705714945e-06,
      "loss": 0.0064,
      "mean_token_accuracy": 0.9987593293190002,
      "num_tokens": 4656604.0,
      "step": 10697
    },
    {
      "epoch": 0.8228597800169217,
      "grad_norm": 33.0,
      "learning_rate": 1.7657103299746175e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4656914.0,
      "step": 10698
    },
    {
      "epoch": 0.8229366971771402,
      "grad_norm": 7.625,
      "learning_rate": 1.76559495423429e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4657472.0,
      "step": 10699
    },
    {
      "epoch": 0.8230136143373586,
      "grad_norm": 67.5,
      "learning_rate": 1.765479578493962e-06,
      "loss": 0.0482,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 4657772.0,
      "step": 10700
    },
    {
      "epoch": 0.8230905314975772,
      "grad_norm": 29.625,
      "learning_rate": 1.7653642027536342e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9858155846595764,
      "num_tokens": 4658056.0,
      "step": 10701
    },
    {
      "epoch": 0.8231674486577956,
      "grad_norm": 67.0,
      "learning_rate": 1.7652488270133067e-06,
      "loss": 0.0438,
      "mean_token_accuracy": 0.9816513657569885,
      "num_tokens": 4658384.0,
      "step": 10702
    },
    {
      "epoch": 0.823244365818014,
      "grad_norm": 22.625,
      "learning_rate": 1.7651334512729792e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 4658674.0,
      "step": 10703
    },
    {
      "epoch": 0.8233212829782325,
      "grad_norm": 36.75,
      "learning_rate": 1.7650180755326514e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4659248.0,
      "step": 10704
    },
    {
      "epoch": 0.8233982001384509,
      "grad_norm": 15.625,
      "learning_rate": 1.7649026997923235e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4659534.0,
      "step": 10705
    },
    {
      "epoch": 0.8234751172986693,
      "grad_norm": 54.5,
      "learning_rate": 1.764787324051996e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4659838.0,
      "step": 10706
    },
    {
      "epoch": 0.8235520344588878,
      "grad_norm": 10.4375,
      "learning_rate": 1.7646719483116684e-06,
      "loss": 0.0054,
      "mean_token_accuracy": 0.9990394115447998,
      "num_tokens": 4660882.0,
      "step": 10707
    },
    {
      "epoch": 0.8236289516191062,
      "grad_norm": 46.75,
      "learning_rate": 1.7645565725713407e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4661438.0,
      "step": 10708
    },
    {
      "epoch": 0.8237058687793246,
      "grad_norm": 35.75,
      "learning_rate": 1.7644411968310132e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9969696998596191,
      "num_tokens": 4661770.0,
      "step": 10709
    },
    {
      "epoch": 0.8237827859395431,
      "grad_norm": 37.25,
      "learning_rate": 1.7643258210906854e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9851301312446594,
      "num_tokens": 4662040.0,
      "step": 10710
    },
    {
      "epoch": 0.8238597030997615,
      "grad_norm": 41.0,
      "learning_rate": 1.7642104453503577e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4662604.0,
      "step": 10711
    },
    {
      "epoch": 0.82393662025998,
      "grad_norm": 10.75,
      "learning_rate": 1.76409506961003e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4662948.0,
      "step": 10712
    },
    {
      "epoch": 0.8240135374201984,
      "grad_norm": 26.875,
      "learning_rate": 1.7639796938697024e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 4663236.0,
      "step": 10713
    },
    {
      "epoch": 0.8240904545804169,
      "grad_norm": 38.25,
      "learning_rate": 1.7638643181293747e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4663798.0,
      "step": 10714
    },
    {
      "epoch": 0.8241673717406354,
      "grad_norm": 14.5625,
      "learning_rate": 1.7637489423890471e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 4664054.0,
      "step": 10715
    },
    {
      "epoch": 0.8242442889008538,
      "grad_norm": 13.25,
      "learning_rate": 1.7636335666487192e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4664638.0,
      "step": 10716
    },
    {
      "epoch": 0.8243212060610722,
      "grad_norm": 3.234375,
      "learning_rate": 1.7635181909083916e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4665184.0,
      "step": 10717
    },
    {
      "epoch": 0.8243981232212907,
      "grad_norm": 15.5625,
      "learning_rate": 1.763402815168064e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4665504.0,
      "step": 10718
    },
    {
      "epoch": 0.8244750403815091,
      "grad_norm": 62.0,
      "learning_rate": 1.7632874394277364e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.97826087474823,
      "num_tokens": 4665782.0,
      "step": 10719
    },
    {
      "epoch": 0.8245519575417276,
      "grad_norm": 35.75,
      "learning_rate": 1.7631720636874088e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4666098.0,
      "step": 10720
    },
    {
      "epoch": 0.824628874701946,
      "grad_norm": 7.3125,
      "learning_rate": 1.763056687947081e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.9974293112754822,
      "num_tokens": 4666878.0,
      "step": 10721
    },
    {
      "epoch": 0.8247057918621644,
      "grad_norm": 17.75,
      "learning_rate": 1.7629413122067531e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4667452.0,
      "step": 10722
    },
    {
      "epoch": 0.8247827090223829,
      "grad_norm": 10.9375,
      "learning_rate": 1.7628259364664256e-06,
      "loss": 0.0058,
      "mean_token_accuracy": 0.9976019263267517,
      "num_tokens": 4668288.0,
      "step": 10723
    },
    {
      "epoch": 0.8248596261826013,
      "grad_norm": 15.0625,
      "learning_rate": 1.762710560726098e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4668806.0,
      "step": 10724
    },
    {
      "epoch": 0.8249365433428197,
      "grad_norm": 12.375,
      "learning_rate": 1.7625951849857703e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4669390.0,
      "step": 10725
    },
    {
      "epoch": 0.8250134605030383,
      "grad_norm": 12.4375,
      "learning_rate": 1.7624798092454428e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4669976.0,
      "step": 10726
    },
    {
      "epoch": 0.8250903776632567,
      "grad_norm": 37.25,
      "learning_rate": 1.7623644335051149e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4670292.0,
      "step": 10727
    },
    {
      "epoch": 0.8251672948234752,
      "grad_norm": 19.375,
      "learning_rate": 1.7622490577647873e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 4670626.0,
      "step": 10728
    },
    {
      "epoch": 0.8252442119836936,
      "grad_norm": 31.25,
      "learning_rate": 1.7621336820244596e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4670930.0,
      "step": 10729
    },
    {
      "epoch": 0.825321129143912,
      "grad_norm": 11.6875,
      "learning_rate": 1.762018306284132e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 4671494.0,
      "step": 10730
    },
    {
      "epoch": 0.8253980463041305,
      "grad_norm": 26.5,
      "learning_rate": 1.7619029305438045e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4671810.0,
      "step": 10731
    },
    {
      "epoch": 0.8254749634643489,
      "grad_norm": 29.625,
      "learning_rate": 1.7617875548034768e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 4672382.0,
      "step": 10732
    },
    {
      "epoch": 0.8255518806245673,
      "grad_norm": 9.3125,
      "learning_rate": 1.7616721790631488e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4672928.0,
      "step": 10733
    },
    {
      "epoch": 0.8256287977847858,
      "grad_norm": 25.875,
      "learning_rate": 1.7615568033228213e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4673228.0,
      "step": 10734
    },
    {
      "epoch": 0.8257057149450042,
      "grad_norm": 12.9375,
      "learning_rate": 1.7614414275824938e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9948849081993103,
      "num_tokens": 4674012.0,
      "step": 10735
    },
    {
      "epoch": 0.8257826321052226,
      "grad_norm": 22.0,
      "learning_rate": 1.761326051842166e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4674268.0,
      "step": 10736
    },
    {
      "epoch": 0.8258595492654411,
      "grad_norm": 41.75,
      "learning_rate": 1.7612106761018385e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4674574.0,
      "step": 10737
    },
    {
      "epoch": 0.8259364664256595,
      "grad_norm": 15.375,
      "learning_rate": 1.7610953003615105e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4674880.0,
      "step": 10738
    },
    {
      "epoch": 0.8260133835858781,
      "grad_norm": 35.75,
      "learning_rate": 1.760979924621183e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4675432.0,
      "step": 10739
    },
    {
      "epoch": 0.8260903007460965,
      "grad_norm": 50.0,
      "learning_rate": 1.7608645488808553e-06,
      "loss": 0.0396,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4675754.0,
      "step": 10740
    },
    {
      "epoch": 0.8261672179063149,
      "grad_norm": 50.0,
      "learning_rate": 1.7607491731405277e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 4676580.0,
      "step": 10741
    },
    {
      "epoch": 0.8262441350665334,
      "grad_norm": 15.875,
      "learning_rate": 1.7606337974002002e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 4676910.0,
      "step": 10742
    },
    {
      "epoch": 0.8263210522267518,
      "grad_norm": 39.25,
      "learning_rate": 1.7605184216598723e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4677460.0,
      "step": 10743
    },
    {
      "epoch": 0.8263979693869702,
      "grad_norm": 90.0,
      "learning_rate": 1.7604030459195445e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4677792.0,
      "step": 10744
    },
    {
      "epoch": 0.8264748865471887,
      "grad_norm": 9.625,
      "learning_rate": 1.760287670179217e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 4678350.0,
      "step": 10745
    },
    {
      "epoch": 0.8265518037074071,
      "grad_norm": 28.5,
      "learning_rate": 1.7601722944388895e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9975728392601013,
      "num_tokens": 4679176.0,
      "step": 10746
    },
    {
      "epoch": 0.8266287208676256,
      "grad_norm": 25.625,
      "learning_rate": 1.7600569186985617e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4679460.0,
      "step": 10747
    },
    {
      "epoch": 0.826705638027844,
      "grad_norm": 32.75,
      "learning_rate": 1.7599415429582342e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 4679788.0,
      "step": 10748
    },
    {
      "epoch": 0.8267825551880624,
      "grad_norm": 5.09375,
      "learning_rate": 1.7598261672179062e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4680374.0,
      "step": 10749
    },
    {
      "epoch": 0.8268594723482809,
      "grad_norm": 26.625,
      "learning_rate": 1.7597107914775787e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4680668.0,
      "step": 10750
    },
    {
      "epoch": 0.8269363895084993,
      "grad_norm": 17.625,
      "learning_rate": 1.759595415737251e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4681012.0,
      "step": 10751
    },
    {
      "epoch": 0.8270133066687178,
      "grad_norm": 6.46875,
      "learning_rate": 1.7594800399969234e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 4681554.0,
      "step": 10752
    },
    {
      "epoch": 0.8270902238289363,
      "grad_norm": 9.9375,
      "learning_rate": 1.759364664256596e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4682102.0,
      "step": 10753
    },
    {
      "epoch": 0.8271671409891547,
      "grad_norm": 11.8125,
      "learning_rate": 1.759249288516268e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4682448.0,
      "step": 10754
    },
    {
      "epoch": 0.8272440581493731,
      "grad_norm": 9.1875,
      "learning_rate": 1.7591339127759402e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 4683044.0,
      "step": 10755
    },
    {
      "epoch": 0.8273209753095916,
      "grad_norm": 12.75,
      "learning_rate": 1.7590185370356127e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4683562.0,
      "step": 10756
    },
    {
      "epoch": 0.82739789246981,
      "grad_norm": 40.75,
      "learning_rate": 1.7589031612952851e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 4683848.0,
      "step": 10757
    },
    {
      "epoch": 0.8274748096300285,
      "grad_norm": 35.0,
      "learning_rate": 1.7587877855549574e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 4684192.0,
      "step": 10758
    },
    {
      "epoch": 0.8275517267902469,
      "grad_norm": 47.5,
      "learning_rate": 1.7586724098146299e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4684498.0,
      "step": 10759
    },
    {
      "epoch": 0.8276286439504653,
      "grad_norm": 29.5,
      "learning_rate": 1.758557034074302e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 4684806.0,
      "step": 10760
    },
    {
      "epoch": 0.8277055611106838,
      "grad_norm": 26.875,
      "learning_rate": 1.7584416583339744e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4685354.0,
      "step": 10761
    },
    {
      "epoch": 0.8277824782709022,
      "grad_norm": 26.625,
      "learning_rate": 1.7583262825936466e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4685924.0,
      "step": 10762
    },
    {
      "epoch": 0.8278593954311206,
      "grad_norm": 40.5,
      "learning_rate": 1.7582109068533191e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4686252.0,
      "step": 10763
    },
    {
      "epoch": 0.8279363125913392,
      "grad_norm": 21.125,
      "learning_rate": 1.7580955311129914e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 4687060.0,
      "step": 10764
    },
    {
      "epoch": 0.8280132297515576,
      "grad_norm": 62.75,
      "learning_rate": 1.7579801553726636e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4687384.0,
      "step": 10765
    },
    {
      "epoch": 0.8280901469117761,
      "grad_norm": 27.625,
      "learning_rate": 1.7578647796323359e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4687682.0,
      "step": 10766
    },
    {
      "epoch": 0.8281670640719945,
      "grad_norm": 12.5,
      "learning_rate": 1.7577494038920084e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 4687990.0,
      "step": 10767
    },
    {
      "epoch": 0.8282439812322129,
      "grad_norm": 22.5,
      "learning_rate": 1.7576340281516806e-06,
      "loss": 0.0427,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 4688270.0,
      "step": 10768
    },
    {
      "epoch": 0.8283208983924314,
      "grad_norm": 9.3125,
      "learning_rate": 1.757518652411353e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 4688608.0,
      "step": 10769
    },
    {
      "epoch": 0.8283978155526498,
      "grad_norm": 12.6875,
      "learning_rate": 1.7574032766710256e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 4688930.0,
      "step": 10770
    },
    {
      "epoch": 0.8284747327128682,
      "grad_norm": 18.25,
      "learning_rate": 1.7572879009306976e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4689476.0,
      "step": 10771
    },
    {
      "epoch": 0.8285516498730867,
      "grad_norm": 16.25,
      "learning_rate": 1.7571725251903699e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4689798.0,
      "step": 10772
    },
    {
      "epoch": 0.8286285670333051,
      "grad_norm": 19.5,
      "learning_rate": 1.7570571494500423e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 4690110.0,
      "step": 10773
    },
    {
      "epoch": 0.8287054841935235,
      "grad_norm": 41.25,
      "learning_rate": 1.7569417737097148e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9967213273048401,
      "num_tokens": 4690416.0,
      "step": 10774
    },
    {
      "epoch": 0.828782401353742,
      "grad_norm": 47.75,
      "learning_rate": 1.756826397969387e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4690732.0,
      "step": 10775
    },
    {
      "epoch": 0.8288593185139604,
      "grad_norm": 13.6875,
      "learning_rate": 1.7567110222290593e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 4691012.0,
      "step": 10776
    },
    {
      "epoch": 0.828936235674179,
      "grad_norm": 24.5,
      "learning_rate": 1.7565956464887316e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4691352.0,
      "step": 10777
    },
    {
      "epoch": 0.8290131528343974,
      "grad_norm": 25.75,
      "learning_rate": 1.756480270748404e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.991055428981781,
      "num_tokens": 4691912.0,
      "step": 10778
    },
    {
      "epoch": 0.8290900699946158,
      "grad_norm": 41.5,
      "learning_rate": 1.7563648950080763e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9923809766769409,
      "num_tokens": 4692438.0,
      "step": 10779
    },
    {
      "epoch": 0.8291669871548343,
      "grad_norm": 25.375,
      "learning_rate": 1.7562495192677488e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9928443431854248,
      "num_tokens": 4692998.0,
      "step": 10780
    },
    {
      "epoch": 0.8292439043150527,
      "grad_norm": 39.75,
      "learning_rate": 1.7561341435274208e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4693314.0,
      "step": 10781
    },
    {
      "epoch": 0.8293208214752711,
      "grad_norm": 21.5,
      "learning_rate": 1.7560187677870933e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4693596.0,
      "step": 10782
    },
    {
      "epoch": 0.8293977386354896,
      "grad_norm": 24.875,
      "learning_rate": 1.7559033920467656e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4693940.0,
      "step": 10783
    },
    {
      "epoch": 0.829474655795708,
      "grad_norm": 10.0,
      "learning_rate": 1.755788016306438e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 4694498.0,
      "step": 10784
    },
    {
      "epoch": 0.8295515729559265,
      "grad_norm": 18.125,
      "learning_rate": 1.7556726405661105e-06,
      "loss": 0.0185,
      "mean_token_accuracy": 0.9967741966247559,
      "num_tokens": 4694810.0,
      "step": 10785
    },
    {
      "epoch": 0.8296284901161449,
      "grad_norm": 34.75,
      "learning_rate": 1.7555572648257827e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4695066.0,
      "step": 10786
    },
    {
      "epoch": 0.8297054072763633,
      "grad_norm": 14.9375,
      "learning_rate": 1.7554418890854548e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4695614.0,
      "step": 10787
    },
    {
      "epoch": 0.8297823244365818,
      "grad_norm": 20.375,
      "learning_rate": 1.7553265133451273e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4695920.0,
      "step": 10788
    },
    {
      "epoch": 0.8298592415968002,
      "grad_norm": 15.1875,
      "learning_rate": 1.7552111376047997e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 4696414.0,
      "step": 10789
    },
    {
      "epoch": 0.8299361587570186,
      "grad_norm": 17.5,
      "learning_rate": 1.755095761864472e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4697000.0,
      "step": 10790
    },
    {
      "epoch": 0.8300130759172372,
      "grad_norm": 14.5625,
      "learning_rate": 1.7549803861241445e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4697310.0,
      "step": 10791
    },
    {
      "epoch": 0.8300899930774556,
      "grad_norm": 18.0,
      "learning_rate": 1.7548650103838165e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 4697652.0,
      "step": 10792
    },
    {
      "epoch": 0.830166910237674,
      "grad_norm": 7.3125,
      "learning_rate": 1.754749634643489e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4698176.0,
      "step": 10793
    },
    {
      "epoch": 0.8302438273978925,
      "grad_norm": 32.75,
      "learning_rate": 1.7546342589031612e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 4698456.0,
      "step": 10794
    },
    {
      "epoch": 0.8303207445581109,
      "grad_norm": 56.5,
      "learning_rate": 1.7545188831628337e-06,
      "loss": 0.0378,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 4698716.0,
      "step": 10795
    },
    {
      "epoch": 0.8303976617183294,
      "grad_norm": 35.75,
      "learning_rate": 1.7544035074225062e-06,
      "loss": 0.0309,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4699020.0,
      "step": 10796
    },
    {
      "epoch": 0.8304745788785478,
      "grad_norm": 24.875,
      "learning_rate": 1.7542881316821784e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4699298.0,
      "step": 10797
    },
    {
      "epoch": 0.8305514960387662,
      "grad_norm": 4.6875,
      "learning_rate": 1.7541727559418505e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4699816.0,
      "step": 10798
    },
    {
      "epoch": 0.8306284131989847,
      "grad_norm": 30.75,
      "learning_rate": 1.754057380201523e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4700378.0,
      "step": 10799
    },
    {
      "epoch": 0.8307053303592031,
      "grad_norm": 26.75,
      "learning_rate": 1.7539420044611954e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4700718.0,
      "step": 10800
    },
    {
      "epoch": 0.8307822475194215,
      "grad_norm": 26.625,
      "learning_rate": 1.7538266287208677e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4701034.0,
      "step": 10801
    },
    {
      "epoch": 0.83085916467964,
      "grad_norm": 16.875,
      "learning_rate": 1.7537112529805402e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4701580.0,
      "step": 10802
    },
    {
      "epoch": 0.8309360818398585,
      "grad_norm": 68.0,
      "learning_rate": 1.7535958772402122e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 4701932.0,
      "step": 10803
    },
    {
      "epoch": 0.831012999000077,
      "grad_norm": 24.625,
      "learning_rate": 1.7534805014998847e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 4702222.0,
      "step": 10804
    },
    {
      "epoch": 0.8310899161602954,
      "grad_norm": 30.625,
      "learning_rate": 1.753365125759557e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 4702758.0,
      "step": 10805
    },
    {
      "epoch": 0.8311668333205138,
      "grad_norm": 13.125,
      "learning_rate": 1.7532497500192294e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4703074.0,
      "step": 10806
    },
    {
      "epoch": 0.8312437504807323,
      "grad_norm": 16.5,
      "learning_rate": 1.7531343742789019e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4703350.0,
      "step": 10807
    },
    {
      "epoch": 0.8313206676409507,
      "grad_norm": 10.5625,
      "learning_rate": 1.753018998538574e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 4704170.0,
      "step": 10808
    },
    {
      "epoch": 0.8313975848011691,
      "grad_norm": 6.25,
      "learning_rate": 1.7529036227982462e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9961832165718079,
      "num_tokens": 4704958.0,
      "step": 10809
    },
    {
      "epoch": 0.8314745019613876,
      "grad_norm": 38.0,
      "learning_rate": 1.7527882470579186e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 4705250.0,
      "step": 10810
    },
    {
      "epoch": 0.831551419121606,
      "grad_norm": 32.5,
      "learning_rate": 1.7526728713175911e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4705566.0,
      "step": 10811
    },
    {
      "epoch": 0.8316283362818244,
      "grad_norm": 14.0625,
      "learning_rate": 1.7525574955772634e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4705866.0,
      "step": 10812
    },
    {
      "epoch": 0.8317052534420429,
      "grad_norm": 32.75,
      "learning_rate": 1.7524421198369358e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4706184.0,
      "step": 10813
    },
    {
      "epoch": 0.8317821706022613,
      "grad_norm": 41.5,
      "learning_rate": 1.7523267440966079e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 4706532.0,
      "step": 10814
    },
    {
      "epoch": 0.8318590877624799,
      "grad_norm": 27.625,
      "learning_rate": 1.7522113683562804e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4706844.0,
      "step": 10815
    },
    {
      "epoch": 0.8319360049226983,
      "grad_norm": 25.0,
      "learning_rate": 1.7520959926159526e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4707166.0,
      "step": 10816
    },
    {
      "epoch": 0.8320129220829167,
      "grad_norm": 14.25,
      "learning_rate": 1.751980616875625e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 4707730.0,
      "step": 10817
    },
    {
      "epoch": 0.8320898392431352,
      "grad_norm": 28.25,
      "learning_rate": 1.7518652411352976e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 4708018.0,
      "step": 10818
    },
    {
      "epoch": 0.8321667564033536,
      "grad_norm": 21.625,
      "learning_rate": 1.7517498653949696e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4708574.0,
      "step": 10819
    },
    {
      "epoch": 0.832243673563572,
      "grad_norm": 25.125,
      "learning_rate": 1.7516344896546419e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4708858.0,
      "step": 10820
    },
    {
      "epoch": 0.8323205907237905,
      "grad_norm": 41.5,
      "learning_rate": 1.7515191139143143e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4709434.0,
      "step": 10821
    },
    {
      "epoch": 0.8323975078840089,
      "grad_norm": 14.4375,
      "learning_rate": 1.7514037381739868e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4709728.0,
      "step": 10822
    },
    {
      "epoch": 0.8324744250442274,
      "grad_norm": 10.875,
      "learning_rate": 1.751288362433659e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 4710512.0,
      "step": 10823
    },
    {
      "epoch": 0.8325513422044458,
      "grad_norm": 27.625,
      "learning_rate": 1.7511729866933315e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 4710816.0,
      "step": 10824
    },
    {
      "epoch": 0.8326282593646642,
      "grad_norm": 57.25,
      "learning_rate": 1.7510576109530036e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9973333477973938,
      "num_tokens": 4711194.0,
      "step": 10825
    },
    {
      "epoch": 0.8327051765248827,
      "grad_norm": 18.25,
      "learning_rate": 1.750942235212676e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4711534.0,
      "step": 10826
    },
    {
      "epoch": 0.8327820936851011,
      "grad_norm": 18.0,
      "learning_rate": 1.7508268594723483e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4712086.0,
      "step": 10827
    },
    {
      "epoch": 0.8328590108453195,
      "grad_norm": 12.875,
      "learning_rate": 1.7507114837320208e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 4712384.0,
      "step": 10828
    },
    {
      "epoch": 0.8329359280055381,
      "grad_norm": 30.5,
      "learning_rate": 1.750596107991693e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9912739992141724,
      "num_tokens": 4712958.0,
      "step": 10829
    },
    {
      "epoch": 0.8330128451657565,
      "grad_norm": 24.5,
      "learning_rate": 1.7504807322513653e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4713294.0,
      "step": 10830
    },
    {
      "epoch": 0.8330897623259749,
      "grad_norm": 32.75,
      "learning_rate": 1.7503653565110375e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 4713570.0,
      "step": 10831
    },
    {
      "epoch": 0.8331666794861934,
      "grad_norm": 30.25,
      "learning_rate": 1.75024998077071e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 4713858.0,
      "step": 10832
    },
    {
      "epoch": 0.8332435966464118,
      "grad_norm": 12.375,
      "learning_rate": 1.7501346050303823e-06,
      "loss": 0.0082,
      "mean_token_accuracy": 0.9963054060935974,
      "num_tokens": 4714672.0,
      "step": 10833
    },
    {
      "epoch": 0.8333205138066303,
      "grad_norm": 21.125,
      "learning_rate": 1.7500192292900547e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.993127167224884,
      "num_tokens": 4714964.0,
      "step": 10834
    },
    {
      "epoch": 0.8333974309668487,
      "grad_norm": 24.5,
      "learning_rate": 1.7499038535497272e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 4715488.0,
      "step": 10835
    },
    {
      "epoch": 0.8334743481270671,
      "grad_norm": 7.90625,
      "learning_rate": 1.7497884778093993e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9982847571372986,
      "num_tokens": 4716072.0,
      "step": 10836
    },
    {
      "epoch": 0.8335512652872856,
      "grad_norm": 27.375,
      "learning_rate": 1.7496731020690715e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 4716410.0,
      "step": 10837
    },
    {
      "epoch": 0.833628182447504,
      "grad_norm": 26.0,
      "learning_rate": 1.749557726328744e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4716708.0,
      "step": 10838
    },
    {
      "epoch": 0.8337050996077224,
      "grad_norm": 32.25,
      "learning_rate": 1.7494423505884165e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 4716984.0,
      "step": 10839
    },
    {
      "epoch": 0.833782016767941,
      "grad_norm": 14.4375,
      "learning_rate": 1.7493269748480887e-06,
      "loss": 0.0076,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 4717780.0,
      "step": 10840
    },
    {
      "epoch": 0.8338589339281594,
      "grad_norm": 13.875,
      "learning_rate": 1.7492115991077608e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 4718564.0,
      "step": 10841
    },
    {
      "epoch": 0.8339358510883779,
      "grad_norm": 13.5625,
      "learning_rate": 1.7490962233674332e-06,
      "loss": 0.0097,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4719372.0,
      "step": 10842
    },
    {
      "epoch": 0.8340127682485963,
      "grad_norm": 13.0,
      "learning_rate": 1.7489808476271057e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9911816716194153,
      "num_tokens": 4719940.0,
      "step": 10843
    },
    {
      "epoch": 0.8340896854088147,
      "grad_norm": 21.375,
      "learning_rate": 1.748865471886778e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9949109554290771,
      "num_tokens": 4720728.0,
      "step": 10844
    },
    {
      "epoch": 0.8341666025690332,
      "grad_norm": 39.75,
      "learning_rate": 1.7487500961464504e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4721010.0,
      "step": 10845
    },
    {
      "epoch": 0.8342435197292516,
      "grad_norm": 5.75,
      "learning_rate": 1.7486347204061225e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4721556.0,
      "step": 10846
    },
    {
      "epoch": 0.83432043688947,
      "grad_norm": 68.0,
      "learning_rate": 1.748519344665795e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4721892.0,
      "step": 10847
    },
    {
      "epoch": 0.8343973540496885,
      "grad_norm": 26.75,
      "learning_rate": 1.7484039689254672e-06,
      "loss": 0.0344,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 4722148.0,
      "step": 10848
    },
    {
      "epoch": 0.8344742712099069,
      "grad_norm": 36.25,
      "learning_rate": 1.7482885931851397e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9975308775901794,
      "num_tokens": 4722960.0,
      "step": 10849
    },
    {
      "epoch": 0.8345511883701254,
      "grad_norm": 4.0,
      "learning_rate": 1.7481732174448121e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4723522.0,
      "step": 10850
    },
    {
      "epoch": 0.8346281055303438,
      "grad_norm": 29.0,
      "learning_rate": 1.7480578417044844e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.994575023651123,
      "num_tokens": 4724076.0,
      "step": 10851
    },
    {
      "epoch": 0.8347050226905622,
      "grad_norm": 78.5,
      "learning_rate": 1.7479424659641564e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.981249988079071,
      "num_tokens": 4724400.0,
      "step": 10852
    },
    {
      "epoch": 0.8347819398507808,
      "grad_norm": 41.5,
      "learning_rate": 1.747827090223829e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4724722.0,
      "step": 10853
    },
    {
      "epoch": 0.8348588570109992,
      "grad_norm": 14.4375,
      "learning_rate": 1.7477117144835014e-06,
      "loss": 0.0102,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4725306.0,
      "step": 10854
    },
    {
      "epoch": 0.8349357741712176,
      "grad_norm": 33.25,
      "learning_rate": 1.7475963387431736e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 4725678.0,
      "step": 10855
    },
    {
      "epoch": 0.8350126913314361,
      "grad_norm": 45.75,
      "learning_rate": 1.7474809630028461e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 4726002.0,
      "step": 10856
    },
    {
      "epoch": 0.8350896084916545,
      "grad_norm": 70.0,
      "learning_rate": 1.7473655872625182e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 4726340.0,
      "step": 10857
    },
    {
      "epoch": 0.8351665256518729,
      "grad_norm": 4.09375,
      "learning_rate": 1.7472502115221906e-06,
      "loss": 0.0051,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 4727136.0,
      "step": 10858
    },
    {
      "epoch": 0.8352434428120914,
      "grad_norm": 69.5,
      "learning_rate": 1.7471348357818629e-06,
      "loss": 0.0387,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 4727436.0,
      "step": 10859
    },
    {
      "epoch": 0.8353203599723098,
      "grad_norm": 25.875,
      "learning_rate": 1.7470194600415354e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4727734.0,
      "step": 10860
    },
    {
      "epoch": 0.8353972771325283,
      "grad_norm": 52.0,
      "learning_rate": 1.7469040843012078e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9809523820877075,
      "num_tokens": 4728050.0,
      "step": 10861
    },
    {
      "epoch": 0.8354741942927467,
      "grad_norm": 25.125,
      "learning_rate": 1.74678870856088e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4728394.0,
      "step": 10862
    },
    {
      "epoch": 0.8355511114529651,
      "grad_norm": 9.0625,
      "learning_rate": 1.7466733328205521e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4728962.0,
      "step": 10863
    },
    {
      "epoch": 0.8356280286131836,
      "grad_norm": 4.0,
      "learning_rate": 1.7465579570802246e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 4729564.0,
      "step": 10864
    },
    {
      "epoch": 0.835704945773402,
      "grad_norm": 15.0,
      "learning_rate": 1.746442581339897e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 4729900.0,
      "step": 10865
    },
    {
      "epoch": 0.8357818629336204,
      "grad_norm": 12.8125,
      "learning_rate": 1.7463272055995693e-06,
      "loss": 0.0166,
      "mean_token_accuracy": 0.994413435459137,
      "num_tokens": 4730260.0,
      "step": 10866
    },
    {
      "epoch": 0.835858780093839,
      "grad_norm": 51.5,
      "learning_rate": 1.7462118298592418e-06,
      "loss": 0.0361,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4730564.0,
      "step": 10867
    },
    {
      "epoch": 0.8359356972540574,
      "grad_norm": 19.125,
      "learning_rate": 1.7460964541189138e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 4730914.0,
      "step": 10868
    },
    {
      "epoch": 0.8360126144142759,
      "grad_norm": 59.0,
      "learning_rate": 1.7459810783785863e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 4731224.0,
      "step": 10869
    },
    {
      "epoch": 0.8360895315744943,
      "grad_norm": 48.25,
      "learning_rate": 1.7458657026382586e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4731534.0,
      "step": 10870
    },
    {
      "epoch": 0.8361664487347127,
      "grad_norm": 17.875,
      "learning_rate": 1.745750326897931e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4731874.0,
      "step": 10871
    },
    {
      "epoch": 0.8362433658949312,
      "grad_norm": 17.875,
      "learning_rate": 1.7456349511576035e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4732422.0,
      "step": 10872
    },
    {
      "epoch": 0.8363202830551496,
      "grad_norm": 21.625,
      "learning_rate": 1.7455195754172758e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4732962.0,
      "step": 10873
    },
    {
      "epoch": 0.836397200215368,
      "grad_norm": 23.375,
      "learning_rate": 1.7454041996769478e-06,
      "loss": 0.0311,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4733240.0,
      "step": 10874
    },
    {
      "epoch": 0.8364741173755865,
      "grad_norm": 43.75,
      "learning_rate": 1.7452888239366203e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4733802.0,
      "step": 10875
    },
    {
      "epoch": 0.8365510345358049,
      "grad_norm": 52.5,
      "learning_rate": 1.7451734481962928e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4734100.0,
      "step": 10876
    },
    {
      "epoch": 0.8366279516960233,
      "grad_norm": 33.5,
      "learning_rate": 1.745058072455965e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4734672.0,
      "step": 10877
    },
    {
      "epoch": 0.8367048688562418,
      "grad_norm": 18.5,
      "learning_rate": 1.7449426967156375e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4734988.0,
      "step": 10878
    },
    {
      "epoch": 0.8367817860164602,
      "grad_norm": 26.875,
      "learning_rate": 1.7448273209753095e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4735558.0,
      "step": 10879
    },
    {
      "epoch": 0.8368587031766788,
      "grad_norm": 25.375,
      "learning_rate": 1.744711945234982e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4735864.0,
      "step": 10880
    },
    {
      "epoch": 0.8369356203368972,
      "grad_norm": 28.5,
      "learning_rate": 1.7445965694946543e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9901960492134094,
      "num_tokens": 4736172.0,
      "step": 10881
    },
    {
      "epoch": 0.8370125374971156,
      "grad_norm": 41.5,
      "learning_rate": 1.7444811937543267e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 4736482.0,
      "step": 10882
    },
    {
      "epoch": 0.8370894546573341,
      "grad_norm": 16.25,
      "learning_rate": 1.744365818013999e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9931740760803223,
      "num_tokens": 4736776.0,
      "step": 10883
    },
    {
      "epoch": 0.8371663718175525,
      "grad_norm": 40.5,
      "learning_rate": 1.7442504422736712e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 4737366.0,
      "step": 10884
    },
    {
      "epoch": 0.8372432889777709,
      "grad_norm": 32.25,
      "learning_rate": 1.7441350665333435e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9900332093238831,
      "num_tokens": 4737668.0,
      "step": 10885
    },
    {
      "epoch": 0.8373202061379894,
      "grad_norm": 27.625,
      "learning_rate": 1.744019690793016e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4737952.0,
      "step": 10886
    },
    {
      "epoch": 0.8373971232982078,
      "grad_norm": 24.375,
      "learning_rate": 1.7439043150526882e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4738500.0,
      "step": 10887
    },
    {
      "epoch": 0.8374740404584263,
      "grad_norm": 51.25,
      "learning_rate": 1.7437889393123607e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4738784.0,
      "step": 10888
    },
    {
      "epoch": 0.8375509576186447,
      "grad_norm": 16.25,
      "learning_rate": 1.7436735635720332e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4739130.0,
      "step": 10889
    },
    {
      "epoch": 0.8376278747788631,
      "grad_norm": 24.0,
      "learning_rate": 1.7435581878317052e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9940476417541504,
      "num_tokens": 4739468.0,
      "step": 10890
    },
    {
      "epoch": 0.8377047919390816,
      "grad_norm": 59.75,
      "learning_rate": 1.7434428120913775e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4740032.0,
      "step": 10891
    },
    {
      "epoch": 0.8377817090993,
      "grad_norm": 10.5,
      "learning_rate": 1.74332743635105e-06,
      "loss": 0.0093,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 4740852.0,
      "step": 10892
    },
    {
      "epoch": 0.8378586262595185,
      "grad_norm": 18.0,
      "learning_rate": 1.7432120606107224e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4741176.0,
      "step": 10893
    },
    {
      "epoch": 0.837935543419737,
      "grad_norm": 6.28125,
      "learning_rate": 1.7430966848703947e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 4741734.0,
      "step": 10894
    },
    {
      "epoch": 0.8380124605799554,
      "grad_norm": 25.375,
      "learning_rate": 1.7429813091300667e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4742046.0,
      "step": 10895
    },
    {
      "epoch": 0.8380893777401738,
      "grad_norm": 6.84375,
      "learning_rate": 1.7428659333897392e-06,
      "loss": 0.009,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 4742856.0,
      "step": 10896
    },
    {
      "epoch": 0.8381662949003923,
      "grad_norm": 15.75,
      "learning_rate": 1.7427505576494117e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4743180.0,
      "step": 10897
    },
    {
      "epoch": 0.8382432120606107,
      "grad_norm": 10.125,
      "learning_rate": 1.742635181909084e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 4743960.0,
      "step": 10898
    },
    {
      "epoch": 0.8383201292208292,
      "grad_norm": 5.46875,
      "learning_rate": 1.7425198061687564e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 4744532.0,
      "step": 10899
    },
    {
      "epoch": 0.8383970463810476,
      "grad_norm": 20.0,
      "learning_rate": 1.7424044304284289e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4744874.0,
      "step": 10900
    },
    {
      "epoch": 0.838473963541266,
      "grad_norm": 24.5,
      "learning_rate": 1.742289054688101e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 4745210.0,
      "step": 10901
    },
    {
      "epoch": 0.8385508807014845,
      "grad_norm": 7.5625,
      "learning_rate": 1.7421736789477732e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 4746030.0,
      "step": 10902
    },
    {
      "epoch": 0.8386277978617029,
      "grad_norm": 46.25,
      "learning_rate": 1.7420583032074456e-06,
      "loss": 0.035,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 4746310.0,
      "step": 10903
    },
    {
      "epoch": 0.8387047150219213,
      "grad_norm": 41.25,
      "learning_rate": 1.741942927467118e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 4746626.0,
      "step": 10904
    },
    {
      "epoch": 0.8387816321821399,
      "grad_norm": 14.9375,
      "learning_rate": 1.7418275517267904e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 4746988.0,
      "step": 10905
    },
    {
      "epoch": 0.8388585493423583,
      "grad_norm": 18.25,
      "learning_rate": 1.7417121759864624e-06,
      "loss": 0.0104,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4747572.0,
      "step": 10906
    },
    {
      "epoch": 0.8389354665025768,
      "grad_norm": 51.5,
      "learning_rate": 1.7415968002461349e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.98591548204422,
      "num_tokens": 4747858.0,
      "step": 10907
    },
    {
      "epoch": 0.8390123836627952,
      "grad_norm": 20.0,
      "learning_rate": 1.7414814245058073e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 4748162.0,
      "step": 10908
    },
    {
      "epoch": 0.8390893008230136,
      "grad_norm": 6.71875,
      "learning_rate": 1.7413660487654796e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4748684.0,
      "step": 10909
    },
    {
      "epoch": 0.8391662179832321,
      "grad_norm": 9.1875,
      "learning_rate": 1.741250673025152e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9962335228919983,
      "num_tokens": 4749216.0,
      "step": 10910
    },
    {
      "epoch": 0.8392431351434505,
      "grad_norm": 43.75,
      "learning_rate": 1.7411352972848241e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.988950252532959,
      "num_tokens": 4749580.0,
      "step": 10911
    },
    {
      "epoch": 0.8393200523036689,
      "grad_norm": 23.375,
      "learning_rate": 1.7410199215444966e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4750126.0,
      "step": 10912
    },
    {
      "epoch": 0.8393969694638874,
      "grad_norm": 5.4375,
      "learning_rate": 1.7409045458041688e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4750644.0,
      "step": 10913
    },
    {
      "epoch": 0.8394738866241058,
      "grad_norm": 37.5,
      "learning_rate": 1.7407891700638413e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 4750968.0,
      "step": 10914
    },
    {
      "epoch": 0.8395508037843242,
      "grad_norm": 39.75,
      "learning_rate": 1.7406737943235138e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4751296.0,
      "step": 10915
    },
    {
      "epoch": 0.8396277209445427,
      "grad_norm": 23.0,
      "learning_rate": 1.740558418583186e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9885057210922241,
      "num_tokens": 4751558.0,
      "step": 10916
    },
    {
      "epoch": 0.8397046381047611,
      "grad_norm": 12.8125,
      "learning_rate": 1.740443042842858e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4751880.0,
      "step": 10917
    },
    {
      "epoch": 0.8397815552649797,
      "grad_norm": 47.5,
      "learning_rate": 1.7403276671025306e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4752202.0,
      "step": 10918
    },
    {
      "epoch": 0.8398584724251981,
      "grad_norm": 16.125,
      "learning_rate": 1.740212291362203e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4752782.0,
      "step": 10919
    },
    {
      "epoch": 0.8399353895854165,
      "grad_norm": 73.5,
      "learning_rate": 1.7400969156218753e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4753122.0,
      "step": 10920
    },
    {
      "epoch": 0.840012306745635,
      "grad_norm": 30.375,
      "learning_rate": 1.7399815398815478e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 4753462.0,
      "step": 10921
    },
    {
      "epoch": 0.8400892239058534,
      "grad_norm": 46.5,
      "learning_rate": 1.7398661641412198e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4753774.0,
      "step": 10922
    },
    {
      "epoch": 0.8401661410660718,
      "grad_norm": 63.0,
      "learning_rate": 1.7397507884008923e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4754084.0,
      "step": 10923
    },
    {
      "epoch": 0.8402430582262903,
      "grad_norm": 49.5,
      "learning_rate": 1.7396354126605645e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4754388.0,
      "step": 10924
    },
    {
      "epoch": 0.8403199753865087,
      "grad_norm": 34.0,
      "learning_rate": 1.739520036920237e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 4754964.0,
      "step": 10925
    },
    {
      "epoch": 0.8403968925467272,
      "grad_norm": 20.0,
      "learning_rate": 1.7394046611799095e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 4755300.0,
      "step": 10926
    },
    {
      "epoch": 0.8404738097069456,
      "grad_norm": 39.0,
      "learning_rate": 1.7392892854395817e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4755600.0,
      "step": 10927
    },
    {
      "epoch": 0.840550726867164,
      "grad_norm": 20.375,
      "learning_rate": 1.7391739096992538e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4756146.0,
      "step": 10928
    },
    {
      "epoch": 0.8406276440273825,
      "grad_norm": 51.5,
      "learning_rate": 1.7390585339589262e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4756428.0,
      "step": 10929
    },
    {
      "epoch": 0.840704561187601,
      "grad_norm": 17.75,
      "learning_rate": 1.7389431582185987e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4757210.0,
      "step": 10930
    },
    {
      "epoch": 0.8407814783478194,
      "grad_norm": 22.0,
      "learning_rate": 1.738827782478271e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 4757790.0,
      "step": 10931
    },
    {
      "epoch": 0.8408583955080379,
      "grad_norm": 8.5625,
      "learning_rate": 1.7387124067379434e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4758312.0,
      "step": 10932
    },
    {
      "epoch": 0.8409353126682563,
      "grad_norm": 8.0625,
      "learning_rate": 1.7385970309976155e-06,
      "loss": 0.0067,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 4759098.0,
      "step": 10933
    },
    {
      "epoch": 0.8410122298284747,
      "grad_norm": 14.625,
      "learning_rate": 1.738481655257288e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4759352.0,
      "step": 10934
    },
    {
      "epoch": 0.8410891469886932,
      "grad_norm": 10.875,
      "learning_rate": 1.7383662795169602e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9905063509941101,
      "num_tokens": 4759670.0,
      "step": 10935
    },
    {
      "epoch": 0.8411660641489116,
      "grad_norm": 54.25,
      "learning_rate": 1.7382509037766327e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4759970.0,
      "step": 10936
    },
    {
      "epoch": 0.8412429813091301,
      "grad_norm": 32.25,
      "learning_rate": 1.738135528036305e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4760316.0,
      "step": 10937
    },
    {
      "epoch": 0.8413198984693485,
      "grad_norm": 53.25,
      "learning_rate": 1.7380201522959774e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4760672.0,
      "step": 10938
    },
    {
      "epoch": 0.8413968156295669,
      "grad_norm": 31.875,
      "learning_rate": 1.7379047765556495e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4760984.0,
      "step": 10939
    },
    {
      "epoch": 0.8414737327897854,
      "grad_norm": 26.375,
      "learning_rate": 1.737789400815322e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4761558.0,
      "step": 10940
    },
    {
      "epoch": 0.8415506499500038,
      "grad_norm": 56.25,
      "learning_rate": 1.7376740250749942e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9904153347015381,
      "num_tokens": 4761872.0,
      "step": 10941
    },
    {
      "epoch": 0.8416275671102222,
      "grad_norm": 41.75,
      "learning_rate": 1.7375586493346667e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4762204.0,
      "step": 10942
    },
    {
      "epoch": 0.8417044842704408,
      "grad_norm": 5.28125,
      "learning_rate": 1.7374432735943391e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 4762724.0,
      "step": 10943
    },
    {
      "epoch": 0.8417814014306592,
      "grad_norm": 29.75,
      "learning_rate": 1.7373278978540112e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9983022212982178,
      "num_tokens": 4763314.0,
      "step": 10944
    },
    {
      "epoch": 0.8418583185908777,
      "grad_norm": 40.5,
      "learning_rate": 1.7372125221136834e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 4763664.0,
      "step": 10945
    },
    {
      "epoch": 0.8419352357510961,
      "grad_norm": 4.15625,
      "learning_rate": 1.737097146373356e-06,
      "loss": 0.006,
      "mean_token_accuracy": 0.9987212419509888,
      "num_tokens": 4764448.0,
      "step": 10946
    },
    {
      "epoch": 0.8420121529113145,
      "grad_norm": 50.5,
      "learning_rate": 1.7369817706330284e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4764780.0,
      "step": 10947
    },
    {
      "epoch": 0.842089070071533,
      "grad_norm": 23.375,
      "learning_rate": 1.7368663948927006e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 4765124.0,
      "step": 10948
    },
    {
      "epoch": 0.8421659872317514,
      "grad_norm": 26.75,
      "learning_rate": 1.7367510191523727e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4765460.0,
      "step": 10949
    },
    {
      "epoch": 0.8422429043919698,
      "grad_norm": 9.875,
      "learning_rate": 1.7366356434120451e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.993914783000946,
      "num_tokens": 4765954.0,
      "step": 10950
    },
    {
      "epoch": 0.8423198215521883,
      "grad_norm": 19.0,
      "learning_rate": 1.7365202676717176e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4766208.0,
      "step": 10951
    },
    {
      "epoch": 0.8423967387124067,
      "grad_norm": 34.0,
      "learning_rate": 1.7364048919313899e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4766462.0,
      "step": 10952
    },
    {
      "epoch": 0.8424736558726252,
      "grad_norm": 14.125,
      "learning_rate": 1.7362895161910623e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4767012.0,
      "step": 10953
    },
    {
      "epoch": 0.8425505730328436,
      "grad_norm": 70.0,
      "learning_rate": 1.7361741404507348e-06,
      "loss": 0.0479,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 4767322.0,
      "step": 10954
    },
    {
      "epoch": 0.842627490193062,
      "grad_norm": 5.75,
      "learning_rate": 1.7360587647104069e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.9987212419509888,
      "num_tokens": 4768106.0,
      "step": 10955
    },
    {
      "epoch": 0.8427044073532806,
      "grad_norm": 38.0,
      "learning_rate": 1.7359433889700791e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 4768366.0,
      "step": 10956
    },
    {
      "epoch": 0.842781324513499,
      "grad_norm": 11.1875,
      "learning_rate": 1.7358280132297516e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 4768654.0,
      "step": 10957
    },
    {
      "epoch": 0.8428582416737174,
      "grad_norm": 43.75,
      "learning_rate": 1.735712637489424e-06,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 4768912.0,
      "step": 10958
    },
    {
      "epoch": 0.8429351588339359,
      "grad_norm": 55.0,
      "learning_rate": 1.7355972617490963e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 4769748.0,
      "step": 10959
    },
    {
      "epoch": 0.8430120759941543,
      "grad_norm": 14.6875,
      "learning_rate": 1.7354818860087684e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4770082.0,
      "step": 10960
    },
    {
      "epoch": 0.8430889931543727,
      "grad_norm": 53.5,
      "learning_rate": 1.7353665102684408e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4770396.0,
      "step": 10961
    },
    {
      "epoch": 0.8431659103145912,
      "grad_norm": 36.0,
      "learning_rate": 1.7352511345281133e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 4771180.0,
      "step": 10962
    },
    {
      "epoch": 0.8432428274748096,
      "grad_norm": 23.75,
      "learning_rate": 1.7351357587877856e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.992277979850769,
      "num_tokens": 4771440.0,
      "step": 10963
    },
    {
      "epoch": 0.8433197446350281,
      "grad_norm": 44.0,
      "learning_rate": 1.735020383047458e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9910233616828918,
      "num_tokens": 4771998.0,
      "step": 10964
    },
    {
      "epoch": 0.8433966617952465,
      "grad_norm": 60.5,
      "learning_rate": 1.7349050073071305e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4772340.0,
      "step": 10965
    },
    {
      "epoch": 0.8434735789554649,
      "grad_norm": 118.5,
      "learning_rate": 1.7347896315668025e-06,
      "loss": 0.0424,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 4772708.0,
      "step": 10966
    },
    {
      "epoch": 0.8435504961156834,
      "grad_norm": 10.875,
      "learning_rate": 1.7346742558264748e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4772992.0,
      "step": 10967
    },
    {
      "epoch": 0.8436274132759018,
      "grad_norm": 45.25,
      "learning_rate": 1.7345588800861473e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9949066042900085,
      "num_tokens": 4773582.0,
      "step": 10968
    },
    {
      "epoch": 0.8437043304361203,
      "grad_norm": 26.75,
      "learning_rate": 1.7344435043458197e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9902597665786743,
      "num_tokens": 4773892.0,
      "step": 10969
    },
    {
      "epoch": 0.8437812475963388,
      "grad_norm": 38.75,
      "learning_rate": 1.734328128605492e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4774170.0,
      "step": 10970
    },
    {
      "epoch": 0.8438581647565572,
      "grad_norm": 3.78125,
      "learning_rate": 1.734212752865164e-06,
      "loss": 0.0059,
      "mean_token_accuracy": 0.9987244606018066,
      "num_tokens": 4774956.0,
      "step": 10971
    },
    {
      "epoch": 0.8439350819167757,
      "grad_norm": 3.46875,
      "learning_rate": 1.7340973771248365e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.997481107711792,
      "num_tokens": 4775752.0,
      "step": 10972
    },
    {
      "epoch": 0.8440119990769941,
      "grad_norm": 39.75,
      "learning_rate": 1.733982001384509e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4776108.0,
      "step": 10973
    },
    {
      "epoch": 0.8440889162372125,
      "grad_norm": 8.375,
      "learning_rate": 1.7338666256441812e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 4776704.0,
      "step": 10974
    },
    {
      "epoch": 0.844165833397431,
      "grad_norm": 35.25,
      "learning_rate": 1.7337512499038537e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4777034.0,
      "step": 10975
    },
    {
      "epoch": 0.8442427505576494,
      "grad_norm": 19.5,
      "learning_rate": 1.7336358741635262e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4777364.0,
      "step": 10976
    },
    {
      "epoch": 0.8443196677178678,
      "grad_norm": 73.5,
      "learning_rate": 1.7335204984231982e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4777670.0,
      "step": 10977
    },
    {
      "epoch": 0.8443965848780863,
      "grad_norm": 20.5,
      "learning_rate": 1.7334051226828705e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 4778020.0,
      "step": 10978
    },
    {
      "epoch": 0.8444735020383047,
      "grad_norm": 12.9375,
      "learning_rate": 1.733289746942543e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 4778576.0,
      "step": 10979
    },
    {
      "epoch": 0.8445504191985231,
      "grad_norm": 26.125,
      "learning_rate": 1.7331743712022154e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9883381724357605,
      "num_tokens": 4778920.0,
      "step": 10980
    },
    {
      "epoch": 0.8446273363587417,
      "grad_norm": 5.4375,
      "learning_rate": 1.7330589954618877e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4779488.0,
      "step": 10981
    },
    {
      "epoch": 0.8447042535189601,
      "grad_norm": 24.125,
      "learning_rate": 1.7329436197215597e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4779808.0,
      "step": 10982
    },
    {
      "epoch": 0.8447811706791786,
      "grad_norm": 10.75,
      "learning_rate": 1.7328282439812322e-06,
      "loss": 0.0065,
      "mean_token_accuracy": 0.9973474740982056,
      "num_tokens": 4780564.0,
      "step": 10983
    },
    {
      "epoch": 0.844858087839397,
      "grad_norm": 40.25,
      "learning_rate": 1.7327128682409047e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4780846.0,
      "step": 10984
    },
    {
      "epoch": 0.8449350049996154,
      "grad_norm": 71.5,
      "learning_rate": 1.732597492500577e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4781168.0,
      "step": 10985
    },
    {
      "epoch": 0.8450119221598339,
      "grad_norm": 39.5,
      "learning_rate": 1.7324821167602494e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 4781750.0,
      "step": 10986
    },
    {
      "epoch": 0.8450888393200523,
      "grad_norm": 26.375,
      "learning_rate": 1.7323667410199215e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 4782078.0,
      "step": 10987
    },
    {
      "epoch": 0.8451657564802707,
      "grad_norm": 51.0,
      "learning_rate": 1.732251365279594e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4782394.0,
      "step": 10988
    },
    {
      "epoch": 0.8452426736404892,
      "grad_norm": 72.5,
      "learning_rate": 1.7321359895392662e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 4782748.0,
      "step": 10989
    },
    {
      "epoch": 0.8453195908007076,
      "grad_norm": 26.375,
      "learning_rate": 1.7320206137989386e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4783322.0,
      "step": 10990
    },
    {
      "epoch": 0.8453965079609261,
      "grad_norm": 54.75,
      "learning_rate": 1.731905238058611e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 4784118.0,
      "step": 10991
    },
    {
      "epoch": 0.8454734251211445,
      "grad_norm": 23.25,
      "learning_rate": 1.7317898623182834e-06,
      "loss": 0.0273,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4784404.0,
      "step": 10992
    },
    {
      "epoch": 0.8455503422813629,
      "grad_norm": 56.25,
      "learning_rate": 1.7316744865779554e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4784726.0,
      "step": 10993
    },
    {
      "epoch": 0.8456272594415815,
      "grad_norm": 17.875,
      "learning_rate": 1.7315591108376279e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 4785278.0,
      "step": 10994
    },
    {
      "epoch": 0.8457041766017999,
      "grad_norm": 21.0,
      "learning_rate": 1.7314437350973002e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9982847571372986,
      "num_tokens": 4785862.0,
      "step": 10995
    },
    {
      "epoch": 0.8457810937620183,
      "grad_norm": 32.5,
      "learning_rate": 1.7313283593569726e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 4786196.0,
      "step": 10996
    },
    {
      "epoch": 0.8458580109222368,
      "grad_norm": 13.0,
      "learning_rate": 1.731212983616645e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 4786776.0,
      "step": 10997
    },
    {
      "epoch": 0.8459349280824552,
      "grad_norm": 37.5,
      "learning_rate": 1.7310976078763171e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 4787116.0,
      "step": 10998
    },
    {
      "epoch": 0.8460118452426736,
      "grad_norm": 45.75,
      "learning_rate": 1.7309822321359894e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 4787452.0,
      "step": 10999
    },
    {
      "epoch": 0.8460887624028921,
      "grad_norm": 82.5,
      "learning_rate": 1.7308668563956619e-06,
      "loss": 0.0364,
      "step": 11000
    },
    {
      "epoch": 0.8460887624028921,
      "eval_loss": 0.019776852801442146,
      "eval_mean_token_accuracy": 0.9922760391235351,
      "eval_num_tokens": 4787726.0,
      "eval_runtime": 21.0714,
      "eval_samples_per_second": 9.492,
      "eval_steps_per_second": 1.186,
      "step": 11000
    },
    {
      "epoch": 0.8461656795631105,
      "grad_norm": 28.5,
      "learning_rate": 1.7307514806553343e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9879718124866486,
      "num_tokens": 4788046.0,
      "step": 11001
    },
    {
      "epoch": 0.846242596723329,
      "grad_norm": 64.5,
      "learning_rate": 1.7306361049150066e-06,
      "loss": 0.0358,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4788328.0,
      "step": 11002
    },
    {
      "epoch": 0.8463195138835474,
      "grad_norm": 36.5,
      "learning_rate": 1.730520729174679e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4788662.0,
      "step": 11003
    },
    {
      "epoch": 0.8463964310437658,
      "grad_norm": 11.0,
      "learning_rate": 1.7304053534343511e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 4789232.0,
      "step": 11004
    },
    {
      "epoch": 0.8464733482039843,
      "grad_norm": 31.0,
      "learning_rate": 1.7302899776940236e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 4789552.0,
      "step": 11005
    },
    {
      "epoch": 0.8465502653642027,
      "grad_norm": 29.25,
      "learning_rate": 1.7301746019536958e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4790068.0,
      "step": 11006
    },
    {
      "epoch": 0.8466271825244212,
      "grad_norm": 79.0,
      "learning_rate": 1.7300592262133683e-06,
      "loss": 0.045,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4790328.0,
      "step": 11007
    },
    {
      "epoch": 0.8467040996846397,
      "grad_norm": 37.75,
      "learning_rate": 1.7299438504730408e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4790610.0,
      "step": 11008
    },
    {
      "epoch": 0.8467810168448581,
      "grad_norm": 46.75,
      "learning_rate": 1.7298284747327128e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 4790934.0,
      "step": 11009
    },
    {
      "epoch": 0.8468579340050766,
      "grad_norm": 21.125,
      "learning_rate": 1.729713098992385e-06,
      "loss": 0.0173,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 4791478.0,
      "step": 11010
    },
    {
      "epoch": 0.846934851165295,
      "grad_norm": 20.0,
      "learning_rate": 1.7295977232520576e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4792026.0,
      "step": 11011
    },
    {
      "epoch": 0.8470117683255134,
      "grad_norm": 26.125,
      "learning_rate": 1.72948234751173e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4792372.0,
      "step": 11012
    },
    {
      "epoch": 0.8470886854857319,
      "grad_norm": 15.25,
      "learning_rate": 1.7293669717714023e-06,
      "loss": 0.0183,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4792946.0,
      "step": 11013
    },
    {
      "epoch": 0.8471656026459503,
      "grad_norm": 32.75,
      "learning_rate": 1.7292515960310743e-06,
      "loss": 0.0346,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4793252.0,
      "step": 11014
    },
    {
      "epoch": 0.8472425198061687,
      "grad_norm": 13.0,
      "learning_rate": 1.7291362202907468e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9963167309761047,
      "num_tokens": 4793796.0,
      "step": 11015
    },
    {
      "epoch": 0.8473194369663872,
      "grad_norm": 7.78125,
      "learning_rate": 1.7290208445504193e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4794338.0,
      "step": 11016
    },
    {
      "epoch": 0.8473963541266056,
      "grad_norm": 18.625,
      "learning_rate": 1.7289054688100915e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4794900.0,
      "step": 11017
    },
    {
      "epoch": 0.847473271286824,
      "grad_norm": 17.875,
      "learning_rate": 1.728790093069764e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4795452.0,
      "step": 11018
    },
    {
      "epoch": 0.8475501884470426,
      "grad_norm": 24.25,
      "learning_rate": 1.7286747173294365e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4796014.0,
      "step": 11019
    },
    {
      "epoch": 0.847627105607261,
      "grad_norm": 37.0,
      "learning_rate": 1.7285593415891085e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4796344.0,
      "step": 11020
    },
    {
      "epoch": 0.8477040227674795,
      "grad_norm": 35.0,
      "learning_rate": 1.7284439658487808e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4796654.0,
      "step": 11021
    },
    {
      "epoch": 0.8477809399276979,
      "grad_norm": 16.75,
      "learning_rate": 1.7283285901084532e-06,
      "loss": 0.0341,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4796908.0,
      "step": 11022
    },
    {
      "epoch": 0.8478578570879163,
      "grad_norm": 33.5,
      "learning_rate": 1.7282132143681257e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 4797234.0,
      "step": 11023
    },
    {
      "epoch": 0.8479347742481348,
      "grad_norm": 40.5,
      "learning_rate": 1.728097838627798e-06,
      "loss": 0.0354,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 4797492.0,
      "step": 11024
    },
    {
      "epoch": 0.8480116914083532,
      "grad_norm": 12.0625,
      "learning_rate": 1.72798246288747e-06,
      "loss": 0.016,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4797836.0,
      "step": 11025
    },
    {
      "epoch": 0.8480886085685716,
      "grad_norm": 27.875,
      "learning_rate": 1.7278670871471425e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4798188.0,
      "step": 11026
    },
    {
      "epoch": 0.8481655257287901,
      "grad_norm": 8.25,
      "learning_rate": 1.727751711406815e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4798752.0,
      "step": 11027
    },
    {
      "epoch": 0.8482424428890085,
      "grad_norm": 17.0,
      "learning_rate": 1.7276363356664872e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 4799038.0,
      "step": 11028
    },
    {
      "epoch": 0.848319360049227,
      "grad_norm": 19.5,
      "learning_rate": 1.7275209599261597e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4799376.0,
      "step": 11029
    },
    {
      "epoch": 0.8483962772094454,
      "grad_norm": 9.3125,
      "learning_rate": 1.7274055841858321e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 4799910.0,
      "step": 11030
    },
    {
      "epoch": 0.8484731943696638,
      "grad_norm": 29.25,
      "learning_rate": 1.7272902084455042e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4800222.0,
      "step": 11031
    },
    {
      "epoch": 0.8485501115298824,
      "grad_norm": 28.0,
      "learning_rate": 1.7271748327051765e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 4800518.0,
      "step": 11032
    },
    {
      "epoch": 0.8486270286901008,
      "grad_norm": 51.0,
      "learning_rate": 1.727059456964849e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 4800822.0,
      "step": 11033
    },
    {
      "epoch": 0.8487039458503192,
      "grad_norm": 10.75,
      "learning_rate": 1.7269440812245214e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4801168.0,
      "step": 11034
    },
    {
      "epoch": 0.8487808630105377,
      "grad_norm": 27.25,
      "learning_rate": 1.7268287054841937e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9961089491844177,
      "num_tokens": 4801426.0,
      "step": 11035
    },
    {
      "epoch": 0.8488577801707561,
      "grad_norm": 32.0,
      "learning_rate": 1.7267133297438657e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9860334992408752,
      "num_tokens": 4801786.0,
      "step": 11036
    },
    {
      "epoch": 0.8489346973309745,
      "grad_norm": 8.5,
      "learning_rate": 1.7265979540035382e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4802348.0,
      "step": 11037
    },
    {
      "epoch": 0.849011614491193,
      "grad_norm": 12.625,
      "learning_rate": 1.7264825782632106e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4802928.0,
      "step": 11038
    },
    {
      "epoch": 0.8490885316514114,
      "grad_norm": 28.125,
      "learning_rate": 1.726367202522883e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4803500.0,
      "step": 11039
    },
    {
      "epoch": 0.8491654488116299,
      "grad_norm": 11.0,
      "learning_rate": 1.7262518267825554e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9922330379486084,
      "num_tokens": 4804016.0,
      "step": 11040
    },
    {
      "epoch": 0.8492423659718483,
      "grad_norm": 11.1875,
      "learning_rate": 1.7261364510422276e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.996610164642334,
      "num_tokens": 4804312.0,
      "step": 11041
    },
    {
      "epoch": 0.8493192831320667,
      "grad_norm": 4.53125,
      "learning_rate": 1.7260210753018999e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 4805092.0,
      "step": 11042
    },
    {
      "epoch": 0.8493962002922852,
      "grad_norm": 10.4375,
      "learning_rate": 1.7259056995615721e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4805610.0,
      "step": 11043
    },
    {
      "epoch": 0.8494731174525036,
      "grad_norm": 44.0,
      "learning_rate": 1.7257903238212446e-06,
      "loss": 0.0326,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4805892.0,
      "step": 11044
    },
    {
      "epoch": 0.849550034612722,
      "grad_norm": 34.5,
      "learning_rate": 1.7256749480809169e-06,
      "loss": 0.0388,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4806174.0,
      "step": 11045
    },
    {
      "epoch": 0.8496269517729406,
      "grad_norm": 14.4375,
      "learning_rate": 1.7255595723405893e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4806482.0,
      "step": 11046
    },
    {
      "epoch": 0.849703868933159,
      "grad_norm": 9.0,
      "learning_rate": 1.7254441966002614e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 4807026.0,
      "step": 11047
    },
    {
      "epoch": 0.8497807860933775,
      "grad_norm": 25.75,
      "learning_rate": 1.7253288208599339e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.992277979850769,
      "num_tokens": 4807286.0,
      "step": 11048
    },
    {
      "epoch": 0.8498577032535959,
      "grad_norm": 42.5,
      "learning_rate": 1.7252134451196061e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 4807618.0,
      "step": 11049
    },
    {
      "epoch": 0.8499346204138143,
      "grad_norm": 17.0,
      "learning_rate": 1.7250980693792786e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 4807958.0,
      "step": 11050
    },
    {
      "epoch": 0.8500115375740328,
      "grad_norm": 8.9375,
      "learning_rate": 1.724982693638951e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.992277979850769,
      "num_tokens": 4808218.0,
      "step": 11051
    },
    {
      "epoch": 0.8500884547342512,
      "grad_norm": 17.75,
      "learning_rate": 1.724867317898623e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4808522.0,
      "step": 11052
    },
    {
      "epoch": 0.8501653718944696,
      "grad_norm": 11.9375,
      "learning_rate": 1.7247519421582954e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4809070.0,
      "step": 11053
    },
    {
      "epoch": 0.8502422890546881,
      "grad_norm": 10.1875,
      "learning_rate": 1.7246365664179678e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4809376.0,
      "step": 11054
    },
    {
      "epoch": 0.8503192062149065,
      "grad_norm": 10.3125,
      "learning_rate": 1.7245211906776403e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 4809684.0,
      "step": 11055
    },
    {
      "epoch": 0.8503961233751249,
      "grad_norm": 65.0,
      "learning_rate": 1.7244058149373126e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4810024.0,
      "step": 11056
    },
    {
      "epoch": 0.8504730405353434,
      "grad_norm": 55.5,
      "learning_rate": 1.724290439196985e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 4810362.0,
      "step": 11057
    },
    {
      "epoch": 0.8505499576955619,
      "grad_norm": 9.1875,
      "learning_rate": 1.724175063456657e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9959595799446106,
      "num_tokens": 4810858.0,
      "step": 11058
    },
    {
      "epoch": 0.8506268748557804,
      "grad_norm": 42.75,
      "learning_rate": 1.7240596877163295e-06,
      "loss": 0.0337,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4811180.0,
      "step": 11059
    },
    {
      "epoch": 0.8507037920159988,
      "grad_norm": 23.625,
      "learning_rate": 1.7239443119760018e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4811524.0,
      "step": 11060
    },
    {
      "epoch": 0.8507807091762172,
      "grad_norm": 13.4375,
      "learning_rate": 1.7238289362356743e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4811864.0,
      "step": 11061
    },
    {
      "epoch": 0.8508576263364357,
      "grad_norm": 13.1875,
      "learning_rate": 1.7237135604953467e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4812406.0,
      "step": 11062
    },
    {
      "epoch": 0.8509345434966541,
      "grad_norm": 17.25,
      "learning_rate": 1.7235981847550188e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9826086759567261,
      "num_tokens": 4812638.0,
      "step": 11063
    },
    {
      "epoch": 0.8510114606568725,
      "grad_norm": 4.8125,
      "learning_rate": 1.723482809014691e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9987244606018066,
      "num_tokens": 4813424.0,
      "step": 11064
    },
    {
      "epoch": 0.851088377817091,
      "grad_norm": 47.5,
      "learning_rate": 1.7233674332743635e-06,
      "loss": 0.0332,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 4813728.0,
      "step": 11065
    },
    {
      "epoch": 0.8511652949773094,
      "grad_norm": 37.25,
      "learning_rate": 1.723252057534036e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4814058.0,
      "step": 11066
    },
    {
      "epoch": 0.8512422121375279,
      "grad_norm": 10.0,
      "learning_rate": 1.7231366817937082e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4814602.0,
      "step": 11067
    },
    {
      "epoch": 0.8513191292977463,
      "grad_norm": 5.34375,
      "learning_rate": 1.7230213060533807e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 4815170.0,
      "step": 11068
    },
    {
      "epoch": 0.8513960464579647,
      "grad_norm": 55.75,
      "learning_rate": 1.7229059303130528e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9917582273483276,
      "num_tokens": 4815536.0,
      "step": 11069
    },
    {
      "epoch": 0.8514729636181833,
      "grad_norm": 7.6875,
      "learning_rate": 1.7227905545727252e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 4816110.0,
      "step": 11070
    },
    {
      "epoch": 0.8515498807784017,
      "grad_norm": 47.5,
      "learning_rate": 1.7226751788323975e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 4816476.0,
      "step": 11071
    },
    {
      "epoch": 0.8516267979386201,
      "grad_norm": 14.5625,
      "learning_rate": 1.72255980309207e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4816806.0,
      "step": 11072
    },
    {
      "epoch": 0.8517037150988386,
      "grad_norm": 16.125,
      "learning_rate": 1.7224444273517424e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4817146.0,
      "step": 11073
    },
    {
      "epoch": 0.851780632259057,
      "grad_norm": 11.5,
      "learning_rate": 1.7223290516114145e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 4817480.0,
      "step": 11074
    },
    {
      "epoch": 0.8518575494192755,
      "grad_norm": 5.71875,
      "learning_rate": 1.7222136758710867e-06,
      "loss": 0.0061,
      "mean_token_accuracy": 0.9975247383117676,
      "num_tokens": 4818290.0,
      "step": 11075
    },
    {
      "epoch": 0.8519344665794939,
      "grad_norm": 23.5,
      "learning_rate": 1.7220983001307592e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9961612224578857,
      "num_tokens": 4818812.0,
      "step": 11076
    },
    {
      "epoch": 0.8520113837397123,
      "grad_norm": 15.625,
      "learning_rate": 1.7219829243904317e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4819334.0,
      "step": 11077
    },
    {
      "epoch": 0.8520883008999308,
      "grad_norm": 19.75,
      "learning_rate": 1.721867548650104e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4819662.0,
      "step": 11078
    },
    {
      "epoch": 0.8521652180601492,
      "grad_norm": 43.0,
      "learning_rate": 1.7217521729097764e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 4820006.0,
      "step": 11079
    },
    {
      "epoch": 0.8522421352203676,
      "grad_norm": 33.5,
      "learning_rate": 1.7216367971694484e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 4820344.0,
      "step": 11080
    },
    {
      "epoch": 0.8523190523805861,
      "grad_norm": 15.4375,
      "learning_rate": 1.721521421429121e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9982847571372986,
      "num_tokens": 4820928.0,
      "step": 11081
    },
    {
      "epoch": 0.8523959695408045,
      "grad_norm": 10.3125,
      "learning_rate": 1.7214060456887932e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4821498.0,
      "step": 11082
    },
    {
      "epoch": 0.852472886701023,
      "grad_norm": 44.0,
      "learning_rate": 1.7212906699484656e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 4821838.0,
      "step": 11083
    },
    {
      "epoch": 0.8525498038612415,
      "grad_norm": 12.125,
      "learning_rate": 1.7211752942081381e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 4822374.0,
      "step": 11084
    },
    {
      "epoch": 0.8526267210214599,
      "grad_norm": 23.375,
      "learning_rate": 1.7210599184678102e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4822682.0,
      "step": 11085
    },
    {
      "epoch": 0.8527036381816784,
      "grad_norm": 23.875,
      "learning_rate": 1.7209445427274824e-06,
      "loss": 0.0325,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4822980.0,
      "step": 11086
    },
    {
      "epoch": 0.8527805553418968,
      "grad_norm": 13.875,
      "learning_rate": 1.7208291669871549e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9975429773330688,
      "num_tokens": 4823796.0,
      "step": 11087
    },
    {
      "epoch": 0.8528574725021152,
      "grad_norm": 15.5625,
      "learning_rate": 1.7207137912468274e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 4824142.0,
      "step": 11088
    },
    {
      "epoch": 0.8529343896623337,
      "grad_norm": 5.25,
      "learning_rate": 1.7205984155064996e-06,
      "loss": 0.0075,
      "mean_token_accuracy": 0.9987684488296509,
      "num_tokens": 4824956.0,
      "step": 11089
    },
    {
      "epoch": 0.8530113068225521,
      "grad_norm": 10.6875,
      "learning_rate": 1.7204830397661717e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9964221715927124,
      "num_tokens": 4825516.0,
      "step": 11090
    },
    {
      "epoch": 0.8530882239827705,
      "grad_norm": 31.25,
      "learning_rate": 1.7203676640258441e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.9885057210922241,
      "num_tokens": 4825778.0,
      "step": 11091
    },
    {
      "epoch": 0.853165141142989,
      "grad_norm": 35.75,
      "learning_rate": 1.7202522882855166e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 4826112.0,
      "step": 11092
    },
    {
      "epoch": 0.8532420583032074,
      "grad_norm": 23.0,
      "learning_rate": 1.7201369125451889e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4826432.0,
      "step": 11093
    },
    {
      "epoch": 0.8533189754634259,
      "grad_norm": 19.875,
      "learning_rate": 1.7200215368048613e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4826744.0,
      "step": 11094
    },
    {
      "epoch": 0.8533958926236443,
      "grad_norm": 10.625,
      "learning_rate": 1.7199061610645336e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4827292.0,
      "step": 11095
    },
    {
      "epoch": 0.8534728097838628,
      "grad_norm": 30.125,
      "learning_rate": 1.7197907853242058e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4827546.0,
      "step": 11096
    },
    {
      "epoch": 0.8535497269440813,
      "grad_norm": 18.125,
      "learning_rate": 1.719675409583878e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 4828080.0,
      "step": 11097
    },
    {
      "epoch": 0.8536266441042997,
      "grad_norm": 19.0,
      "learning_rate": 1.7195600338435506e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 4828876.0,
      "step": 11098
    },
    {
      "epoch": 0.8537035612645181,
      "grad_norm": 13.375,
      "learning_rate": 1.7194446581032228e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4829160.0,
      "step": 11099
    },
    {
      "epoch": 0.8537804784247366,
      "grad_norm": 25.0,
      "learning_rate": 1.7193292823628953e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4829740.0,
      "step": 11100
    },
    {
      "epoch": 0.853857395584955,
      "grad_norm": 37.75,
      "learning_rate": 1.7192139066225673e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4830000.0,
      "step": 11101
    },
    {
      "epoch": 0.8539343127451734,
      "grad_norm": 15.5,
      "learning_rate": 1.7190985308822398e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4830332.0,
      "step": 11102
    },
    {
      "epoch": 0.8540112299053919,
      "grad_norm": 19.875,
      "learning_rate": 1.718983155141912e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 4830892.0,
      "step": 11103
    },
    {
      "epoch": 0.8540881470656103,
      "grad_norm": 25.25,
      "learning_rate": 1.7188677794015845e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4831198.0,
      "step": 11104
    },
    {
      "epoch": 0.8541650642258288,
      "grad_norm": 36.25,
      "learning_rate": 1.718752403661257e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4831538.0,
      "step": 11105
    },
    {
      "epoch": 0.8542419813860472,
      "grad_norm": 37.5,
      "learning_rate": 1.7186370279209293e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4831866.0,
      "step": 11106
    },
    {
      "epoch": 0.8543188985462656,
      "grad_norm": 25.5,
      "learning_rate": 1.7185216521806013e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9934640526771545,
      "num_tokens": 4832174.0,
      "step": 11107
    },
    {
      "epoch": 0.8543958157064842,
      "grad_norm": 23.75,
      "learning_rate": 1.7184062764402738e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9932659864425659,
      "num_tokens": 4832472.0,
      "step": 11108
    },
    {
      "epoch": 0.8544727328667026,
      "grad_norm": 33.0,
      "learning_rate": 1.7182909006999463e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4832800.0,
      "step": 11109
    },
    {
      "epoch": 0.854549650026921,
      "grad_norm": 5.25,
      "learning_rate": 1.7181755249596185e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 4833324.0,
      "step": 11110
    },
    {
      "epoch": 0.8546265671871395,
      "grad_norm": 38.0,
      "learning_rate": 1.718060149219291e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 4833924.0,
      "step": 11111
    },
    {
      "epoch": 0.8547034843473579,
      "grad_norm": 9.75,
      "learning_rate": 1.717944773478963e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4834240.0,
      "step": 11112
    },
    {
      "epoch": 0.8547804015075764,
      "grad_norm": 21.875,
      "learning_rate": 1.7178293977386355e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4834548.0,
      "step": 11113
    },
    {
      "epoch": 0.8548573186677948,
      "grad_norm": 11.375,
      "learning_rate": 1.7177140219983078e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4835070.0,
      "step": 11114
    },
    {
      "epoch": 0.8549342358280132,
      "grad_norm": 53.25,
      "learning_rate": 1.7175986462579802e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4835406.0,
      "step": 11115
    },
    {
      "epoch": 0.8550111529882317,
      "grad_norm": 27.25,
      "learning_rate": 1.7174832705176527e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4835734.0,
      "step": 11116
    },
    {
      "epoch": 0.8550880701484501,
      "grad_norm": 11.5625,
      "learning_rate": 1.7173678947773247e-06,
      "loss": 0.0154,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4836280.0,
      "step": 11117
    },
    {
      "epoch": 0.8551649873086685,
      "grad_norm": 25.5,
      "learning_rate": 1.717252519036997e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4836598.0,
      "step": 11118
    },
    {
      "epoch": 0.855241904468887,
      "grad_norm": 19.625,
      "learning_rate": 1.7171371432966695e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 4837154.0,
      "step": 11119
    },
    {
      "epoch": 0.8553188216291054,
      "grad_norm": 25.375,
      "learning_rate": 1.717021767556342e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9940652847290039,
      "num_tokens": 4837492.0,
      "step": 11120
    },
    {
      "epoch": 0.8553957387893238,
      "grad_norm": 64.5,
      "learning_rate": 1.7169063918160142e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4837778.0,
      "step": 11121
    },
    {
      "epoch": 0.8554726559495424,
      "grad_norm": 13.75,
      "learning_rate": 1.7167910160756867e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.9890109896659851,
      "num_tokens": 4838052.0,
      "step": 11122
    },
    {
      "epoch": 0.8555495731097608,
      "grad_norm": 5.25,
      "learning_rate": 1.7166756403353587e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 4838594.0,
      "step": 11123
    },
    {
      "epoch": 0.8556264902699793,
      "grad_norm": 13.9375,
      "learning_rate": 1.7165602645950312e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 4839114.0,
      "step": 11124
    },
    {
      "epoch": 0.8557034074301977,
      "grad_norm": 36.0,
      "learning_rate": 1.7164448888547034e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9872204661369324,
      "num_tokens": 4839428.0,
      "step": 11125
    },
    {
      "epoch": 0.8557803245904161,
      "grad_norm": 5.40625,
      "learning_rate": 1.716329513114376e-06,
      "loss": 0.0084,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 4840208.0,
      "step": 11126
    },
    {
      "epoch": 0.8558572417506346,
      "grad_norm": 15.6875,
      "learning_rate": 1.7162141373740484e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4840526.0,
      "step": 11127
    },
    {
      "epoch": 0.855934158910853,
      "grad_norm": 10.1875,
      "learning_rate": 1.7160987616337204e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4840812.0,
      "step": 11128
    },
    {
      "epoch": 0.8560110760710714,
      "grad_norm": 19.0,
      "learning_rate": 1.7159833858933927e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4841392.0,
      "step": 11129
    },
    {
      "epoch": 0.8560879932312899,
      "grad_norm": 28.625,
      "learning_rate": 1.7158680101530652e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4841648.0,
      "step": 11130
    },
    {
      "epoch": 0.8561649103915083,
      "grad_norm": 38.75,
      "learning_rate": 1.7157526344127376e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 4841948.0,
      "step": 11131
    },
    {
      "epoch": 0.8562418275517268,
      "grad_norm": 17.0,
      "learning_rate": 1.7156372586724099e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4842266.0,
      "step": 11132
    },
    {
      "epoch": 0.8563187447119452,
      "grad_norm": 22.75,
      "learning_rate": 1.7155218829320824e-06,
      "loss": 0.028,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4842550.0,
      "step": 11133
    },
    {
      "epoch": 0.8563956618721636,
      "grad_norm": 4.1875,
      "learning_rate": 1.7154065071917544e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 4843068.0,
      "step": 11134
    },
    {
      "epoch": 0.8564725790323822,
      "grad_norm": 19.75,
      "learning_rate": 1.7152911314514269e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 4843652.0,
      "step": 11135
    },
    {
      "epoch": 0.8565494961926006,
      "grad_norm": 30.5,
      "learning_rate": 1.7151757557110991e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4843930.0,
      "step": 11136
    },
    {
      "epoch": 0.856626413352819,
      "grad_norm": 21.25,
      "learning_rate": 1.7150603799707716e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 4844280.0,
      "step": 11137
    },
    {
      "epoch": 0.8567033305130375,
      "grad_norm": 22.875,
      "learning_rate": 1.714945004230444e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4844596.0,
      "step": 11138
    },
    {
      "epoch": 0.8567802476732559,
      "grad_norm": 17.0,
      "learning_rate": 1.7148296284901161e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 4844850.0,
      "step": 11139
    },
    {
      "epoch": 0.8568571648334743,
      "grad_norm": 31.5,
      "learning_rate": 1.7147142527497884e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9856630563735962,
      "num_tokens": 4845130.0,
      "step": 11140
    },
    {
      "epoch": 0.8569340819936928,
      "grad_norm": 13.5,
      "learning_rate": 1.7145988770094608e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 4845676.0,
      "step": 11141
    },
    {
      "epoch": 0.8570109991539112,
      "grad_norm": 38.0,
      "learning_rate": 1.7144835012691333e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 4846032.0,
      "step": 11142
    },
    {
      "epoch": 0.8570879163141297,
      "grad_norm": 11.0625,
      "learning_rate": 1.7143681255288056e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9921568632125854,
      "num_tokens": 4846288.0,
      "step": 11143
    },
    {
      "epoch": 0.8571648334743481,
      "grad_norm": 7.5625,
      "learning_rate": 1.714252749788478e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9921568632125854,
      "num_tokens": 4846544.0,
      "step": 11144
    },
    {
      "epoch": 0.8572417506345665,
      "grad_norm": 21.75,
      "learning_rate": 1.71413737404815e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 4846858.0,
      "step": 11145
    },
    {
      "epoch": 0.857318667794785,
      "grad_norm": 4.71875,
      "learning_rate": 1.7140219983078226e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9981516003608704,
      "num_tokens": 4847400.0,
      "step": 11146
    },
    {
      "epoch": 0.8573955849550035,
      "grad_norm": 5.6875,
      "learning_rate": 1.7139066225674948e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 4847684.0,
      "step": 11147
    },
    {
      "epoch": 0.8574725021152219,
      "grad_norm": 56.75,
      "learning_rate": 1.7137912468271673e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9908814430236816,
      "num_tokens": 4848014.0,
      "step": 11148
    },
    {
      "epoch": 0.8575494192754404,
      "grad_norm": 8.6875,
      "learning_rate": 1.7136758710868398e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4848556.0,
      "step": 11149
    },
    {
      "epoch": 0.8576263364356588,
      "grad_norm": 5.8125,
      "learning_rate": 1.7135604953465118e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4849102.0,
      "step": 11150
    },
    {
      "epoch": 0.8577032535958773,
      "grad_norm": 41.0,
      "learning_rate": 1.713445119606184e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4849398.0,
      "step": 11151
    },
    {
      "epoch": 0.8577801707560957,
      "grad_norm": 16.875,
      "learning_rate": 1.7133297438658565e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 4849916.0,
      "step": 11152
    },
    {
      "epoch": 0.8578570879163141,
      "grad_norm": 30.0,
      "learning_rate": 1.713214368125529e-06,
      "loss": 0.0333,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4850214.0,
      "step": 11153
    },
    {
      "epoch": 0.8579340050765326,
      "grad_norm": 15.4375,
      "learning_rate": 1.7130989923852013e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9938900470733643,
      "num_tokens": 4850706.0,
      "step": 11154
    },
    {
      "epoch": 0.858010922236751,
      "grad_norm": 11.0625,
      "learning_rate": 1.7129836166448733e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 4851040.0,
      "step": 11155
    },
    {
      "epoch": 0.8580878393969694,
      "grad_norm": 4.28125,
      "learning_rate": 1.7128682409045458e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4851588.0,
      "step": 11156
    },
    {
      "epoch": 0.8581647565571879,
      "grad_norm": 23.625,
      "learning_rate": 1.7127528651642182e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9873417615890503,
      "num_tokens": 4851906.0,
      "step": 11157
    },
    {
      "epoch": 0.8582416737174063,
      "grad_norm": 16.25,
      "learning_rate": 1.7126374894238905e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4852452.0,
      "step": 11158
    },
    {
      "epoch": 0.8583185908776247,
      "grad_norm": 23.875,
      "learning_rate": 1.712522113683563e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 4853036.0,
      "step": 11159
    },
    {
      "epoch": 0.8583955080378433,
      "grad_norm": 36.5,
      "learning_rate": 1.7124067379432352e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4853320.0,
      "step": 11160
    },
    {
      "epoch": 0.8584724251980617,
      "grad_norm": 11.125,
      "learning_rate": 1.7122913622029075e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4853894.0,
      "step": 11161
    },
    {
      "epoch": 0.8585493423582802,
      "grad_norm": 54.5,
      "learning_rate": 1.7121759864625797e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.983818769454956,
      "num_tokens": 4854204.0,
      "step": 11162
    },
    {
      "epoch": 0.8586262595184986,
      "grad_norm": 38.25,
      "learning_rate": 1.7120606107222522e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 4854800.0,
      "step": 11163
    },
    {
      "epoch": 0.858703176678717,
      "grad_norm": 50.5,
      "learning_rate": 1.7119452349819245e-06,
      "loss": 0.0439,
      "mean_token_accuracy": 0.9805194735527039,
      "num_tokens": 4855110.0,
      "step": 11164
    },
    {
      "epoch": 0.8587800938389355,
      "grad_norm": 15.9375,
      "learning_rate": 1.711829859241597e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9959431886672974,
      "num_tokens": 4855604.0,
      "step": 11165
    },
    {
      "epoch": 0.8588570109991539,
      "grad_norm": 11.0625,
      "learning_rate": 1.711714483501269e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 4855890.0,
      "step": 11166
    },
    {
      "epoch": 0.8589339281593723,
      "grad_norm": 30.25,
      "learning_rate": 1.7115991077609415e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 4856244.0,
      "step": 11167
    },
    {
      "epoch": 0.8590108453195908,
      "grad_norm": 17.5,
      "learning_rate": 1.7114837320206137e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.989130437374115,
      "num_tokens": 4856522.0,
      "step": 11168
    },
    {
      "epoch": 0.8590877624798092,
      "grad_norm": 11.1875,
      "learning_rate": 1.7113683562802862e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 4856808.0,
      "step": 11169
    },
    {
      "epoch": 0.8591646796400277,
      "grad_norm": 30.25,
      "learning_rate": 1.7112529805399587e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4857128.0,
      "step": 11170
    },
    {
      "epoch": 0.8592415968002461,
      "grad_norm": 35.25,
      "learning_rate": 1.711137604799631e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 4857686.0,
      "step": 11171
    },
    {
      "epoch": 0.8593185139604645,
      "grad_norm": 21.25,
      "learning_rate": 1.711022229059303e-06,
      "loss": 0.0385,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 4857966.0,
      "step": 11172
    },
    {
      "epoch": 0.8593954311206831,
      "grad_norm": 20.0,
      "learning_rate": 1.7109068533189754e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 4858250.0,
      "step": 11173
    },
    {
      "epoch": 0.8594723482809015,
      "grad_norm": 30.0,
      "learning_rate": 1.710791477578648e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9961832165718079,
      "num_tokens": 4859038.0,
      "step": 11174
    },
    {
      "epoch": 0.8595492654411199,
      "grad_norm": 22.625,
      "learning_rate": 1.7106761018383202e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9868995547294617,
      "num_tokens": 4859268.0,
      "step": 11175
    },
    {
      "epoch": 0.8596261826013384,
      "grad_norm": 21.625,
      "learning_rate": 1.7105607260979926e-06,
      "loss": 0.0318,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4859550.0,
      "step": 11176
    },
    {
      "epoch": 0.8597030997615568,
      "grad_norm": 32.25,
      "learning_rate": 1.7104453503576647e-06,
      "loss": 0.045,
      "mean_token_accuracy": 0.9801980257034302,
      "num_tokens": 4859854.0,
      "step": 11177
    },
    {
      "epoch": 0.8597800169217753,
      "grad_norm": 16.625,
      "learning_rate": 1.7103299746173371e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4860136.0,
      "step": 11178
    },
    {
      "epoch": 0.8598569340819937,
      "grad_norm": 33.25,
      "learning_rate": 1.7102145988770094e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 4860420.0,
      "step": 11179
    },
    {
      "epoch": 0.8599338512422121,
      "grad_norm": 17.0,
      "learning_rate": 1.7100992231366819e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 4860724.0,
      "step": 11180
    },
    {
      "epoch": 0.8600107684024306,
      "grad_norm": 38.5,
      "learning_rate": 1.7099838473963543e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4861024.0,
      "step": 11181
    },
    {
      "epoch": 0.860087685562649,
      "grad_norm": 38.75,
      "learning_rate": 1.7098684716560266e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4861322.0,
      "step": 11182
    },
    {
      "epoch": 0.8601646027228674,
      "grad_norm": 42.5,
      "learning_rate": 1.7097530959156987e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9831649661064148,
      "num_tokens": 4861622.0,
      "step": 11183
    },
    {
      "epoch": 0.860241519883086,
      "grad_norm": 72.0,
      "learning_rate": 1.7096377201753711e-06,
      "loss": 0.0365,
      "mean_token_accuracy": 0.9858757257461548,
      "num_tokens": 4861978.0,
      "step": 11184
    },
    {
      "epoch": 0.8603184370433044,
      "grad_norm": 29.25,
      "learning_rate": 1.7095223444350436e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4862786.0,
      "step": 11185
    },
    {
      "epoch": 0.8603953542035228,
      "grad_norm": 15.5625,
      "learning_rate": 1.7094069686947158e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4863070.0,
      "step": 11186
    },
    {
      "epoch": 0.8604722713637413,
      "grad_norm": 22.0,
      "learning_rate": 1.7092915929543883e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 4864114.0,
      "step": 11187
    },
    {
      "epoch": 0.8605491885239597,
      "grad_norm": 40.5,
      "learning_rate": 1.7091762172140604e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9933554530143738,
      "num_tokens": 4864416.0,
      "step": 11188
    },
    {
      "epoch": 0.8606261056841782,
      "grad_norm": 24.5,
      "learning_rate": 1.7090608414737328e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4864738.0,
      "step": 11189
    },
    {
      "epoch": 0.8607030228443966,
      "grad_norm": 20.125,
      "learning_rate": 1.708945465733405e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4865068.0,
      "step": 11190
    },
    {
      "epoch": 0.860779940004615,
      "grad_norm": 42.75,
      "learning_rate": 1.7088300899930776e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9886040091514587,
      "num_tokens": 4865420.0,
      "step": 11191
    },
    {
      "epoch": 0.8608568571648335,
      "grad_norm": 21.75,
      "learning_rate": 1.70871471425275e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4865772.0,
      "step": 11192
    },
    {
      "epoch": 0.8609337743250519,
      "grad_norm": 15.9375,
      "learning_rate": 1.708599338512422e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4866294.0,
      "step": 11193
    },
    {
      "epoch": 0.8610106914852703,
      "grad_norm": 61.25,
      "learning_rate": 1.7084839627720943e-06,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9806950092315674,
      "num_tokens": 4866554.0,
      "step": 11194
    },
    {
      "epoch": 0.8610876086454888,
      "grad_norm": 19.875,
      "learning_rate": 1.7083685870317668e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 4866844.0,
      "step": 11195
    },
    {
      "epoch": 0.8611645258057072,
      "grad_norm": 14.25,
      "learning_rate": 1.7082532112914393e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4867390.0,
      "step": 11196
    },
    {
      "epoch": 0.8612414429659258,
      "grad_norm": 3.421875,
      "learning_rate": 1.7081378355511115e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4867974.0,
      "step": 11197
    },
    {
      "epoch": 0.8613183601261442,
      "grad_norm": 29.75,
      "learning_rate": 1.708022459810784e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4868268.0,
      "step": 11198
    },
    {
      "epoch": 0.8613952772863626,
      "grad_norm": 4.90625,
      "learning_rate": 1.707907084070456e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4868820.0,
      "step": 11199
    },
    {
      "epoch": 0.8614721944465811,
      "grad_norm": 29.75,
      "learning_rate": 1.7077917083301285e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9889298677444458,
      "num_tokens": 4869092.0,
      "step": 11200
    },
    {
      "epoch": 0.8615491116067995,
      "grad_norm": 20.25,
      "learning_rate": 1.7076763325898008e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 4869408.0,
      "step": 11201
    },
    {
      "epoch": 0.8616260287670179,
      "grad_norm": 36.5,
      "learning_rate": 1.7075609568494732e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 4869692.0,
      "step": 11202
    },
    {
      "epoch": 0.8617029459272364,
      "grad_norm": 8.625,
      "learning_rate": 1.7074455811091457e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.996632993221283,
      "num_tokens": 4869990.0,
      "step": 11203
    },
    {
      "epoch": 0.8617798630874548,
      "grad_norm": 19.0,
      "learning_rate": 1.7073302053688178e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4870296.0,
      "step": 11204
    },
    {
      "epoch": 0.8618567802476732,
      "grad_norm": 13.1875,
      "learning_rate": 1.70721482962849e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4870606.0,
      "step": 11205
    },
    {
      "epoch": 0.8619336974078917,
      "grad_norm": 20.5,
      "learning_rate": 1.7070994538881625e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 4871366.0,
      "step": 11206
    },
    {
      "epoch": 0.8620106145681101,
      "grad_norm": 32.0,
      "learning_rate": 1.706984078147835e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 4871688.0,
      "step": 11207
    },
    {
      "epoch": 0.8620875317283286,
      "grad_norm": 41.75,
      "learning_rate": 1.7068687024075072e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4872034.0,
      "step": 11208
    },
    {
      "epoch": 0.862164448888547,
      "grad_norm": 21.875,
      "learning_rate": 1.7067533266671797e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4872362.0,
      "step": 11209
    },
    {
      "epoch": 0.8622413660487654,
      "grad_norm": 13.6875,
      "learning_rate": 1.7066379509268517e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4872908.0,
      "step": 11210
    },
    {
      "epoch": 0.862318283208984,
      "grad_norm": 14.4375,
      "learning_rate": 1.7065225751865242e-06,
      "loss": 0.0263,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 4873162.0,
      "step": 11211
    },
    {
      "epoch": 0.8623952003692024,
      "grad_norm": 33.0,
      "learning_rate": 1.7064071994461965e-06,
      "loss": 0.0323,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4873504.0,
      "step": 11212
    },
    {
      "epoch": 0.8624721175294208,
      "grad_norm": 16.625,
      "learning_rate": 1.706291823705869e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9929078221321106,
      "num_tokens": 4873788.0,
      "step": 11213
    },
    {
      "epoch": 0.8625490346896393,
      "grad_norm": 31.875,
      "learning_rate": 1.7061764479655412e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 4874082.0,
      "step": 11214
    },
    {
      "epoch": 0.8626259518498577,
      "grad_norm": 20.375,
      "learning_rate": 1.7060610722252135e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9930434823036194,
      "num_tokens": 4874658.0,
      "step": 11215
    },
    {
      "epoch": 0.8627028690100762,
      "grad_norm": 20.25,
      "learning_rate": 1.7059456964848857e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4874958.0,
      "step": 11216
    },
    {
      "epoch": 0.8627797861702946,
      "grad_norm": 37.5,
      "learning_rate": 1.7058303207445582e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 4875258.0,
      "step": 11217
    },
    {
      "epoch": 0.862856703330513,
      "grad_norm": 13.5,
      "learning_rate": 1.7057149450042304e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9963570237159729,
      "num_tokens": 4875808.0,
      "step": 11218
    },
    {
      "epoch": 0.8629336204907315,
      "grad_norm": 6.15625,
      "learning_rate": 1.705599569263903e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4876090.0,
      "step": 11219
    },
    {
      "epoch": 0.8630105376509499,
      "grad_norm": 43.75,
      "learning_rate": 1.7054841935235754e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 4876462.0,
      "step": 11220
    },
    {
      "epoch": 0.8630874548111683,
      "grad_norm": 20.0,
      "learning_rate": 1.7053688177832474e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 4876760.0,
      "step": 11221
    },
    {
      "epoch": 0.8631643719713868,
      "grad_norm": 8.125,
      "learning_rate": 1.7052534420429197e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4877330.0,
      "step": 11222
    },
    {
      "epoch": 0.8632412891316052,
      "grad_norm": 6.5625,
      "learning_rate": 1.7051380663025922e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4877870.0,
      "step": 11223
    },
    {
      "epoch": 0.8633182062918237,
      "grad_norm": 11.9375,
      "learning_rate": 1.7050226905622646e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 4878200.0,
      "step": 11224
    },
    {
      "epoch": 0.8633951234520422,
      "grad_norm": 8.0625,
      "learning_rate": 1.7049073148219369e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4878522.0,
      "step": 11225
    },
    {
      "epoch": 0.8634720406122606,
      "grad_norm": 8.0625,
      "learning_rate": 1.704791939081609e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 4879016.0,
      "step": 11226
    },
    {
      "epoch": 0.8635489577724791,
      "grad_norm": 12.0625,
      "learning_rate": 1.7046765633412814e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4879372.0,
      "step": 11227
    },
    {
      "epoch": 0.8636258749326975,
      "grad_norm": 19.125,
      "learning_rate": 1.7045611876009539e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4879922.0,
      "step": 11228
    },
    {
      "epoch": 0.8637027920929159,
      "grad_norm": 8.625,
      "learning_rate": 1.7044458118606261e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4880444.0,
      "step": 11229
    },
    {
      "epoch": 0.8637797092531344,
      "grad_norm": 8.875,
      "learning_rate": 1.7043304361202986e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4880784.0,
      "step": 11230
    },
    {
      "epoch": 0.8638566264133528,
      "grad_norm": 38.25,
      "learning_rate": 1.7042150603799706e-06,
      "loss": 0.0399,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 4881044.0,
      "step": 11231
    },
    {
      "epoch": 0.8639335435735712,
      "grad_norm": 15.4375,
      "learning_rate": 1.7040996846396431e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 4881336.0,
      "step": 11232
    },
    {
      "epoch": 0.8640104607337897,
      "grad_norm": 8.9375,
      "learning_rate": 1.7039843088993154e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 4881632.0,
      "step": 11233
    },
    {
      "epoch": 0.8640873778940081,
      "grad_norm": 6.65625,
      "learning_rate": 1.7038689331589878e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4881950.0,
      "step": 11234
    },
    {
      "epoch": 0.8641642950542266,
      "grad_norm": 14.5,
      "learning_rate": 1.7037535574186603e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 4882212.0,
      "step": 11235
    },
    {
      "epoch": 0.864241212214445,
      "grad_norm": 13.625,
      "learning_rate": 1.7036381816783326e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4882552.0,
      "step": 11236
    },
    {
      "epoch": 0.8643181293746635,
      "grad_norm": 7.6875,
      "learning_rate": 1.7035228059380046e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 4882892.0,
      "step": 11237
    },
    {
      "epoch": 0.864395046534882,
      "grad_norm": 8.5625,
      "learning_rate": 1.703407430197677e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9950371980667114,
      "num_tokens": 4883700.0,
      "step": 11238
    },
    {
      "epoch": 0.8644719636951004,
      "grad_norm": 6.53125,
      "learning_rate": 1.7032920544573496e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 4884248.0,
      "step": 11239
    },
    {
      "epoch": 0.8645488808553188,
      "grad_norm": 22.25,
      "learning_rate": 1.7031766787170218e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4884600.0,
      "step": 11240
    },
    {
      "epoch": 0.8646257980155373,
      "grad_norm": 12.8125,
      "learning_rate": 1.7030613029766943e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4884908.0,
      "step": 11241
    },
    {
      "epoch": 0.8647027151757557,
      "grad_norm": 18.125,
      "learning_rate": 1.7029459272363663e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9951338171958923,
      "num_tokens": 4885732.0,
      "step": 11242
    },
    {
      "epoch": 0.8647796323359741,
      "grad_norm": 10.6875,
      "learning_rate": 1.7028305514960388e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 4886302.0,
      "step": 11243
    },
    {
      "epoch": 0.8648565494961926,
      "grad_norm": 20.25,
      "learning_rate": 1.702715175755711e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4886612.0,
      "step": 11244
    },
    {
      "epoch": 0.864933466656411,
      "grad_norm": 32.0,
      "learning_rate": 1.7025998000153835e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4886918.0,
      "step": 11245
    },
    {
      "epoch": 0.8650103838166295,
      "grad_norm": 26.375,
      "learning_rate": 1.702484424275056e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 4887274.0,
      "step": 11246
    },
    {
      "epoch": 0.8650873009768479,
      "grad_norm": 26.5,
      "learning_rate": 1.7023690485347283e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 4887622.0,
      "step": 11247
    },
    {
      "epoch": 0.8651642181370663,
      "grad_norm": 23.5,
      "learning_rate": 1.7022536727944003e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4887932.0,
      "step": 11248
    },
    {
      "epoch": 0.8652411352972849,
      "grad_norm": 8.75,
      "learning_rate": 1.7021382970540728e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4888246.0,
      "step": 11249
    },
    {
      "epoch": 0.8653180524575033,
      "grad_norm": 7.5,
      "learning_rate": 1.7020229213137452e-06,
      "loss": 0.0175,
      "mean_token_accuracy": 0.9919517040252686,
      "num_tokens": 4888744.0,
      "step": 11250
    },
    {
      "epoch": 0.8653949696177217,
      "grad_norm": 19.25,
      "learning_rate": 1.7019075455734175e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 4889050.0,
      "step": 11251
    },
    {
      "epoch": 0.8654718867779402,
      "grad_norm": 32.25,
      "learning_rate": 1.70179216983309e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4889612.0,
      "step": 11252
    },
    {
      "epoch": 0.8655488039381586,
      "grad_norm": 17.875,
      "learning_rate": 1.701676794092762e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 4889896.0,
      "step": 11253
    },
    {
      "epoch": 0.8656257210983771,
      "grad_norm": 14.9375,
      "learning_rate": 1.7015614183524345e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4890218.0,
      "step": 11254
    },
    {
      "epoch": 0.8657026382585955,
      "grad_norm": 20.75,
      "learning_rate": 1.7014460426121067e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4890536.0,
      "step": 11255
    },
    {
      "epoch": 0.8657795554188139,
      "grad_norm": 7.5,
      "learning_rate": 1.7013306668717792e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 4890826.0,
      "step": 11256
    },
    {
      "epoch": 0.8658564725790324,
      "grad_norm": 6.21875,
      "learning_rate": 1.7012152911314517e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4891138.0,
      "step": 11257
    },
    {
      "epoch": 0.8659333897392508,
      "grad_norm": 10.75,
      "learning_rate": 1.7010999153911237e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4891446.0,
      "step": 11258
    },
    {
      "epoch": 0.8660103068994692,
      "grad_norm": 67.0,
      "learning_rate": 1.700984539650796e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4891758.0,
      "step": 11259
    },
    {
      "epoch": 0.8660872240596877,
      "grad_norm": 19.625,
      "learning_rate": 1.7008691639104685e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 4892046.0,
      "step": 11260
    },
    {
      "epoch": 0.8661641412199061,
      "grad_norm": 3.515625,
      "learning_rate": 1.700753788170141e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4892614.0,
      "step": 11261
    },
    {
      "epoch": 0.8662410583801246,
      "grad_norm": 6.5,
      "learning_rate": 1.7006384124298132e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4892926.0,
      "step": 11262
    },
    {
      "epoch": 0.8663179755403431,
      "grad_norm": 13.0,
      "learning_rate": 1.7005230366894857e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 4893722.0,
      "step": 11263
    },
    {
      "epoch": 0.8663948927005615,
      "grad_norm": 7.5625,
      "learning_rate": 1.7004076609491577e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 4894216.0,
      "step": 11264
    },
    {
      "epoch": 0.86647180986078,
      "grad_norm": 9.8125,
      "learning_rate": 1.7002922852088302e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4894762.0,
      "step": 11265
    },
    {
      "epoch": 0.8665487270209984,
      "grad_norm": 23.125,
      "learning_rate": 1.7001769094685024e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4895018.0,
      "step": 11266
    },
    {
      "epoch": 0.8666256441812168,
      "grad_norm": 5.5625,
      "learning_rate": 1.700061533728175e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4895568.0,
      "step": 11267
    },
    {
      "epoch": 0.8667025613414353,
      "grad_norm": 10.75,
      "learning_rate": 1.6999461579878472e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 4895928.0,
      "step": 11268
    },
    {
      "epoch": 0.8667794785016537,
      "grad_norm": 9.125,
      "learning_rate": 1.6998307822475194e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 4896262.0,
      "step": 11269
    },
    {
      "epoch": 0.8668563956618721,
      "grad_norm": 11.25,
      "learning_rate": 1.6997154065071917e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9936708807945251,
      "num_tokens": 4896580.0,
      "step": 11270
    },
    {
      "epoch": 0.8669333128220906,
      "grad_norm": 6.09375,
      "learning_rate": 1.6996000307668641e-06,
      "loss": 0.0113,
      "mean_token_accuracy": 0.9948006868362427,
      "num_tokens": 4897158.0,
      "step": 11271
    },
    {
      "epoch": 0.867010229982309,
      "grad_norm": 22.5,
      "learning_rate": 1.6994846550265364e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9940119981765747,
      "num_tokens": 4897494.0,
      "step": 11272
    },
    {
      "epoch": 0.8670871471425275,
      "grad_norm": 14.125,
      "learning_rate": 1.6993692792862089e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 4898304.0,
      "step": 11273
    },
    {
      "epoch": 0.867164064302746,
      "grad_norm": 34.75,
      "learning_rate": 1.6992539035458813e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 4898580.0,
      "step": 11274
    },
    {
      "epoch": 0.8672409814629644,
      "grad_norm": 11.375,
      "learning_rate": 1.6991385278055534e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 4898914.0,
      "step": 11275
    },
    {
      "epoch": 0.8673178986231829,
      "grad_norm": 7.78125,
      "learning_rate": 1.6990231520652256e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4899236.0,
      "step": 11276
    },
    {
      "epoch": 0.8673948157834013,
      "grad_norm": 26.75,
      "learning_rate": 1.6989077763248981e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 4899820.0,
      "step": 11277
    },
    {
      "epoch": 0.8674717329436197,
      "grad_norm": 31.5,
      "learning_rate": 1.6987924005845706e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 4900152.0,
      "step": 11278
    },
    {
      "epoch": 0.8675486501038382,
      "grad_norm": 4.65625,
      "learning_rate": 1.6986770248442428e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4900696.0,
      "step": 11279
    },
    {
      "epoch": 0.8676255672640566,
      "grad_norm": 9.0,
      "learning_rate": 1.6985616491039149e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4901264.0,
      "step": 11280
    },
    {
      "epoch": 0.8677024844242751,
      "grad_norm": 4.84375,
      "learning_rate": 1.6984462733635874e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4901808.0,
      "step": 11281
    },
    {
      "epoch": 0.8677794015844935,
      "grad_norm": 27.5,
      "learning_rate": 1.6983308976232598e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4902090.0,
      "step": 11282
    },
    {
      "epoch": 0.8678563187447119,
      "grad_norm": 16.25,
      "learning_rate": 1.698215521882932e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9929824471473694,
      "num_tokens": 4902376.0,
      "step": 11283
    },
    {
      "epoch": 0.8679332359049304,
      "grad_norm": 16.125,
      "learning_rate": 1.6981001461426046e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9967213273048401,
      "num_tokens": 4902682.0,
      "step": 11284
    },
    {
      "epoch": 0.8680101530651488,
      "grad_norm": 9.3125,
      "learning_rate": 1.697984770402277e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4903228.0,
      "step": 11285
    },
    {
      "epoch": 0.8680870702253672,
      "grad_norm": 8.875,
      "learning_rate": 1.697869394661949e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9935275316238403,
      "num_tokens": 4903538.0,
      "step": 11286
    },
    {
      "epoch": 0.8681639873855858,
      "grad_norm": 12.375,
      "learning_rate": 1.6977540189216213e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.99303138256073,
      "num_tokens": 4903826.0,
      "step": 11287
    },
    {
      "epoch": 0.8682409045458042,
      "grad_norm": 9.1875,
      "learning_rate": 1.6976386431812938e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4904108.0,
      "step": 11288
    },
    {
      "epoch": 0.8683178217060226,
      "grad_norm": 21.125,
      "learning_rate": 1.6975232674409663e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4904670.0,
      "step": 11289
    },
    {
      "epoch": 0.8683947388662411,
      "grad_norm": 57.0,
      "learning_rate": 1.6974078917006385e-06,
      "loss": 0.0303,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 4904996.0,
      "step": 11290
    },
    {
      "epoch": 0.8684716560264595,
      "grad_norm": 30.5,
      "learning_rate": 1.6972925159603106e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.9971510171890259,
      "num_tokens": 4905348.0,
      "step": 11291
    },
    {
      "epoch": 0.868548573186678,
      "grad_norm": 13.3125,
      "learning_rate": 1.697177140219983e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4905684.0,
      "step": 11292
    },
    {
      "epoch": 0.8686254903468964,
      "grad_norm": 25.375,
      "learning_rate": 1.6970617644796555e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 4906254.0,
      "step": 11293
    },
    {
      "epoch": 0.8687024075071148,
      "grad_norm": 27.125,
      "learning_rate": 1.6969463887393278e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 4907034.0,
      "step": 11294
    },
    {
      "epoch": 0.8687793246673333,
      "grad_norm": 20.375,
      "learning_rate": 1.6968310129990002e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 4907354.0,
      "step": 11295
    },
    {
      "epoch": 0.8688562418275517,
      "grad_norm": 41.25,
      "learning_rate": 1.6967156372586723e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 4907666.0,
      "step": 11296
    },
    {
      "epoch": 0.8689331589877701,
      "grad_norm": 28.0,
      "learning_rate": 1.6966002615183448e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4907986.0,
      "step": 11297
    },
    {
      "epoch": 0.8690100761479886,
      "grad_norm": 32.5,
      "learning_rate": 1.696484885778017e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 4908294.0,
      "step": 11298
    },
    {
      "epoch": 0.869086993308207,
      "grad_norm": 42.75,
      "learning_rate": 1.6963695100376895e-06,
      "loss": 0.0297,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4908622.0,
      "step": 11299
    },
    {
      "epoch": 0.8691639104684256,
      "grad_norm": 44.5,
      "learning_rate": 1.696254134297362e-06,
      "loss": 0.0313,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 4908926.0,
      "step": 11300
    },
    {
      "epoch": 0.869240827628644,
      "grad_norm": 29.125,
      "learning_rate": 1.6961387585570342e-06,
      "loss": 0.0304,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4909220.0,
      "step": 11301
    },
    {
      "epoch": 0.8693177447888624,
      "grad_norm": 28.5,
      "learning_rate": 1.6960233828167063e-06,
      "loss": 0.0186,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 4909558.0,
      "step": 11302
    },
    {
      "epoch": 0.8693946619490809,
      "grad_norm": 23.0,
      "learning_rate": 1.6959080070763787e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4910144.0,
      "step": 11303
    },
    {
      "epoch": 0.8694715791092993,
      "grad_norm": 13.4375,
      "learning_rate": 1.6957926313360512e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4910690.0,
      "step": 11304
    },
    {
      "epoch": 0.8695484962695177,
      "grad_norm": 28.0,
      "learning_rate": 1.6956772555957235e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4911006.0,
      "step": 11305
    },
    {
      "epoch": 0.8696254134297362,
      "grad_norm": 27.125,
      "learning_rate": 1.695561879855396e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 4911586.0,
      "step": 11306
    },
    {
      "epoch": 0.8697023305899546,
      "grad_norm": 18.875,
      "learning_rate": 1.695446504115068e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4912138.0,
      "step": 11307
    },
    {
      "epoch": 0.869779247750173,
      "grad_norm": 39.75,
      "learning_rate": 1.6953311283747404e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4912414.0,
      "step": 11308
    },
    {
      "epoch": 0.8698561649103915,
      "grad_norm": 43.0,
      "learning_rate": 1.6952157526344127e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9938837885856628,
      "num_tokens": 4912742.0,
      "step": 11309
    },
    {
      "epoch": 0.8699330820706099,
      "grad_norm": 36.25,
      "learning_rate": 1.6951003768940852e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 4913088.0,
      "step": 11310
    },
    {
      "epoch": 0.8700099992308284,
      "grad_norm": 32.0,
      "learning_rate": 1.6949850011537576e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.997183084487915,
      "num_tokens": 4913444.0,
      "step": 11311
    },
    {
      "epoch": 0.8700869163910468,
      "grad_norm": 23.5,
      "learning_rate": 1.69486962541343e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9856114983558655,
      "num_tokens": 4913724.0,
      "step": 11312
    },
    {
      "epoch": 0.8701638335512653,
      "grad_norm": 21.25,
      "learning_rate": 1.694754249673102e-06,
      "loss": 0.0188,
      "mean_token_accuracy": 0.9884726405143738,
      "num_tokens": 4914074.0,
      "step": 11313
    },
    {
      "epoch": 0.8702407507114838,
      "grad_norm": 11.875,
      "learning_rate": 1.6946388739327744e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9972899556159973,
      "num_tokens": 4914446.0,
      "step": 11314
    },
    {
      "epoch": 0.8703176678717022,
      "grad_norm": 20.75,
      "learning_rate": 1.6945234981924469e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4914994.0,
      "step": 11315
    },
    {
      "epoch": 0.8703945850319206,
      "grad_norm": 22.375,
      "learning_rate": 1.6944081224521191e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4915278.0,
      "step": 11316
    },
    {
      "epoch": 0.8704715021921391,
      "grad_norm": 18.125,
      "learning_rate": 1.6942927467117916e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4915600.0,
      "step": 11317
    },
    {
      "epoch": 0.8705484193523575,
      "grad_norm": 5.15625,
      "learning_rate": 1.6941773709714637e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 4916120.0,
      "step": 11318
    },
    {
      "epoch": 0.870625336512576,
      "grad_norm": 19.125,
      "learning_rate": 1.6940619952311361e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4916706.0,
      "step": 11319
    },
    {
      "epoch": 0.8707022536727944,
      "grad_norm": 10.125,
      "learning_rate": 1.6939466194908084e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 4917030.0,
      "step": 11320
    },
    {
      "epoch": 0.8707791708330128,
      "grad_norm": 4.8125,
      "learning_rate": 1.6938312437504809e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 4917790.0,
      "step": 11321
    },
    {
      "epoch": 0.8708560879932313,
      "grad_norm": 13.1875,
      "learning_rate": 1.6937158680101531e-06,
      "loss": 0.033,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4918098.0,
      "step": 11322
    },
    {
      "epoch": 0.8709330051534497,
      "grad_norm": 22.625,
      "learning_rate": 1.6936004922698256e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9970845580101013,
      "num_tokens": 4918442.0,
      "step": 11323
    },
    {
      "epoch": 0.8710099223136681,
      "grad_norm": 30.875,
      "learning_rate": 1.6934851165294976e-06,
      "loss": 0.0277,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4918740.0,
      "step": 11324
    },
    {
      "epoch": 0.8710868394738867,
      "grad_norm": 19.125,
      "learning_rate": 1.69336974078917e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 4919050.0,
      "step": 11325
    },
    {
      "epoch": 0.8711637566341051,
      "grad_norm": 10.375,
      "learning_rate": 1.6932543650488424e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.994350254535675,
      "num_tokens": 4919406.0,
      "step": 11326
    },
    {
      "epoch": 0.8712406737943235,
      "grad_norm": 12.875,
      "learning_rate": 1.6931389893085148e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4919950.0,
      "step": 11327
    },
    {
      "epoch": 0.871317590954542,
      "grad_norm": 11.875,
      "learning_rate": 1.6930236135681873e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4920512.0,
      "step": 11328
    },
    {
      "epoch": 0.8713945081147604,
      "grad_norm": 10.5625,
      "learning_rate": 1.6929082378278593e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4921052.0,
      "step": 11329
    },
    {
      "epoch": 0.8714714252749789,
      "grad_norm": 4.3125,
      "learning_rate": 1.6927928620875316e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 4921832.0,
      "step": 11330
    },
    {
      "epoch": 0.8715483424351973,
      "grad_norm": 13.25,
      "learning_rate": 1.692677486347204e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4922110.0,
      "step": 11331
    },
    {
      "epoch": 0.8716252595954157,
      "grad_norm": 11.75,
      "learning_rate": 1.6925621106068765e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9965397715568542,
      "num_tokens": 4922400.0,
      "step": 11332
    },
    {
      "epoch": 0.8717021767556342,
      "grad_norm": 52.5,
      "learning_rate": 1.6924467348665488e-06,
      "loss": 0.0232,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 4922746.0,
      "step": 11333
    },
    {
      "epoch": 0.8717790939158526,
      "grad_norm": 29.75,
      "learning_rate": 1.6923313591262208e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9967426657676697,
      "num_tokens": 4923054.0,
      "step": 11334
    },
    {
      "epoch": 0.871856011076071,
      "grad_norm": 15.8125,
      "learning_rate": 1.6922159833858933e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4923572.0,
      "step": 11335
    },
    {
      "epoch": 0.8719329282362895,
      "grad_norm": 43.25,
      "learning_rate": 1.6921006076455658e-06,
      "loss": 0.0327,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 4923902.0,
      "step": 11336
    },
    {
      "epoch": 0.8720098453965079,
      "grad_norm": 23.75,
      "learning_rate": 1.691985231905238e-06,
      "loss": 0.0092,
      "mean_token_accuracy": 0.9983633160591125,
      "num_tokens": 4924514.0,
      "step": 11337
    },
    {
      "epoch": 0.8720867625567265,
      "grad_norm": 10.125,
      "learning_rate": 1.6918698561649105e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4924822.0,
      "step": 11338
    },
    {
      "epoch": 0.8721636797169449,
      "grad_norm": 12.6875,
      "learning_rate": 1.691754480424583e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 4925368.0,
      "step": 11339
    },
    {
      "epoch": 0.8722405968771633,
      "grad_norm": 13.9375,
      "learning_rate": 1.691639104684255e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 4925656.0,
      "step": 11340
    },
    {
      "epoch": 0.8723175140373818,
      "grad_norm": 89.0,
      "learning_rate": 1.6915237289439273e-06,
      "loss": 0.0506,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 4925980.0,
      "step": 11341
    },
    {
      "epoch": 0.8723944311976002,
      "grad_norm": 49.75,
      "learning_rate": 1.6914083532035998e-06,
      "loss": 0.0321,
      "mean_token_accuracy": 0.982332170009613,
      "num_tokens": 4926264.0,
      "step": 11342
    },
    {
      "epoch": 0.8724713483578186,
      "grad_norm": 10.875,
      "learning_rate": 1.6912929774632722e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4926838.0,
      "step": 11343
    },
    {
      "epoch": 0.8725482655180371,
      "grad_norm": 39.0,
      "learning_rate": 1.6911776017229445e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9842271208763123,
      "num_tokens": 4927156.0,
      "step": 11344
    },
    {
      "epoch": 0.8726251826782555,
      "grad_norm": 12.9375,
      "learning_rate": 1.6910622259826165e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 4927414.0,
      "step": 11345
    },
    {
      "epoch": 0.8727020998384739,
      "grad_norm": 19.75,
      "learning_rate": 1.690946850242289e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.9982905983924866,
      "num_tokens": 4928000.0,
      "step": 11346
    },
    {
      "epoch": 0.8727790169986924,
      "grad_norm": 18.875,
      "learning_rate": 1.6908314745019615e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 4928564.0,
      "step": 11347
    },
    {
      "epoch": 0.8728559341589108,
      "grad_norm": 12.25,
      "learning_rate": 1.6907160987616337e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9982728958129883,
      "num_tokens": 4929144.0,
      "step": 11348
    },
    {
      "epoch": 0.8729328513191293,
      "grad_norm": 22.875,
      "learning_rate": 1.6906007230213062e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4929466.0,
      "step": 11349
    },
    {
      "epoch": 0.8730097684793477,
      "grad_norm": 53.75,
      "learning_rate": 1.6904853472809787e-06,
      "loss": 0.0402,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 4929768.0,
      "step": 11350
    },
    {
      "epoch": 0.8730866856395662,
      "grad_norm": 71.5,
      "learning_rate": 1.6903699715406507e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9889298677444458,
      "num_tokens": 4930040.0,
      "step": 11351
    },
    {
      "epoch": 0.8731636027997847,
      "grad_norm": 57.75,
      "learning_rate": 1.690254595800323e-06,
      "loss": 0.0319,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4930350.0,
      "step": 11352
    },
    {
      "epoch": 0.8732405199600031,
      "grad_norm": 21.0,
      "learning_rate": 1.6901392200599954e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 4930952.0,
      "step": 11353
    },
    {
      "epoch": 0.8733174371202215,
      "grad_norm": 47.75,
      "learning_rate": 1.690023844319668e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4931274.0,
      "step": 11354
    },
    {
      "epoch": 0.87339435428044,
      "grad_norm": 8.3125,
      "learning_rate": 1.6899084685793402e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4931854.0,
      "step": 11355
    },
    {
      "epoch": 0.8734712714406584,
      "grad_norm": 13.1875,
      "learning_rate": 1.6897930928390122e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.9965457916259766,
      "num_tokens": 4932434.0,
      "step": 11356
    },
    {
      "epoch": 0.8735481886008769,
      "grad_norm": 32.5,
      "learning_rate": 1.6896777170986847e-06,
      "loss": 0.0307,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4932766.0,
      "step": 11357
    },
    {
      "epoch": 0.8736251057610953,
      "grad_norm": 19.25,
      "learning_rate": 1.6895623413583572e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9912434220314026,
      "num_tokens": 4933338.0,
      "step": 11358
    },
    {
      "epoch": 0.8737020229213137,
      "grad_norm": 38.0,
      "learning_rate": 1.6894469656180294e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 4933658.0,
      "step": 11359
    },
    {
      "epoch": 0.8737789400815322,
      "grad_norm": 5.5,
      "learning_rate": 1.6893315898777019e-06,
      "loss": 0.005,
      "mean_token_accuracy": 1.0,
      "num_tokens": 4934462.0,
      "step": 11360
    },
    {
      "epoch": 0.8738558572417506,
      "grad_norm": 8.125,
      "learning_rate": 1.689216214137374e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9981982111930847,
      "num_tokens": 4935018.0,
      "step": 11361
    },
    {
      "epoch": 0.873932774401969,
      "grad_norm": 34.25,
      "learning_rate": 1.6891008383970464e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9911242723464966,
      "num_tokens": 4935358.0,
      "step": 11362
    },
    {
      "epoch": 0.8740096915621876,
      "grad_norm": 6.53125,
      "learning_rate": 1.6889854626567187e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4935680.0,
      "step": 11363
    },
    {
      "epoch": 0.874086608722406,
      "grad_norm": 38.0,
      "learning_rate": 1.6888700869163911e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 4935986.0,
      "step": 11364
    },
    {
      "epoch": 0.8741635258826244,
      "grad_norm": 48.25,
      "learning_rate": 1.6887547111760636e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4936286.0,
      "step": 11365
    },
    {
      "epoch": 0.8742404430428429,
      "grad_norm": 30.0,
      "learning_rate": 1.6886393354357359e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 4936608.0,
      "step": 11366
    },
    {
      "epoch": 0.8743173602030613,
      "grad_norm": 7.1875,
      "learning_rate": 1.688523959695408e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 4937152.0,
      "step": 11367
    },
    {
      "epoch": 0.8743942773632798,
      "grad_norm": 50.75,
      "learning_rate": 1.6884085839550804e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9825783967971802,
      "num_tokens": 4937440.0,
      "step": 11368
    },
    {
      "epoch": 0.8744711945234982,
      "grad_norm": 35.25,
      "learning_rate": 1.6882932082147528e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4937750.0,
      "step": 11369
    },
    {
      "epoch": 0.8745481116837166,
      "grad_norm": 16.125,
      "learning_rate": 1.688177832474425e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 4938036.0,
      "step": 11370
    },
    {
      "epoch": 0.8746250288439351,
      "grad_norm": 30.5,
      "learning_rate": 1.6880624567340976e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9909090995788574,
      "num_tokens": 4938368.0,
      "step": 11371
    },
    {
      "epoch": 0.8747019460041535,
      "grad_norm": 51.0,
      "learning_rate": 1.6879470809937696e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9914529919624329,
      "num_tokens": 4938720.0,
      "step": 11372
    },
    {
      "epoch": 0.8747788631643719,
      "grad_norm": 48.25,
      "learning_rate": 1.687831705253442e-06,
      "loss": 0.0377,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 4939052.0,
      "step": 11373
    },
    {
      "epoch": 0.8748557803245904,
      "grad_norm": 14.375,
      "learning_rate": 1.6877163295131143e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4939370.0,
      "step": 11374
    },
    {
      "epoch": 0.8749326974848088,
      "grad_norm": 25.25,
      "learning_rate": 1.6876009537727868e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 4939664.0,
      "step": 11375
    },
    {
      "epoch": 0.8750096146450274,
      "grad_norm": 29.25,
      "learning_rate": 1.687485578032459e-06,
      "loss": 0.0134,
      "mean_token_accuracy": 0.9925233721733093,
      "num_tokens": 4940200.0,
      "step": 11376
    },
    {
      "epoch": 0.8750865318052458,
      "grad_norm": 32.0,
      "learning_rate": 1.6873702022921315e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9929701089859009,
      "num_tokens": 4940770.0,
      "step": 11377
    },
    {
      "epoch": 0.8751634489654642,
      "grad_norm": 85.5,
      "learning_rate": 1.6872548265518036e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 4941108.0,
      "step": 11378
    },
    {
      "epoch": 0.8752403661256827,
      "grad_norm": 43.25,
      "learning_rate": 1.687139450811476e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9936102032661438,
      "num_tokens": 4941422.0,
      "step": 11379
    },
    {
      "epoch": 0.8753172832859011,
      "grad_norm": 20.5,
      "learning_rate": 1.6870240750711483e-06,
      "loss": 0.0259,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 4941686.0,
      "step": 11380
    },
    {
      "epoch": 0.8753942004461195,
      "grad_norm": 34.25,
      "learning_rate": 1.6869086993308208e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 4942248.0,
      "step": 11381
    },
    {
      "epoch": 0.875471117606338,
      "grad_norm": 46.5,
      "learning_rate": 1.6867933235904933e-06,
      "loss": 0.0306,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 4942558.0,
      "step": 11382
    },
    {
      "epoch": 0.8755480347665564,
      "grad_norm": 27.5,
      "learning_rate": 1.6866779478501653e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 4942894.0,
      "step": 11383
    },
    {
      "epoch": 0.8756249519267748,
      "grad_norm": 17.5,
      "learning_rate": 1.6865625721098376e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9965694546699524,
      "num_tokens": 4943478.0,
      "step": 11384
    },
    {
      "epoch": 0.8757018690869933,
      "grad_norm": 12.9375,
      "learning_rate": 1.68644719636951e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 4944018.0,
      "step": 11385
    },
    {
      "epoch": 0.8757787862472117,
      "grad_norm": 7.65625,
      "learning_rate": 1.6863318206291825e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9961758852005005,
      "num_tokens": 4944542.0,
      "step": 11386
    },
    {
      "epoch": 0.8758557034074302,
      "grad_norm": 19.5,
      "learning_rate": 1.6862164448888548e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 4944862.0,
      "step": 11387
    },
    {
      "epoch": 0.8759326205676486,
      "grad_norm": 19.625,
      "learning_rate": 1.6861010691485272e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 4945180.0,
      "step": 11388
    },
    {
      "epoch": 0.876009537727867,
      "grad_norm": 37.25,
      "learning_rate": 1.6859856934081993e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9873015880584717,
      "num_tokens": 4945498.0,
      "step": 11389
    },
    {
      "epoch": 0.8760864548880856,
      "grad_norm": 76.5,
      "learning_rate": 1.6858703176678717e-06,
      "loss": 0.0328,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4945782.0,
      "step": 11390
    },
    {
      "epoch": 0.876163372048304,
      "grad_norm": 12.5,
      "learning_rate": 1.685754941927544e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 4946350.0,
      "step": 11391
    },
    {
      "epoch": 0.8762402892085224,
      "grad_norm": 3.765625,
      "learning_rate": 1.6856395661872165e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4947108.0,
      "step": 11392
    },
    {
      "epoch": 0.8763172063687409,
      "grad_norm": 31.375,
      "learning_rate": 1.685524190446889e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4947688.0,
      "step": 11393
    },
    {
      "epoch": 0.8763941235289593,
      "grad_norm": 29.625,
      "learning_rate": 1.685408814706561e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 4948016.0,
      "step": 11394
    },
    {
      "epoch": 0.8764710406891778,
      "grad_norm": 23.0,
      "learning_rate": 1.6852934389662332e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4948326.0,
      "step": 11395
    },
    {
      "epoch": 0.8765479578493962,
      "grad_norm": 10.5625,
      "learning_rate": 1.6851780632259057e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 4948634.0,
      "step": 11396
    },
    {
      "epoch": 0.8766248750096146,
      "grad_norm": 27.5,
      "learning_rate": 1.6850626874855782e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9818181991577148,
      "num_tokens": 4948966.0,
      "step": 11397
    },
    {
      "epoch": 0.8767017921698331,
      "grad_norm": 23.25,
      "learning_rate": 1.6849473117452504e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4949248.0,
      "step": 11398
    },
    {
      "epoch": 0.8767787093300515,
      "grad_norm": 11.375,
      "learning_rate": 1.6848319360049225e-06,
      "loss": 0.026,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 4949554.0,
      "step": 11399
    },
    {
      "epoch": 0.8768556264902699,
      "grad_norm": 13.25,
      "learning_rate": 1.684716560264595e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9757575988769531,
      "num_tokens": 4949886.0,
      "step": 11400
    },
    {
      "epoch": 0.8769325436504884,
      "grad_norm": 33.5,
      "learning_rate": 1.6846011845242674e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9961636662483215,
      "num_tokens": 4950670.0,
      "step": 11401
    },
    {
      "epoch": 0.8770094608107069,
      "grad_norm": 34.75,
      "learning_rate": 1.6844858087839397e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9932318329811096,
      "num_tokens": 4951262.0,
      "step": 11402
    },
    {
      "epoch": 0.8770863779709254,
      "grad_norm": 22.75,
      "learning_rate": 1.6843704330436122e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 4951564.0,
      "step": 11403
    },
    {
      "epoch": 0.8771632951311438,
      "grad_norm": 33.75,
      "learning_rate": 1.6842550573032846e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 4952126.0,
      "step": 11404
    },
    {
      "epoch": 0.8772402122913622,
      "grad_norm": 30.375,
      "learning_rate": 1.6841396815629567e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 4952680.0,
      "step": 11405
    },
    {
      "epoch": 0.8773171294515807,
      "grad_norm": 10.5,
      "learning_rate": 1.684024305822629e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 4953024.0,
      "step": 11406
    },
    {
      "epoch": 0.8773940466117991,
      "grad_norm": 31.625,
      "learning_rate": 1.6839089300823014e-06,
      "loss": 0.0374,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4953300.0,
      "step": 11407
    },
    {
      "epoch": 0.8774709637720175,
      "grad_norm": 25.875,
      "learning_rate": 1.6837935543419739e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 4953850.0,
      "step": 11408
    },
    {
      "epoch": 0.877547880932236,
      "grad_norm": 6.71875,
      "learning_rate": 1.6836781786016461e-06,
      "loss": 0.0094,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 4954422.0,
      "step": 11409
    },
    {
      "epoch": 0.8776247980924544,
      "grad_norm": 25.875,
      "learning_rate": 1.6835628028613182e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.996960461139679,
      "num_tokens": 4954752.0,
      "step": 11410
    },
    {
      "epoch": 0.8777017152526728,
      "grad_norm": 60.25,
      "learning_rate": 1.6834474271209906e-06,
      "loss": 0.0255,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4955062.0,
      "step": 11411
    },
    {
      "epoch": 0.8777786324128913,
      "grad_norm": 24.625,
      "learning_rate": 1.6833320513806631e-06,
      "loss": 0.0072,
      "mean_token_accuracy": 0.9975247383117676,
      "num_tokens": 4955872.0,
      "step": 11412
    },
    {
      "epoch": 0.8778555495731097,
      "grad_norm": 35.25,
      "learning_rate": 1.6832166756403354e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4956458.0,
      "step": 11413
    },
    {
      "epoch": 0.8779324667333283,
      "grad_norm": 13.125,
      "learning_rate": 1.6831012999000078e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 4957006.0,
      "step": 11414
    },
    {
      "epoch": 0.8780093838935467,
      "grad_norm": 32.0,
      "learning_rate": 1.6829859241596803e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9903537034988403,
      "num_tokens": 4957318.0,
      "step": 11415
    },
    {
      "epoch": 0.8780863010537651,
      "grad_norm": 23.25,
      "learning_rate": 1.6828705484193524e-06,
      "loss": 0.03,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 4957616.0,
      "step": 11416
    },
    {
      "epoch": 0.8781632182139836,
      "grad_norm": 68.0,
      "learning_rate": 1.6827551726790246e-06,
      "loss": 0.0448,
      "mean_token_accuracy": 0.9781249761581421,
      "num_tokens": 4957940.0,
      "step": 11417
    },
    {
      "epoch": 0.878240135374202,
      "grad_norm": 43.5,
      "learning_rate": 1.682639796938697e-06,
      "loss": 0.0363,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4958250.0,
      "step": 11418
    },
    {
      "epoch": 0.8783170525344204,
      "grad_norm": 10.1875,
      "learning_rate": 1.6825244211983696e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4958798.0,
      "step": 11419
    },
    {
      "epoch": 0.8783939696946389,
      "grad_norm": 80.5,
      "learning_rate": 1.6824090454580418e-06,
      "loss": 0.0384,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 4959108.0,
      "step": 11420
    },
    {
      "epoch": 0.8784708868548573,
      "grad_norm": 7.21875,
      "learning_rate": 1.6822936697177139e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 4959656.0,
      "step": 11421
    },
    {
      "epoch": 0.8785478040150758,
      "grad_norm": 31.125,
      "learning_rate": 1.6821782939773863e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 4959978.0,
      "step": 11422
    },
    {
      "epoch": 0.8786247211752942,
      "grad_norm": 8.8125,
      "learning_rate": 1.6820629182370588e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4960524.0,
      "step": 11423
    },
    {
      "epoch": 0.8787016383355126,
      "grad_norm": 6.34375,
      "learning_rate": 1.681947542496731e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9961612224578857,
      "num_tokens": 4961046.0,
      "step": 11424
    },
    {
      "epoch": 0.8787785554957311,
      "grad_norm": 34.75,
      "learning_rate": 1.6818321667564035e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.97826087474823,
      "num_tokens": 4961324.0,
      "step": 11425
    },
    {
      "epoch": 0.8788554726559495,
      "grad_norm": 17.625,
      "learning_rate": 1.6817167910160758e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 4961878.0,
      "step": 11426
    },
    {
      "epoch": 0.878932389816168,
      "grad_norm": 6.0,
      "learning_rate": 1.681601415275748e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4962424.0,
      "step": 11427
    },
    {
      "epoch": 0.8790093069763865,
      "grad_norm": 9.0,
      "learning_rate": 1.6814860395354203e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9934853315353394,
      "num_tokens": 4962732.0,
      "step": 11428
    },
    {
      "epoch": 0.8790862241366049,
      "grad_norm": 14.4375,
      "learning_rate": 1.6813706637950928e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9927404522895813,
      "num_tokens": 4963284.0,
      "step": 11429
    },
    {
      "epoch": 0.8791631412968233,
      "grad_norm": 30.0,
      "learning_rate": 1.681255288054765e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4963604.0,
      "step": 11430
    },
    {
      "epoch": 0.8792400584570418,
      "grad_norm": 30.375,
      "learning_rate": 1.6811399123144375e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.9930070042610168,
      "num_tokens": 4963892.0,
      "step": 11431
    },
    {
      "epoch": 0.8793169756172602,
      "grad_norm": 24.375,
      "learning_rate": 1.6810245365741096e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 4964174.0,
      "step": 11432
    },
    {
      "epoch": 0.8793938927774787,
      "grad_norm": 25.375,
      "learning_rate": 1.680909160833782e-06,
      "loss": 0.037,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4964468.0,
      "step": 11433
    },
    {
      "epoch": 0.8794708099376971,
      "grad_norm": 13.25,
      "learning_rate": 1.6807937850934543e-06,
      "loss": 0.0366,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4964782.0,
      "step": 11434
    },
    {
      "epoch": 0.8795477270979155,
      "grad_norm": 7.25,
      "learning_rate": 1.6806784093531267e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 4965072.0,
      "step": 11435
    },
    {
      "epoch": 0.879624644258134,
      "grad_norm": 16.0,
      "learning_rate": 1.6805630336127992e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9928186535835266,
      "num_tokens": 4965630.0,
      "step": 11436
    },
    {
      "epoch": 0.8797015614183524,
      "grad_norm": 50.75,
      "learning_rate": 1.6804476578724713e-06,
      "loss": 0.0414,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 4965906.0,
      "step": 11437
    },
    {
      "epoch": 0.8797784785785708,
      "grad_norm": 3.921875,
      "learning_rate": 1.6803322821321435e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9961538314819336,
      "num_tokens": 4966688.0,
      "step": 11438
    },
    {
      "epoch": 0.8798553957387893,
      "grad_norm": 19.25,
      "learning_rate": 1.680216906391816e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9937965273857117,
      "num_tokens": 4967496.0,
      "step": 11439
    },
    {
      "epoch": 0.8799323128990078,
      "grad_norm": 14.8125,
      "learning_rate": 1.6801015306514885e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 4967812.0,
      "step": 11440
    },
    {
      "epoch": 0.8800092300592263,
      "grad_norm": 62.75,
      "learning_rate": 1.6799861549111607e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9862637519836426,
      "num_tokens": 4968178.0,
      "step": 11441
    },
    {
      "epoch": 0.8800861472194447,
      "grad_norm": 19.125,
      "learning_rate": 1.6798707791708332e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9827089309692383,
      "num_tokens": 4968526.0,
      "step": 11442
    },
    {
      "epoch": 0.8801630643796631,
      "grad_norm": 13.375,
      "learning_rate": 1.6797554034305052e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9961734414100647,
      "num_tokens": 4969312.0,
      "step": 11443
    },
    {
      "epoch": 0.8802399815398816,
      "grad_norm": 9.4375,
      "learning_rate": 1.6796400276901777e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4969852.0,
      "step": 11444
    },
    {
      "epoch": 0.8803168987001,
      "grad_norm": 15.0,
      "learning_rate": 1.67952465194985e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9907692074775696,
      "num_tokens": 4970178.0,
      "step": 11445
    },
    {
      "epoch": 0.8803938158603184,
      "grad_norm": 28.5,
      "learning_rate": 1.6794092762095224e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 4970450.0,
      "step": 11446
    },
    {
      "epoch": 0.8804707330205369,
      "grad_norm": 10.1875,
      "learning_rate": 1.679293900469195e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 4971014.0,
      "step": 11447
    },
    {
      "epoch": 0.8805476501807553,
      "grad_norm": 51.75,
      "learning_rate": 1.679178524728867e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 4971284.0,
      "step": 11448
    },
    {
      "epoch": 0.8806245673409737,
      "grad_norm": 4.875,
      "learning_rate": 1.6790631489885392e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9975669384002686,
      "num_tokens": 4972108.0,
      "step": 11449
    },
    {
      "epoch": 0.8807014845011922,
      "grad_norm": 20.375,
      "learning_rate": 1.6789477732482117e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9874607920646667,
      "num_tokens": 4972428.0,
      "step": 11450
    },
    {
      "epoch": 0.8807784016614106,
      "grad_norm": 4.375,
      "learning_rate": 1.6788323975078842e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9962476491928101,
      "num_tokens": 4972962.0,
      "step": 11451
    },
    {
      "epoch": 0.8808553188216292,
      "grad_norm": 18.0,
      "learning_rate": 1.6787170217675564e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 4973558.0,
      "step": 11452
    },
    {
      "epoch": 0.8809322359818476,
      "grad_norm": 10.875,
      "learning_rate": 1.6786016460272289e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 4974104.0,
      "step": 11453
    },
    {
      "epoch": 0.881009153142066,
      "grad_norm": 9.125,
      "learning_rate": 1.678486270286901e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9926062822341919,
      "num_tokens": 4974646.0,
      "step": 11454
    },
    {
      "epoch": 0.8810860703022845,
      "grad_norm": 9.5,
      "learning_rate": 1.6783708945465734e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 4974980.0,
      "step": 11455
    },
    {
      "epoch": 0.8811629874625029,
      "grad_norm": 27.75,
      "learning_rate": 1.6782555188062457e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9933110475540161,
      "num_tokens": 4975280.0,
      "step": 11456
    },
    {
      "epoch": 0.8812399046227213,
      "grad_norm": 10.3125,
      "learning_rate": 1.6781401430659181e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 4975848.0,
      "step": 11457
    },
    {
      "epoch": 0.8813168217829398,
      "grad_norm": 7.625,
      "learning_rate": 1.6780247673255906e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9862542748451233,
      "num_tokens": 4976140.0,
      "step": 11458
    },
    {
      "epoch": 0.8813937389431582,
      "grad_norm": 20.0,
      "learning_rate": 1.6779093915852626e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 4976422.0,
      "step": 11459
    },
    {
      "epoch": 0.8814706561033767,
      "grad_norm": 16.75,
      "learning_rate": 1.677794015844935e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 4976996.0,
      "step": 11460
    },
    {
      "epoch": 0.8815475732635951,
      "grad_norm": 7.9375,
      "learning_rate": 1.6776786401046074e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 4977540.0,
      "step": 11461
    },
    {
      "epoch": 0.8816244904238135,
      "grad_norm": 46.25,
      "learning_rate": 1.6775632643642798e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9964664578437805,
      "num_tokens": 4977824.0,
      "step": 11462
    },
    {
      "epoch": 0.881701407584032,
      "grad_norm": 39.25,
      "learning_rate": 1.677447888623952e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9855072498321533,
      "num_tokens": 4978102.0,
      "step": 11463
    },
    {
      "epoch": 0.8817783247442504,
      "grad_norm": 22.0,
      "learning_rate": 1.6773325128836241e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9969696998596191,
      "num_tokens": 4978434.0,
      "step": 11464
    },
    {
      "epoch": 0.8818552419044688,
      "grad_norm": 29.25,
      "learning_rate": 1.6772171371432966e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 4978956.0,
      "step": 11465
    },
    {
      "epoch": 0.8819321590646874,
      "grad_norm": 33.75,
      "learning_rate": 1.677101761402969e-06,
      "loss": 0.0336,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 4979238.0,
      "step": 11466
    },
    {
      "epoch": 0.8820090762249058,
      "grad_norm": 21.75,
      "learning_rate": 1.6769863856626413e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9851632118225098,
      "num_tokens": 4979576.0,
      "step": 11467
    },
    {
      "epoch": 0.8820859933851242,
      "grad_norm": 24.5,
      "learning_rate": 1.6768710099223138e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 4979916.0,
      "step": 11468
    },
    {
      "epoch": 0.8821629105453427,
      "grad_norm": 44.5,
      "learning_rate": 1.6767556341819863e-06,
      "loss": 0.0218,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 4980212.0,
      "step": 11469
    },
    {
      "epoch": 0.8822398277055611,
      "grad_norm": 12.875,
      "learning_rate": 1.6766402584416583e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 4980518.0,
      "step": 11470
    },
    {
      "epoch": 0.8823167448657796,
      "grad_norm": 51.0,
      "learning_rate": 1.6765248827013306e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9809523820877075,
      "num_tokens": 4980834.0,
      "step": 11471
    },
    {
      "epoch": 0.882393662025998,
      "grad_norm": 31.875,
      "learning_rate": 1.676409506961003e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9930915236473083,
      "num_tokens": 4981414.0,
      "step": 11472
    },
    {
      "epoch": 0.8824705791862164,
      "grad_norm": 9.875,
      "learning_rate": 1.6762941312206755e-06,
      "loss": 0.0284,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 4981702.0,
      "step": 11473
    },
    {
      "epoch": 0.8825474963464349,
      "grad_norm": 37.5,
      "learning_rate": 1.6761787554803478e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9908592104911804,
      "num_tokens": 4982250.0,
      "step": 11474
    },
    {
      "epoch": 0.8826244135066533,
      "grad_norm": 20.75,
      "learning_rate": 1.6760633797400198e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.9933993220329285,
      "num_tokens": 4982554.0,
      "step": 11475
    },
    {
      "epoch": 0.8827013306668717,
      "grad_norm": 13.4375,
      "learning_rate": 1.6759480039996923e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 4982868.0,
      "step": 11476
    },
    {
      "epoch": 0.8827782478270902,
      "grad_norm": 15.625,
      "learning_rate": 1.6758326282593648e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 4983176.0,
      "step": 11477
    },
    {
      "epoch": 0.8828551649873086,
      "grad_norm": 9.625,
      "learning_rate": 1.675717252519037e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 4983722.0,
      "step": 11478
    },
    {
      "epoch": 0.8829320821475272,
      "grad_norm": 8.3125,
      "learning_rate": 1.6756018767787095e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9931153059005737,
      "num_tokens": 4984304.0,
      "step": 11479
    },
    {
      "epoch": 0.8830089993077456,
      "grad_norm": 13.1875,
      "learning_rate": 1.6754865010383818e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9964221715927124,
      "num_tokens": 4984864.0,
      "step": 11480
    },
    {
      "epoch": 0.883085916467964,
      "grad_norm": 12.6875,
      "learning_rate": 1.675371125298054e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 4985226.0,
      "step": 11481
    },
    {
      "epoch": 0.8831628336281825,
      "grad_norm": 30.375,
      "learning_rate": 1.6752557495577263e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9879879951477051,
      "num_tokens": 4985560.0,
      "step": 11482
    },
    {
      "epoch": 0.8832397507884009,
      "grad_norm": 5.8125,
      "learning_rate": 1.6751403738173987e-06,
      "loss": 0.006,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 4986380.0,
      "step": 11483
    },
    {
      "epoch": 0.8833166679486193,
      "grad_norm": 21.125,
      "learning_rate": 1.675024998077071e-06,
      "loss": 0.0216,
      "mean_token_accuracy": 0.9913544654846191,
      "num_tokens": 4986728.0,
      "step": 11484
    },
    {
      "epoch": 0.8833935851088378,
      "grad_norm": 15.4375,
      "learning_rate": 1.6749096223367435e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 4987244.0,
      "step": 11485
    },
    {
      "epoch": 0.8834705022690562,
      "grad_norm": 7.71875,
      "learning_rate": 1.6747942465964155e-06,
      "loss": 0.008,
      "mean_token_accuracy": 0.9966722130775452,
      "num_tokens": 4987846.0,
      "step": 11486
    },
    {
      "epoch": 0.8835474194292746,
      "grad_norm": 50.25,
      "learning_rate": 1.674678870856088e-06,
      "loss": 0.0411,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 4988152.0,
      "step": 11487
    },
    {
      "epoch": 0.8836243365894931,
      "grad_norm": 16.25,
      "learning_rate": 1.6745634951157602e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 4988492.0,
      "step": 11488
    },
    {
      "epoch": 0.8837012537497115,
      "grad_norm": 37.25,
      "learning_rate": 1.6744481193754327e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9870550036430359,
      "num_tokens": 4988802.0,
      "step": 11489
    },
    {
      "epoch": 0.88377817090993,
      "grad_norm": 7.6875,
      "learning_rate": 1.6743327436351052e-06,
      "loss": 0.0083,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 4989388.0,
      "step": 11490
    },
    {
      "epoch": 0.8838550880701485,
      "grad_norm": 40.5,
      "learning_rate": 1.6742173678947774e-06,
      "loss": 0.0241,
      "mean_token_accuracy": 0.9852070808410645,
      "num_tokens": 4989728.0,
      "step": 11491
    },
    {
      "epoch": 0.8839320052303669,
      "grad_norm": 18.25,
      "learning_rate": 1.6741019921544495e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9927140474319458,
      "num_tokens": 4990278.0,
      "step": 11492
    },
    {
      "epoch": 0.8840089223905854,
      "grad_norm": 44.75,
      "learning_rate": 1.673986616414122e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 4990602.0,
      "step": 11493
    },
    {
      "epoch": 0.8840858395508038,
      "grad_norm": 14.25,
      "learning_rate": 1.6738712406737944e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4991142.0,
      "step": 11494
    },
    {
      "epoch": 0.8841627567110222,
      "grad_norm": 22.75,
      "learning_rate": 1.6737558649334667e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 4991712.0,
      "step": 11495
    },
    {
      "epoch": 0.8842396738712407,
      "grad_norm": 8.1875,
      "learning_rate": 1.6736404891931392e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 4991996.0,
      "step": 11496
    },
    {
      "epoch": 0.8843165910314591,
      "grad_norm": 3.921875,
      "learning_rate": 1.6735251134528112e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 4992538.0,
      "step": 11497
    },
    {
      "epoch": 0.8843935081916776,
      "grad_norm": 58.25,
      "learning_rate": 1.6734097377124837e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 4992862.0,
      "step": 11498
    },
    {
      "epoch": 0.884470425351896,
      "grad_norm": 35.75,
      "learning_rate": 1.673294361972156e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 4993178.0,
      "step": 11499
    },
    {
      "epoch": 0.8845473425121144,
      "grad_norm": 15.375,
      "learning_rate": 1.6731789862318284e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9930675625801086,
      "num_tokens": 4993756.0,
      "step": 11500
    },
    {
      "epoch": 0.8846242596723329,
      "grad_norm": 17.125,
      "learning_rate": 1.6730636104915009e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 4994042.0,
      "step": 11501
    },
    {
      "epoch": 0.8847011768325513,
      "grad_norm": 25.75,
      "learning_rate": 1.672948234751173e-06,
      "loss": 0.0125,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 4994564.0,
      "step": 11502
    },
    {
      "epoch": 0.8847780939927697,
      "grad_norm": 17.125,
      "learning_rate": 1.6728328590108452e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 4995346.0,
      "step": 11503
    },
    {
      "epoch": 0.8848550111529883,
      "grad_norm": 29.125,
      "learning_rate": 1.6727174832705176e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 4995656.0,
      "step": 11504
    },
    {
      "epoch": 0.8849319283132067,
      "grad_norm": 11.9375,
      "learning_rate": 1.6726021075301901e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 4996196.0,
      "step": 11505
    },
    {
      "epoch": 0.8850088454734252,
      "grad_norm": 22.625,
      "learning_rate": 1.6724867317898624e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9922630786895752,
      "num_tokens": 4996714.0,
      "step": 11506
    },
    {
      "epoch": 0.8850857626336436,
      "grad_norm": 27.625,
      "learning_rate": 1.6723713560495348e-06,
      "loss": 0.0362,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4996970.0,
      "step": 11507
    },
    {
      "epoch": 0.885162679793862,
      "grad_norm": 9.125,
      "learning_rate": 1.6722559803092069e-06,
      "loss": 0.0246,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 4997264.0,
      "step": 11508
    },
    {
      "epoch": 0.8852395969540805,
      "grad_norm": 28.0,
      "learning_rate": 1.6721406045688794e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 4997576.0,
      "step": 11509
    },
    {
      "epoch": 0.8853165141142989,
      "grad_norm": 37.25,
      "learning_rate": 1.6720252288285516e-06,
      "loss": 0.0261,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 4997904.0,
      "step": 11510
    },
    {
      "epoch": 0.8853934312745173,
      "grad_norm": 14.4375,
      "learning_rate": 1.671909853088224e-06,
      "loss": 0.0074,
      "mean_token_accuracy": 0.9963325262069702,
      "num_tokens": 4998724.0,
      "step": 11511
    },
    {
      "epoch": 0.8854703484347358,
      "grad_norm": 16.0,
      "learning_rate": 1.6717944773478966e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.980079710483551,
      "num_tokens": 4998976.0,
      "step": 11512
    },
    {
      "epoch": 0.8855472655949542,
      "grad_norm": 12.8125,
      "learning_rate": 1.6716791016075686e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 4999548.0,
      "step": 11513
    },
    {
      "epoch": 0.8856241827551726,
      "grad_norm": 64.5,
      "learning_rate": 1.6715637258672409e-06,
      "loss": 0.0312,
      "mean_token_accuracy": 0.9803921580314636,
      "num_tokens": 4999856.0,
      "step": 11514
    },
    {
      "epoch": 0.8857010999153911,
      "grad_norm": 24.875,
      "learning_rate": 1.6714483501269133e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9910979270935059,
      "num_tokens": 5000194.0,
      "step": 11515
    },
    {
      "epoch": 0.8857780170756095,
      "grad_norm": 11.4375,
      "learning_rate": 1.6713329743865858e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 5000516.0,
      "step": 11516
    },
    {
      "epoch": 0.8858549342358281,
      "grad_norm": 20.75,
      "learning_rate": 1.671217598646258e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9967213273048401,
      "num_tokens": 5000822.0,
      "step": 11517
    },
    {
      "epoch": 0.8859318513960465,
      "grad_norm": 47.0,
      "learning_rate": 1.6711022229059305e-06,
      "loss": 0.0361,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 5001124.0,
      "step": 11518
    },
    {
      "epoch": 0.8860087685562649,
      "grad_norm": 35.75,
      "learning_rate": 1.6709868471656026e-06,
      "loss": 0.0275,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 5001408.0,
      "step": 11519
    },
    {
      "epoch": 0.8860856857164834,
      "grad_norm": 38.0,
      "learning_rate": 1.670871471425275e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 5001988.0,
      "step": 11520
    },
    {
      "epoch": 0.8861626028767018,
      "grad_norm": 18.25,
      "learning_rate": 1.6707560956849473e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9961795806884766,
      "num_tokens": 5003038.0,
      "step": 11521
    },
    {
      "epoch": 0.8862395200369202,
      "grad_norm": 8.5,
      "learning_rate": 1.6706407199446198e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 5003294.0,
      "step": 11522
    },
    {
      "epoch": 0.8863164371971387,
      "grad_norm": 31.625,
      "learning_rate": 1.6705253442042922e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 5004054.0,
      "step": 11523
    },
    {
      "epoch": 0.8863933543573571,
      "grad_norm": 13.125,
      "learning_rate": 1.6704099684639643e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9970674514770508,
      "num_tokens": 5004396.0,
      "step": 11524
    },
    {
      "epoch": 0.8864702715175756,
      "grad_norm": 58.75,
      "learning_rate": 1.6702945927236365e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 5005008.0,
      "step": 11525
    },
    {
      "epoch": 0.886547188677794,
      "grad_norm": 7.53125,
      "learning_rate": 1.670179216983309e-06,
      "loss": 0.0225,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 5005342.0,
      "step": 11526
    },
    {
      "epoch": 0.8866241058380124,
      "grad_norm": 80.0,
      "learning_rate": 1.6700638412429815e-06,
      "loss": 0.0226,
      "mean_token_accuracy": 0.9933444261550903,
      "num_tokens": 5005944.0,
      "step": 11527
    },
    {
      "epoch": 0.886701022998231,
      "grad_norm": 23.5,
      "learning_rate": 1.6699484655026537e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 5006268.0,
      "step": 11528
    },
    {
      "epoch": 0.8867779401584494,
      "grad_norm": 10.25,
      "learning_rate": 1.6698330897623262e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 5006788.0,
      "step": 11529
    },
    {
      "epoch": 0.8868548573186678,
      "grad_norm": 30.5,
      "learning_rate": 1.6697177140219983e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9950124621391296,
      "num_tokens": 5007592.0,
      "step": 11530
    },
    {
      "epoch": 0.8869317744788863,
      "grad_norm": 28.875,
      "learning_rate": 1.6696023382816707e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9975185990333557,
      "num_tokens": 5008400.0,
      "step": 11531
    },
    {
      "epoch": 0.8870086916391047,
      "grad_norm": 4.65625,
      "learning_rate": 1.669486962541343e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 5008946.0,
      "step": 11532
    },
    {
      "epoch": 0.8870856087993231,
      "grad_norm": 34.75,
      "learning_rate": 1.6693715868010155e-06,
      "loss": 0.0245,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 5009292.0,
      "step": 11533
    },
    {
      "epoch": 0.8871625259595416,
      "grad_norm": 7.1875,
      "learning_rate": 1.669256211060688e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9964726567268372,
      "num_tokens": 5009860.0,
      "step": 11534
    },
    {
      "epoch": 0.88723944311976,
      "grad_norm": 7.25,
      "learning_rate": 1.66914083532036e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9910714030265808,
      "num_tokens": 5010198.0,
      "step": 11535
    },
    {
      "epoch": 0.8873163602799785,
      "grad_norm": 32.75,
      "learning_rate": 1.6690254595800322e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9899665713310242,
      "num_tokens": 5010498.0,
      "step": 11536
    },
    {
      "epoch": 0.8873932774401969,
      "grad_norm": 20.5,
      "learning_rate": 1.6689100838397047e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 5010810.0,
      "step": 11537
    },
    {
      "epoch": 0.8874701946004153,
      "grad_norm": 36.25,
      "learning_rate": 1.6687947080993772e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9909909963607788,
      "num_tokens": 5011144.0,
      "step": 11538
    },
    {
      "epoch": 0.8875471117606338,
      "grad_norm": 32.5,
      "learning_rate": 1.6686793323590494e-06,
      "loss": 0.0381,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 5011432.0,
      "step": 11539
    },
    {
      "epoch": 0.8876240289208522,
      "grad_norm": 2.375,
      "learning_rate": 1.6685639566187215e-06,
      "loss": 0.0065,
      "mean_token_accuracy": 0.9990167021751404,
      "num_tokens": 5012452.0,
      "step": 11540
    },
    {
      "epoch": 0.8877009460810706,
      "grad_norm": 14.25,
      "learning_rate": 1.668448580878394e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 5012798.0,
      "step": 11541
    },
    {
      "epoch": 0.8877778632412892,
      "grad_norm": 33.75,
      "learning_rate": 1.6683332051380664e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 5013108.0,
      "step": 11542
    },
    {
      "epoch": 0.8878547804015076,
      "grad_norm": 5.53125,
      "learning_rate": 1.6682178293977387e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9963964223861694,
      "num_tokens": 5013664.0,
      "step": 11543
    },
    {
      "epoch": 0.8879316975617261,
      "grad_norm": 12.5625,
      "learning_rate": 1.6681024536574111e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 5014222.0,
      "step": 11544
    },
    {
      "epoch": 0.8880086147219445,
      "grad_norm": 25.875,
      "learning_rate": 1.6679870779170834e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 5014822.0,
      "step": 11545
    },
    {
      "epoch": 0.8880855318821629,
      "grad_norm": 13.9375,
      "learning_rate": 1.6678717021767557e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 5015408.0,
      "step": 11546
    },
    {
      "epoch": 0.8881624490423814,
      "grad_norm": 14.0,
      "learning_rate": 1.667756326436428e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 5015904.0,
      "step": 11547
    },
    {
      "epoch": 0.8882393662025998,
      "grad_norm": 41.75,
      "learning_rate": 1.6676409506961004e-06,
      "loss": 0.0289,
      "mean_token_accuracy": 0.9860140085220337,
      "num_tokens": 5016192.0,
      "step": 11548
    },
    {
      "epoch": 0.8883162833628182,
      "grad_norm": 15.1875,
      "learning_rate": 1.6675255749557726e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.987261176109314,
      "num_tokens": 5016508.0,
      "step": 11549
    },
    {
      "epoch": 0.8883932005230367,
      "grad_norm": 41.75,
      "learning_rate": 1.6674101992154451e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9869706630706787,
      "num_tokens": 5016816.0,
      "step": 11550
    },
    {
      "epoch": 0.8884701176832551,
      "grad_norm": 36.25,
      "learning_rate": 1.6672948234751172e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9861111044883728,
      "num_tokens": 5017106.0,
      "step": 11551
    },
    {
      "epoch": 0.8885470348434735,
      "grad_norm": 18.0,
      "learning_rate": 1.6671794477347896e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 5017414.0,
      "step": 11552
    },
    {
      "epoch": 0.888623952003692,
      "grad_norm": 7.875,
      "learning_rate": 1.6670640719944619e-06,
      "loss": 0.0078,
      "mean_token_accuracy": 0.9987684488296509,
      "num_tokens": 5018228.0,
      "step": 11553
    },
    {
      "epoch": 0.8887008691639104,
      "grad_norm": 39.75,
      "learning_rate": 1.6669486962541344e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9881423115730286,
      "num_tokens": 5018482.0,
      "step": 11554
    },
    {
      "epoch": 0.888777786324129,
      "grad_norm": 34.75,
      "learning_rate": 1.6668333205138068e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5019068.0,
      "step": 11555
    },
    {
      "epoch": 0.8888547034843474,
      "grad_norm": 7.8125,
      "learning_rate": 1.666717944773479e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 5019328.0,
      "step": 11556
    },
    {
      "epoch": 0.8889316206445658,
      "grad_norm": 9.3125,
      "learning_rate": 1.6666025690331511e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 5019636.0,
      "step": 11557
    },
    {
      "epoch": 0.8890085378047843,
      "grad_norm": 6.65625,
      "learning_rate": 1.6664871932928236e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 5020182.0,
      "step": 11558
    },
    {
      "epoch": 0.8890854549650027,
      "grad_norm": 48.0,
      "learning_rate": 1.666371817552496e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 5020484.0,
      "step": 11559
    },
    {
      "epoch": 0.8891623721252211,
      "grad_norm": 27.5,
      "learning_rate": 1.6662564418121683e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9896373152732849,
      "num_tokens": 5021064.0,
      "step": 11560
    },
    {
      "epoch": 0.8892392892854396,
      "grad_norm": 24.125,
      "learning_rate": 1.6661410660718408e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 5021380.0,
      "step": 11561
    },
    {
      "epoch": 0.889316206445658,
      "grad_norm": 45.75,
      "learning_rate": 1.6660256903315128e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 5021696.0,
      "step": 11562
    },
    {
      "epoch": 0.8893931236058765,
      "grad_norm": 6.9375,
      "learning_rate": 1.6659103145911853e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9859648942947388,
      "num_tokens": 5021982.0,
      "step": 11563
    },
    {
      "epoch": 0.8894700407660949,
      "grad_norm": 17.5,
      "learning_rate": 1.6657949388508576e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 5022506.0,
      "step": 11564
    },
    {
      "epoch": 0.8895469579263133,
      "grad_norm": 27.75,
      "learning_rate": 1.66567956311053e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 5023054.0,
      "step": 11565
    },
    {
      "epoch": 0.8896238750865318,
      "grad_norm": 12.625,
      "learning_rate": 1.6655641873702025e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 5023578.0,
      "step": 11566
    },
    {
      "epoch": 0.8897007922467502,
      "grad_norm": 52.75,
      "learning_rate": 1.6654488116298746e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 5023902.0,
      "step": 11567
    },
    {
      "epoch": 0.8897777094069687,
      "grad_norm": 10.75,
      "learning_rate": 1.6653334358895468e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9948006868362427,
      "num_tokens": 5024480.0,
      "step": 11568
    },
    {
      "epoch": 0.8898546265671872,
      "grad_norm": 22.375,
      "learning_rate": 1.6652180601492193e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 5024788.0,
      "step": 11569
    },
    {
      "epoch": 0.8899315437274056,
      "grad_norm": 34.5,
      "learning_rate": 1.6651026844088918e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.9934425950050354,
      "num_tokens": 5025094.0,
      "step": 11570
    },
    {
      "epoch": 0.890008460887624,
      "grad_norm": 25.625,
      "learning_rate": 1.664987308668564e-06,
      "loss": 0.0254,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 5025418.0,
      "step": 11571
    },
    {
      "epoch": 0.8900853780478425,
      "grad_norm": 14.4375,
      "learning_rate": 1.6648719329282365e-06,
      "loss": 0.0229,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 5025740.0,
      "step": 11572
    },
    {
      "epoch": 0.8901622952080609,
      "grad_norm": 17.375,
      "learning_rate": 1.6647565571879085e-06,
      "loss": 0.0192,
      "mean_token_accuracy": 0.9940828680992126,
      "num_tokens": 5026080.0,
      "step": 11573
    },
    {
      "epoch": 0.8902392123682794,
      "grad_norm": 67.5,
      "learning_rate": 1.664641181447581e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 5026428.0,
      "step": 11574
    },
    {
      "epoch": 0.8903161295284978,
      "grad_norm": 9.0,
      "learning_rate": 1.6645258057072533e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9869281053543091,
      "num_tokens": 5026736.0,
      "step": 11575
    },
    {
      "epoch": 0.8903930466887162,
      "grad_norm": 8.3125,
      "learning_rate": 1.6644104299669257e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 5027276.0,
      "step": 11576
    },
    {
      "epoch": 0.8904699638489347,
      "grad_norm": 16.5,
      "learning_rate": 1.6642950542265982e-06,
      "loss": 0.0209,
      "mean_token_accuracy": 0.9938461780548096,
      "num_tokens": 5027602.0,
      "step": 11577
    },
    {
      "epoch": 0.8905468810091531,
      "grad_norm": 23.75,
      "learning_rate": 1.6641796784862702e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9896193742752075,
      "num_tokens": 5027892.0,
      "step": 11578
    },
    {
      "epoch": 0.8906237981693715,
      "grad_norm": 18.5,
      "learning_rate": 1.6640643027459425e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 5028442.0,
      "step": 11579
    },
    {
      "epoch": 0.89070071532959,
      "grad_norm": 70.0,
      "learning_rate": 1.663948927005615e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 5028798.0,
      "step": 11580
    },
    {
      "epoch": 0.8907776324898085,
      "grad_norm": 19.25,
      "learning_rate": 1.6638335512652874e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.9820788502693176,
      "num_tokens": 5029078.0,
      "step": 11581
    },
    {
      "epoch": 0.890854549650027,
      "grad_norm": 16.5,
      "learning_rate": 1.6637181755249597e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9911190271377563,
      "num_tokens": 5029642.0,
      "step": 11582
    },
    {
      "epoch": 0.8909314668102454,
      "grad_norm": 21.625,
      "learning_rate": 1.6636027997846322e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 5029950.0,
      "step": 11583
    },
    {
      "epoch": 0.8910083839704638,
      "grad_norm": 43.75,
      "learning_rate": 1.6634874240443042e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9943714737892151,
      "num_tokens": 5030484.0,
      "step": 11584
    },
    {
      "epoch": 0.8910853011306823,
      "grad_norm": 12.5,
      "learning_rate": 1.6633720483039767e-06,
      "loss": 0.0168,
      "mean_token_accuracy": 0.990791916847229,
      "num_tokens": 5031028.0,
      "step": 11585
    },
    {
      "epoch": 0.8911622182909007,
      "grad_norm": 57.75,
      "learning_rate": 1.663256672563649e-06,
      "loss": 0.0283,
      "mean_token_accuracy": 0.9856733679771423,
      "num_tokens": 5031378.0,
      "step": 11586
    },
    {
      "epoch": 0.8912391354511191,
      "grad_norm": 12.0,
      "learning_rate": 1.6631412968233214e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9928951859474182,
      "num_tokens": 5031942.0,
      "step": 11587
    },
    {
      "epoch": 0.8913160526113376,
      "grad_norm": 33.5,
      "learning_rate": 1.6630259210829939e-06,
      "loss": 0.0239,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 5032246.0,
      "step": 11588
    },
    {
      "epoch": 0.891392969771556,
      "grad_norm": 53.5,
      "learning_rate": 1.662910545342666e-06,
      "loss": 0.0317,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 5032528.0,
      "step": 11589
    },
    {
      "epoch": 0.8914698869317744,
      "grad_norm": 36.75,
      "learning_rate": 1.6627951696023382e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 5032812.0,
      "step": 11590
    },
    {
      "epoch": 0.8915468040919929,
      "grad_norm": 7.0625,
      "learning_rate": 1.6626797938620107e-06,
      "loss": 0.0088,
      "mean_token_accuracy": 0.9948586225509644,
      "num_tokens": 5033592.0,
      "step": 11591
    },
    {
      "epoch": 0.8916237212522113,
      "grad_norm": 11.875,
      "learning_rate": 1.6625644181216831e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9937304258346558,
      "num_tokens": 5033912.0,
      "step": 11592
    },
    {
      "epoch": 0.8917006384124299,
      "grad_norm": 55.75,
      "learning_rate": 1.6624490423813554e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 5034214.0,
      "step": 11593
    },
    {
      "epoch": 0.8917775555726483,
      "grad_norm": 17.5,
      "learning_rate": 1.6623336666410279e-06,
      "loss": 0.0199,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 5034554.0,
      "step": 11594
    },
    {
      "epoch": 0.8918544727328667,
      "grad_norm": 21.0,
      "learning_rate": 1.6622182909007e-06,
      "loss": 0.0067,
      "mean_token_accuracy": 0.9961575269699097,
      "num_tokens": 5035598.0,
      "step": 11595
    },
    {
      "epoch": 0.8919313898930852,
      "grad_norm": 41.0,
      "learning_rate": 1.6621029151603724e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 5036160.0,
      "step": 11596
    },
    {
      "epoch": 0.8920083070533036,
      "grad_norm": 83.5,
      "learning_rate": 1.6619875394200446e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 5036458.0,
      "step": 11597
    },
    {
      "epoch": 0.892085224213522,
      "grad_norm": 47.25,
      "learning_rate": 1.661872163679717e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9848484992980957,
      "num_tokens": 5036790.0,
      "step": 11598
    },
    {
      "epoch": 0.8921621413737405,
      "grad_norm": 19.125,
      "learning_rate": 1.6617567879393894e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 5037374.0,
      "step": 11599
    },
    {
      "epoch": 0.8922390585339589,
      "grad_norm": 10.9375,
      "learning_rate": 1.6616414121990616e-06,
      "loss": 0.0195,
      "mean_token_accuracy": 0.9969879388809204,
      "num_tokens": 5037708.0,
      "step": 11600
    },
    {
      "epoch": 0.8923159756941774,
      "grad_norm": 11.875,
      "learning_rate": 1.6615260364587339e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 5038292.0,
      "step": 11601
    },
    {
      "epoch": 0.8923928928543958,
      "grad_norm": 12.625,
      "learning_rate": 1.6614106607184063e-06,
      "loss": 0.0314,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 5038546.0,
      "step": 11602
    },
    {
      "epoch": 0.8924698100146142,
      "grad_norm": 10.75,
      "learning_rate": 1.6612952849780786e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9913644194602966,
      "num_tokens": 5039126.0,
      "step": 11603
    },
    {
      "epoch": 0.8925467271748327,
      "grad_norm": 53.75,
      "learning_rate": 1.661179909237751e-06,
      "loss": 0.0335,
      "mean_token_accuracy": 0.9836065769195557,
      "num_tokens": 5039432.0,
      "step": 11604
    },
    {
      "epoch": 0.8926236443350511,
      "grad_norm": 5.03125,
      "learning_rate": 1.6610645334974231e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 5040018.0,
      "step": 11605
    },
    {
      "epoch": 0.8927005614952696,
      "grad_norm": 46.25,
      "learning_rate": 1.6609491577570956e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9839228391647339,
      "num_tokens": 5040330.0,
      "step": 11606
    },
    {
      "epoch": 0.8927774786554881,
      "grad_norm": 14.5625,
      "learning_rate": 1.6608337820167678e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 5040886.0,
      "step": 11607
    },
    {
      "epoch": 0.8928543958157065,
      "grad_norm": 55.75,
      "learning_rate": 1.6607184062764403e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 5041182.0,
      "step": 11608
    },
    {
      "epoch": 0.8929313129759249,
      "grad_norm": 6.6875,
      "learning_rate": 1.6606030305361128e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 5041996.0,
      "step": 11609
    },
    {
      "epoch": 0.8930082301361434,
      "grad_norm": 49.5,
      "learning_rate": 1.660487654795785e-06,
      "loss": 0.0292,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 5042300.0,
      "step": 11610
    },
    {
      "epoch": 0.8930851472963618,
      "grad_norm": 20.5,
      "learning_rate": 1.660372279055457e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 5042612.0,
      "step": 11611
    },
    {
      "epoch": 0.8931620644565803,
      "grad_norm": 23.25,
      "learning_rate": 1.6602569033151296e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 5042888.0,
      "step": 11612
    },
    {
      "epoch": 0.8932389816167987,
      "grad_norm": 25.5,
      "learning_rate": 1.660141527574802e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9834983348846436,
      "num_tokens": 5043192.0,
      "step": 11613
    },
    {
      "epoch": 0.8933158987770171,
      "grad_norm": 33.5,
      "learning_rate": 1.6600261518344743e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9892086386680603,
      "num_tokens": 5043472.0,
      "step": 11614
    },
    {
      "epoch": 0.8933928159372356,
      "grad_norm": 64.0,
      "learning_rate": 1.6599107760941468e-06,
      "loss": 0.0435,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 5043810.0,
      "step": 11615
    },
    {
      "epoch": 0.893469733097454,
      "grad_norm": 47.0,
      "learning_rate": 1.6597954003538188e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 5044400.0,
      "step": 11616
    },
    {
      "epoch": 0.8935466502576724,
      "grad_norm": 19.375,
      "learning_rate": 1.6596800246134913e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 5044940.0,
      "step": 11617
    },
    {
      "epoch": 0.893623567417891,
      "grad_norm": 94.0,
      "learning_rate": 1.6595646488731635e-06,
      "loss": 0.0264,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 5045276.0,
      "step": 11618
    },
    {
      "epoch": 0.8937004845781094,
      "grad_norm": 59.5,
      "learning_rate": 1.659449273132836e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5045570.0,
      "step": 11619
    },
    {
      "epoch": 0.8937774017383279,
      "grad_norm": 19.125,
      "learning_rate": 1.6593338973925085e-06,
      "loss": 0.0208,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 5046154.0,
      "step": 11620
    },
    {
      "epoch": 0.8938543188985463,
      "grad_norm": 13.625,
      "learning_rate": 1.6592185216521807e-06,
      "loss": 0.0257,
      "mean_token_accuracy": 0.989051103591919,
      "num_tokens": 5046430.0,
      "step": 11621
    },
    {
      "epoch": 0.8939312360587647,
      "grad_norm": 44.75,
      "learning_rate": 1.6591031459118528e-06,
      "loss": 0.0238,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 5047026.0,
      "step": 11622
    },
    {
      "epoch": 0.8940081532189832,
      "grad_norm": 18.5,
      "learning_rate": 1.6589877701715252e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 5047566.0,
      "step": 11623
    },
    {
      "epoch": 0.8940850703792016,
      "grad_norm": 20.375,
      "learning_rate": 1.6588723944311977e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 5047876.0,
      "step": 11624
    },
    {
      "epoch": 0.89416198753942,
      "grad_norm": 13.75,
      "learning_rate": 1.65875701869087e-06,
      "loss": 0.0293,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 5048132.0,
      "step": 11625
    },
    {
      "epoch": 0.8942389046996385,
      "grad_norm": 25.625,
      "learning_rate": 1.6586416429505424e-06,
      "loss": 0.0251,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 5048414.0,
      "step": 11626
    },
    {
      "epoch": 0.8943158218598569,
      "grad_norm": 83.5,
      "learning_rate": 1.6585262672102145e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 5048774.0,
      "step": 11627
    },
    {
      "epoch": 0.8943927390200754,
      "grad_norm": 36.25,
      "learning_rate": 1.658410891469887e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9876161217689514,
      "num_tokens": 5049098.0,
      "step": 11628
    },
    {
      "epoch": 0.8944696561802938,
      "grad_norm": 36.75,
      "learning_rate": 1.6582955157295592e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 5049420.0,
      "step": 11629
    },
    {
      "epoch": 0.8945465733405122,
      "grad_norm": 59.5,
      "learning_rate": 1.6581801399892317e-06,
      "loss": 0.0298,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 5049744.0,
      "step": 11630
    },
    {
      "epoch": 0.8946234905007308,
      "grad_norm": 37.75,
      "learning_rate": 1.6580647642489042e-06,
      "loss": 0.0281,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 5050026.0,
      "step": 11631
    },
    {
      "epoch": 0.8947004076609492,
      "grad_norm": 8.5,
      "learning_rate": 1.6579493885085764e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 5050782.0,
      "step": 11632
    },
    {
      "epoch": 0.8947773248211676,
      "grad_norm": 12.25,
      "learning_rate": 1.6578340127682485e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9944547414779663,
      "num_tokens": 5051324.0,
      "step": 11633
    },
    {
      "epoch": 0.8948542419813861,
      "grad_norm": 36.0,
      "learning_rate": 1.657718637027921e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 5051616.0,
      "step": 11634
    },
    {
      "epoch": 0.8949311591416045,
      "grad_norm": 10.375,
      "learning_rate": 1.6576032612875934e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 5051962.0,
      "step": 11635
    },
    {
      "epoch": 0.8950080763018229,
      "grad_norm": 20.0,
      "learning_rate": 1.6574878855472657e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 5052548.0,
      "step": 11636
    },
    {
      "epoch": 0.8950849934620414,
      "grad_norm": 18.5,
      "learning_rate": 1.6573725098069381e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9969230890274048,
      "num_tokens": 5052874.0,
      "step": 11637
    },
    {
      "epoch": 0.8951619106222598,
      "grad_norm": 14.625,
      "learning_rate": 1.6572571340666102e-06,
      "loss": 0.021,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 5053196.0,
      "step": 11638
    },
    {
      "epoch": 0.8952388277824783,
      "grad_norm": 7.4375,
      "learning_rate": 1.6571417583262826e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 5053744.0,
      "step": 11639
    },
    {
      "epoch": 0.8953157449426967,
      "grad_norm": 20.375,
      "learning_rate": 1.657026382585955e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.9922928810119629,
      "num_tokens": 5054264.0,
      "step": 11640
    },
    {
      "epoch": 0.8953926621029151,
      "grad_norm": 24.625,
      "learning_rate": 1.6569110068456274e-06,
      "loss": 0.0348,
      "mean_token_accuracy": 0.9877675771713257,
      "num_tokens": 5054592.0,
      "step": 11641
    },
    {
      "epoch": 0.8954695792631336,
      "grad_norm": 14.3125,
      "learning_rate": 1.6567956311052998e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 5055150.0,
      "step": 11642
    },
    {
      "epoch": 0.895546496423352,
      "grad_norm": 12.8125,
      "learning_rate": 1.6566802553649719e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 5055736.0,
      "step": 11643
    },
    {
      "epoch": 0.8956234135835704,
      "grad_norm": 6.125,
      "learning_rate": 1.6565648796246442e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5056322.0,
      "step": 11644
    },
    {
      "epoch": 0.895700330743789,
      "grad_norm": 29.5,
      "learning_rate": 1.6564495038843166e-06,
      "loss": 0.0329,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 5056646.0,
      "step": 11645
    },
    {
      "epoch": 0.8957772479040074,
      "grad_norm": 28.875,
      "learning_rate": 1.656334128143989e-06,
      "loss": 0.0223,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 5056944.0,
      "step": 11646
    },
    {
      "epoch": 0.8958541650642259,
      "grad_norm": 19.125,
      "learning_rate": 1.6562187524036613e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 5057488.0,
      "step": 11647
    },
    {
      "epoch": 0.8959310822244443,
      "grad_norm": 49.0,
      "learning_rate": 1.6561033766633338e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9847908616065979,
      "num_tokens": 5057752.0,
      "step": 11648
    },
    {
      "epoch": 0.8960079993846627,
      "grad_norm": 6.15625,
      "learning_rate": 1.6559880009230059e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 5058292.0,
      "step": 11649
    },
    {
      "epoch": 0.8960849165448812,
      "grad_norm": 50.0,
      "learning_rate": 1.6558726251826783e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 5058592.0,
      "step": 11650
    },
    {
      "epoch": 0.8961618337050996,
      "grad_norm": 43.25,
      "learning_rate": 1.6557572494423506e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 5058914.0,
      "step": 11651
    },
    {
      "epoch": 0.896238750865318,
      "grad_norm": 47.5,
      "learning_rate": 1.655641873702023e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 5059212.0,
      "step": 11652
    },
    {
      "epoch": 0.8963156680255365,
      "grad_norm": 18.875,
      "learning_rate": 1.6555264979616953e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 5059796.0,
      "step": 11653
    },
    {
      "epoch": 0.8963925851857549,
      "grad_norm": 9.375,
      "learning_rate": 1.6554111222213676e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9926335215568542,
      "num_tokens": 5060340.0,
      "step": 11654
    },
    {
      "epoch": 0.8964695023459733,
      "grad_norm": 28.125,
      "learning_rate": 1.6552957464810398e-06,
      "loss": 0.0101,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5061122.0,
      "step": 11655
    },
    {
      "epoch": 0.8965464195061918,
      "grad_norm": 4.6875,
      "learning_rate": 1.6551803707407123e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 5061662.0,
      "step": 11656
    },
    {
      "epoch": 0.8966233366664103,
      "grad_norm": 7.125,
      "learning_rate": 1.6550649950003846e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.998236358165741,
      "num_tokens": 5062230.0,
      "step": 11657
    },
    {
      "epoch": 0.8967002538266288,
      "grad_norm": 50.0,
      "learning_rate": 1.654949619260057e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9886685609817505,
      "num_tokens": 5062584.0,
      "step": 11658
    },
    {
      "epoch": 0.8967771709868472,
      "grad_norm": 5.09375,
      "learning_rate": 1.6548342435197295e-06,
      "loss": 0.0212,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 5062892.0,
      "step": 11659
    },
    {
      "epoch": 0.8968540881470656,
      "grad_norm": 9.9375,
      "learning_rate": 1.6547188677794016e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9927927851676941,
      "num_tokens": 5063448.0,
      "step": 11660
    },
    {
      "epoch": 0.8969310053072841,
      "grad_norm": 19.0,
      "learning_rate": 1.6546034920390738e-06,
      "loss": 0.0114,
      "mean_token_accuracy": 0.9964093565940857,
      "num_tokens": 5064006.0,
      "step": 11661
    },
    {
      "epoch": 0.8970079224675025,
      "grad_norm": 14.4375,
      "learning_rate": 1.6544881162987463e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 5064552.0,
      "step": 11662
    },
    {
      "epoch": 0.8970848396277209,
      "grad_norm": 5.75,
      "learning_rate": 1.6543727405584187e-06,
      "loss": 0.0096,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 5065154.0,
      "step": 11663
    },
    {
      "epoch": 0.8971617567879394,
      "grad_norm": 28.125,
      "learning_rate": 1.654257364818091e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 5065486.0,
      "step": 11664
    },
    {
      "epoch": 0.8972386739481578,
      "grad_norm": 15.125,
      "learning_rate": 1.654141989077763e-06,
      "loss": 0.015,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 5066010.0,
      "step": 11665
    },
    {
      "epoch": 0.8973155911083763,
      "grad_norm": 7.15625,
      "learning_rate": 1.6540266133374355e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9946332573890686,
      "num_tokens": 5066570.0,
      "step": 11666
    },
    {
      "epoch": 0.8973925082685947,
      "grad_norm": 19.5,
      "learning_rate": 1.653911237597108e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.988304078578949,
      "num_tokens": 5066914.0,
      "step": 11667
    },
    {
      "epoch": 0.8974694254288131,
      "grad_norm": 18.875,
      "learning_rate": 1.6537958618567803e-06,
      "loss": 0.0271,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 5067254.0,
      "step": 11668
    },
    {
      "epoch": 0.8975463425890317,
      "grad_norm": 27.5,
      "learning_rate": 1.6536804861164527e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5067548.0,
      "step": 11669
    },
    {
      "epoch": 0.8976232597492501,
      "grad_norm": 11.0625,
      "learning_rate": 1.6535651103761248e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 5067848.0,
      "step": 11670
    },
    {
      "epoch": 0.8977001769094685,
      "grad_norm": 21.5,
      "learning_rate": 1.6534497346357972e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 5068130.0,
      "step": 11671
    },
    {
      "epoch": 0.897777094069687,
      "grad_norm": 43.5,
      "learning_rate": 1.6533343588954695e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 5068742.0,
      "step": 11672
    },
    {
      "epoch": 0.8978540112299054,
      "grad_norm": 24.75,
      "learning_rate": 1.653218983155142e-06,
      "loss": 0.0156,
      "mean_token_accuracy": 0.997032642364502,
      "num_tokens": 5069080.0,
      "step": 11673
    },
    {
      "epoch": 0.8979309283901238,
      "grad_norm": 11.0625,
      "learning_rate": 1.6531036074148144e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.9939393997192383,
      "num_tokens": 5069576.0,
      "step": 11674
    },
    {
      "epoch": 0.8980078455503423,
      "grad_norm": 4.90625,
      "learning_rate": 1.6529882316744867e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9961315393447876,
      "num_tokens": 5070094.0,
      "step": 11675
    },
    {
      "epoch": 0.8980847627105607,
      "grad_norm": 26.0,
      "learning_rate": 1.6528728559341587e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 5070614.0,
      "step": 11676
    },
    {
      "epoch": 0.8981616798707792,
      "grad_norm": 8.5625,
      "learning_rate": 1.6527574801938312e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 5070954.0,
      "step": 11677
    },
    {
      "epoch": 0.8982385970309976,
      "grad_norm": 22.5,
      "learning_rate": 1.6526421044535037e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 5071252.0,
      "step": 11678
    },
    {
      "epoch": 0.898315514191216,
      "grad_norm": 9.75,
      "learning_rate": 1.652526728713176e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9969969987869263,
      "num_tokens": 5071586.0,
      "step": 11679
    },
    {
      "epoch": 0.8983924313514345,
      "grad_norm": 34.25,
      "learning_rate": 1.6524113529728484e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 5072132.0,
      "step": 11680
    },
    {
      "epoch": 0.8984693485116529,
      "grad_norm": 45.5,
      "learning_rate": 1.6522959772325205e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 5072434.0,
      "step": 11681
    },
    {
      "epoch": 0.8985462656718713,
      "grad_norm": 21.375,
      "learning_rate": 1.652180601492193e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 5072716.0,
      "step": 11682
    },
    {
      "epoch": 0.8986231828320899,
      "grad_norm": 10.0625,
      "learning_rate": 1.6520652257518652e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9925788640975952,
      "num_tokens": 5073256.0,
      "step": 11683
    },
    {
      "epoch": 0.8987000999923083,
      "grad_norm": 15.3125,
      "learning_rate": 1.6519498500115377e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9947460889816284,
      "num_tokens": 5073828.0,
      "step": 11684
    },
    {
      "epoch": 0.8987770171525268,
      "grad_norm": 26.0,
      "learning_rate": 1.6518344742712101e-06,
      "loss": 0.0424,
      "mean_token_accuracy": 0.97826087474823,
      "num_tokens": 5074106.0,
      "step": 11685
    },
    {
      "epoch": 0.8988539343127452,
      "grad_norm": 19.0,
      "learning_rate": 1.6517190985308824e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 5074626.0,
      "step": 11686
    },
    {
      "epoch": 0.8989308514729636,
      "grad_norm": 50.75,
      "learning_rate": 1.6516037227905544e-06,
      "loss": 0.0296,
      "mean_token_accuracy": 0.9843260049819946,
      "num_tokens": 5074946.0,
      "step": 11687
    },
    {
      "epoch": 0.8990077686331821,
      "grad_norm": 61.0,
      "learning_rate": 1.651488347050227e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9932773113250732,
      "num_tokens": 5075542.0,
      "step": 11688
    },
    {
      "epoch": 0.8990846857934005,
      "grad_norm": 10.8125,
      "learning_rate": 1.6513729713098994e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9935064911842346,
      "num_tokens": 5075852.0,
      "step": 11689
    },
    {
      "epoch": 0.8991616029536189,
      "grad_norm": 25.0,
      "learning_rate": 1.6512575955695716e-06,
      "loss": 0.0345,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 5076108.0,
      "step": 11690
    },
    {
      "epoch": 0.8992385201138374,
      "grad_norm": 95.5,
      "learning_rate": 1.651142219829244e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.9887323975563049,
      "num_tokens": 5076464.0,
      "step": 11691
    },
    {
      "epoch": 0.8993154372740558,
      "grad_norm": 8.1875,
      "learning_rate": 1.6510268440889161e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9967637658119202,
      "num_tokens": 5076774.0,
      "step": 11692
    },
    {
      "epoch": 0.8993923544342742,
      "grad_norm": 14.9375,
      "learning_rate": 1.6509114683485886e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9975369572639465,
      "num_tokens": 5077588.0,
      "step": 11693
    },
    {
      "epoch": 0.8994692715944927,
      "grad_norm": 24.75,
      "learning_rate": 1.6507960926082609e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 5077904.0,
      "step": 11694
    },
    {
      "epoch": 0.8995461887547112,
      "grad_norm": 43.5,
      "learning_rate": 1.6506807168679333e-06,
      "loss": 0.0249,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5078198.0,
      "step": 11695
    },
    {
      "epoch": 0.8996231059149297,
      "grad_norm": 33.75,
      "learning_rate": 1.6505653411276058e-06,
      "loss": 0.031,
      "mean_token_accuracy": 0.9867987036705017,
      "num_tokens": 5078502.0,
      "step": 11696
    },
    {
      "epoch": 0.8997000230751481,
      "grad_norm": 21.125,
      "learning_rate": 1.650449965387278e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9927536249160767,
      "num_tokens": 5078780.0,
      "step": 11697
    },
    {
      "epoch": 0.8997769402353665,
      "grad_norm": 14.6875,
      "learning_rate": 1.6503345896469501e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 5079300.0,
      "step": 11698
    },
    {
      "epoch": 0.899853857395585,
      "grad_norm": 18.875,
      "learning_rate": 1.6502192139066226e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 5079618.0,
      "step": 11699
    },
    {
      "epoch": 0.8999307745558034,
      "grad_norm": 13.1875,
      "learning_rate": 1.650103838166295e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9968253970146179,
      "num_tokens": 5079936.0,
      "step": 11700
    },
    {
      "epoch": 0.9000076917160218,
      "grad_norm": 14.5,
      "learning_rate": 1.6499884624259673e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 5080484.0,
      "step": 11701
    },
    {
      "epoch": 0.9000846088762403,
      "grad_norm": 47.0,
      "learning_rate": 1.6498730866856398e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.988095223903656,
      "num_tokens": 5080822.0,
      "step": 11702
    },
    {
      "epoch": 0.9001615260364587,
      "grad_norm": 22.0,
      "learning_rate": 1.6497577109453118e-06,
      "loss": 0.0123,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 5081376.0,
      "step": 11703
    },
    {
      "epoch": 0.9002384431966772,
      "grad_norm": 13.75,
      "learning_rate": 1.6496423352049843e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 5081632.0,
      "step": 11704
    },
    {
      "epoch": 0.9003153603568956,
      "grad_norm": 6.28125,
      "learning_rate": 1.6495269594646566e-06,
      "loss": 0.0112,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 5082202.0,
      "step": 11705
    },
    {
      "epoch": 0.900392277517114,
      "grad_norm": 17.125,
      "learning_rate": 1.649411583724329e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 5082538.0,
      "step": 11706
    },
    {
      "epoch": 0.9004691946773326,
      "grad_norm": 26.375,
      "learning_rate": 1.6492962079840013e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 5083078.0,
      "step": 11707
    },
    {
      "epoch": 0.900546111837551,
      "grad_norm": 18.25,
      "learning_rate": 1.6491808322436735e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9936708807945251,
      "num_tokens": 5083396.0,
      "step": 11708
    },
    {
      "epoch": 0.9006230289977694,
      "grad_norm": 6.9375,
      "learning_rate": 1.6490654565033458e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5083982.0,
      "step": 11709
    },
    {
      "epoch": 0.9006999461579879,
      "grad_norm": 38.0,
      "learning_rate": 1.6489500807630183e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 5084268.0,
      "step": 11710
    },
    {
      "epoch": 0.9007768633182063,
      "grad_norm": 5.34375,
      "learning_rate": 1.6488347050226905e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5084854.0,
      "step": 11711
    },
    {
      "epoch": 0.9008537804784247,
      "grad_norm": 13.8125,
      "learning_rate": 1.648719329282363e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9908257126808167,
      "num_tokens": 5085182.0,
      "step": 11712
    },
    {
      "epoch": 0.9009306976386432,
      "grad_norm": 29.25,
      "learning_rate": 1.6486039535420355e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 5085526.0,
      "step": 11713
    },
    {
      "epoch": 0.9010076147988616,
      "grad_norm": 21.375,
      "learning_rate": 1.6484885778017075e-06,
      "loss": 0.0161,
      "mean_token_accuracy": 0.9894921183586121,
      "num_tokens": 5086098.0,
      "step": 11714
    },
    {
      "epoch": 0.9010845319590801,
      "grad_norm": 36.5,
      "learning_rate": 1.6483732020613798e-06,
      "loss": 0.0367,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 5086404.0,
      "step": 11715
    },
    {
      "epoch": 0.9011614491192985,
      "grad_norm": 14.375,
      "learning_rate": 1.6482578263210522e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 5086722.0,
      "step": 11716
    },
    {
      "epoch": 0.9012383662795169,
      "grad_norm": 13.4375,
      "learning_rate": 1.6481424505807247e-06,
      "loss": 0.0182,
      "mean_token_accuracy": 0.9942029118537903,
      "num_tokens": 5087068.0,
      "step": 11717
    },
    {
      "epoch": 0.9013152834397354,
      "grad_norm": 24.125,
      "learning_rate": 1.648027074840397e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9844357967376709,
      "num_tokens": 5087326.0,
      "step": 11718
    },
    {
      "epoch": 0.9013922005999538,
      "grad_norm": 27.375,
      "learning_rate": 1.647911699100069e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9820144176483154,
      "num_tokens": 5087606.0,
      "step": 11719
    },
    {
      "epoch": 0.9014691177601722,
      "grad_norm": 13.75,
      "learning_rate": 1.6477963233597415e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.994962215423584,
      "num_tokens": 5088402.0,
      "step": 11720
    },
    {
      "epoch": 0.9015460349203908,
      "grad_norm": 17.875,
      "learning_rate": 1.647680947619414e-06,
      "loss": 0.0157,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 5088898.0,
      "step": 11721
    },
    {
      "epoch": 0.9016229520806092,
      "grad_norm": 15.125,
      "learning_rate": 1.6475655718790862e-06,
      "loss": 0.0167,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 5089432.0,
      "step": 11722
    },
    {
      "epoch": 0.9016998692408277,
      "grad_norm": 73.5,
      "learning_rate": 1.6474501961387587e-06,
      "loss": 0.0315,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 5089738.0,
      "step": 11723
    },
    {
      "epoch": 0.9017767864010461,
      "grad_norm": 31.75,
      "learning_rate": 1.6473348203984312e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 5090300.0,
      "step": 11724
    },
    {
      "epoch": 0.9018537035612645,
      "grad_norm": 6.84375,
      "learning_rate": 1.6472194446581032e-06,
      "loss": 0.0143,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 5090824.0,
      "step": 11725
    },
    {
      "epoch": 0.901930620721483,
      "grad_norm": 22.875,
      "learning_rate": 1.6471040689177755e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9928825497627258,
      "num_tokens": 5091106.0,
      "step": 11726
    },
    {
      "epoch": 0.9020075378817014,
      "grad_norm": 12.75,
      "learning_rate": 1.646988693177448e-06,
      "loss": 0.0087,
      "mean_token_accuracy": 0.9947229623794556,
      "num_tokens": 5091866.0,
      "step": 11727
    },
    {
      "epoch": 0.9020844550419198,
      "grad_norm": 19.125,
      "learning_rate": 1.6468733174371204e-06,
      "loss": 0.0159,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 5092406.0,
      "step": 11728
    },
    {
      "epoch": 0.9021613722021383,
      "grad_norm": 7.3125,
      "learning_rate": 1.6467579416967927e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9929328560829163,
      "num_tokens": 5092690.0,
      "step": 11729
    },
    {
      "epoch": 0.9022382893623567,
      "grad_norm": 8.6875,
      "learning_rate": 1.6466425659564647e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9970059990882874,
      "num_tokens": 5093026.0,
      "step": 11730
    },
    {
      "epoch": 0.9023152065225752,
      "grad_norm": 15.625,
      "learning_rate": 1.6465271902161372e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 5093344.0,
      "step": 11731
    },
    {
      "epoch": 0.9023921236827936,
      "grad_norm": 13.9375,
      "learning_rate": 1.6464118144758096e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 5093600.0,
      "step": 11732
    },
    {
      "epoch": 0.902469040843012,
      "grad_norm": 75.0,
      "learning_rate": 1.646296438735482e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9862637519836426,
      "num_tokens": 5093966.0,
      "step": 11733
    },
    {
      "epoch": 0.9025459580032306,
      "grad_norm": 36.5,
      "learning_rate": 1.6461810629951544e-06,
      "loss": 0.0272,
      "mean_token_accuracy": 0.9814241528511047,
      "num_tokens": 5094290.0,
      "step": 11734
    },
    {
      "epoch": 0.902622875163449,
      "grad_norm": 15.375,
      "learning_rate": 1.6460656872548268e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9890710115432739,
      "num_tokens": 5094840.0,
      "step": 11735
    },
    {
      "epoch": 0.9026997923236674,
      "grad_norm": 9.0,
      "learning_rate": 1.6459503115144989e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 5095110.0,
      "step": 11736
    },
    {
      "epoch": 0.9027767094838859,
      "grad_norm": 34.5,
      "learning_rate": 1.6458349357741711e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 5095426.0,
      "step": 11737
    },
    {
      "epoch": 0.9028536266441043,
      "grad_norm": 3.328125,
      "learning_rate": 1.6457195600338436e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 5095972.0,
      "step": 11738
    },
    {
      "epoch": 0.9029305438043227,
      "grad_norm": 7.40625,
      "learning_rate": 1.645604184293516e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 5096494.0,
      "step": 11739
    },
    {
      "epoch": 0.9030074609645412,
      "grad_norm": 40.0,
      "learning_rate": 1.6454888085531883e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 5096822.0,
      "step": 11740
    },
    {
      "epoch": 0.9030843781247596,
      "grad_norm": 24.5,
      "learning_rate": 1.6453734328128604e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 5097396.0,
      "step": 11741
    },
    {
      "epoch": 0.9031612952849781,
      "grad_norm": 22.375,
      "learning_rate": 1.6452580570725329e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 5097944.0,
      "step": 11742
    },
    {
      "epoch": 0.9032382124451965,
      "grad_norm": 20.5,
      "learning_rate": 1.6451426813322053e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 5098260.0,
      "step": 11743
    },
    {
      "epoch": 0.9033151296054149,
      "grad_norm": 22.875,
      "learning_rate": 1.6450273055918776e-06,
      "loss": 0.0324,
      "mean_token_accuracy": 0.9882352948188782,
      "num_tokens": 5098516.0,
      "step": 11744
    },
    {
      "epoch": 0.9033920467656334,
      "grad_norm": 38.75,
      "learning_rate": 1.64491192985155e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5098810.0,
      "step": 11745
    },
    {
      "epoch": 0.9034689639258519,
      "grad_norm": 20.25,
      "learning_rate": 1.644796554111222e-06,
      "loss": 0.016,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 5099178.0,
      "step": 11746
    },
    {
      "epoch": 0.9035458810860703,
      "grad_norm": 28.125,
      "learning_rate": 1.6446811783708946e-06,
      "loss": 0.0287,
      "mean_token_accuracy": 0.9830508232116699,
      "num_tokens": 5099474.0,
      "step": 11747
    },
    {
      "epoch": 0.9036227982462888,
      "grad_norm": 3.328125,
      "learning_rate": 1.6445658026305668e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9963833689689636,
      "num_tokens": 5100028.0,
      "step": 11748
    },
    {
      "epoch": 0.9036997154065072,
      "grad_norm": 21.0,
      "learning_rate": 1.6444504268902393e-06,
      "loss": 0.0278,
      "mean_token_accuracy": 0.9878787994384766,
      "num_tokens": 5100360.0,
      "step": 11749
    },
    {
      "epoch": 0.9037766325667257,
      "grad_norm": 12.625,
      "learning_rate": 1.6443350511499118e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.9949579834938049,
      "num_tokens": 5100956.0,
      "step": 11750
    },
    {
      "epoch": 0.9038535497269441,
      "grad_norm": 9.75,
      "learning_rate": 1.644219675409584e-06,
      "loss": 0.02,
      "mean_token_accuracy": 0.993220329284668,
      "num_tokens": 5101252.0,
      "step": 11751
    },
    {
      "epoch": 0.9039304668871625,
      "grad_norm": 6.0,
      "learning_rate": 1.644104299669256e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9965096116065979,
      "num_tokens": 5101826.0,
      "step": 11752
    },
    {
      "epoch": 0.904007384047381,
      "grad_norm": 9.75,
      "learning_rate": 1.6439889239289285e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9879518151283264,
      "num_tokens": 5102160.0,
      "step": 11753
    },
    {
      "epoch": 0.9040843012075994,
      "grad_norm": 9.125,
      "learning_rate": 1.643873548188601e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.991525411605835,
      "num_tokens": 5102516.0,
      "step": 11754
    },
    {
      "epoch": 0.9041612183678178,
      "grad_norm": 33.25,
      "learning_rate": 1.6437581724482733e-06,
      "loss": 0.0266,
      "mean_token_accuracy": 0.9893238544464111,
      "num_tokens": 5102798.0,
      "step": 11755
    },
    {
      "epoch": 0.9042381355280363,
      "grad_norm": 10.6875,
      "learning_rate": 1.6436427967079457e-06,
      "loss": 0.0091,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 5103606.0,
      "step": 11756
    },
    {
      "epoch": 0.9043150526882547,
      "grad_norm": 19.5,
      "learning_rate": 1.6435274209676178e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 5103922.0,
      "step": 11757
    },
    {
      "epoch": 0.9043919698484731,
      "grad_norm": 15.5,
      "learning_rate": 1.6434120452272903e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 5104502.0,
      "step": 11758
    },
    {
      "epoch": 0.9044688870086917,
      "grad_norm": 18.5,
      "learning_rate": 1.6432966694869625e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 5105070.0,
      "step": 11759
    },
    {
      "epoch": 0.9045458041689101,
      "grad_norm": 10.875,
      "learning_rate": 1.643181293746635e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9968847632408142,
      "num_tokens": 5105392.0,
      "step": 11760
    },
    {
      "epoch": 0.9046227213291286,
      "grad_norm": 14.875,
      "learning_rate": 1.6430659180063072e-06,
      "loss": 0.0176,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 5105740.0,
      "step": 11761
    },
    {
      "epoch": 0.904699638489347,
      "grad_norm": 7.3125,
      "learning_rate": 1.6429505422659797e-06,
      "loss": 0.0165,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 5106262.0,
      "step": 11762
    },
    {
      "epoch": 0.9047765556495654,
      "grad_norm": 5.375,
      "learning_rate": 1.6428351665256518e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9927667379379272,
      "num_tokens": 5106816.0,
      "step": 11763
    },
    {
      "epoch": 0.9048534728097839,
      "grad_norm": 10.4375,
      "learning_rate": 1.6427197907853242e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9972066879272461,
      "num_tokens": 5107176.0,
      "step": 11764
    },
    {
      "epoch": 0.9049303899700023,
      "grad_norm": 50.25,
      "learning_rate": 1.6426044150449965e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 5107432.0,
      "step": 11765
    },
    {
      "epoch": 0.9050073071302207,
      "grad_norm": 35.25,
      "learning_rate": 1.642489039304669e-06,
      "loss": 0.0285,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 5107744.0,
      "step": 11766
    },
    {
      "epoch": 0.9050842242904392,
      "grad_norm": 30.125,
      "learning_rate": 1.6423736635643414e-06,
      "loss": 0.0169,
      "mean_token_accuracy": 0.9918864369392395,
      "num_tokens": 5108238.0,
      "step": 11767
    },
    {
      "epoch": 0.9051611414506576,
      "grad_norm": 24.375,
      "learning_rate": 1.6422582878240135e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 5108756.0,
      "step": 11768
    },
    {
      "epoch": 0.9052380586108761,
      "grad_norm": 46.0,
      "learning_rate": 1.6421429120836857e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 5109318.0,
      "step": 11769
    },
    {
      "epoch": 0.9053149757710945,
      "grad_norm": 13.5,
      "learning_rate": 1.6420275363433582e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9941002726554871,
      "num_tokens": 5109658.0,
      "step": 11770
    },
    {
      "epoch": 0.905391892931313,
      "grad_norm": 48.0,
      "learning_rate": 1.6419121606030307e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9841772317886353,
      "num_tokens": 5109976.0,
      "step": 11771
    },
    {
      "epoch": 0.9054688100915315,
      "grad_norm": 10.125,
      "learning_rate": 1.641796784862703e-06,
      "loss": 0.0137,
      "mean_token_accuracy": 0.9923518300056458,
      "num_tokens": 5110500.0,
      "step": 11772
    },
    {
      "epoch": 0.9055457272517499,
      "grad_norm": 22.75,
      "learning_rate": 1.6416814091223754e-06,
      "loss": 0.0203,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5110794.0,
      "step": 11773
    },
    {
      "epoch": 0.9056226444119683,
      "grad_norm": 2.578125,
      "learning_rate": 1.6415660333820474e-06,
      "loss": 0.01,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 5111334.0,
      "step": 11774
    },
    {
      "epoch": 0.9056995615721868,
      "grad_norm": 9.0,
      "learning_rate": 1.64145065764172e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9860627055168152,
      "num_tokens": 5111622.0,
      "step": 11775
    },
    {
      "epoch": 0.9057764787324052,
      "grad_norm": 15.6875,
      "learning_rate": 1.6413352819013922e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.990227997303009,
      "num_tokens": 5111930.0,
      "step": 11776
    },
    {
      "epoch": 0.9058533958926236,
      "grad_norm": 9.3125,
      "learning_rate": 1.6412199061610646e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.9902912378311157,
      "num_tokens": 5112240.0,
      "step": 11777
    },
    {
      "epoch": 0.9059303130528421,
      "grad_norm": 9.6875,
      "learning_rate": 1.6411045304207371e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9975369572639465,
      "num_tokens": 5113054.0,
      "step": 11778
    },
    {
      "epoch": 0.9060072302130605,
      "grad_norm": 12.4375,
      "learning_rate": 1.6409891546804092e-06,
      "loss": 0.036,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 5113308.0,
      "step": 11779
    },
    {
      "epoch": 0.906084147373279,
      "grad_norm": 34.5,
      "learning_rate": 1.6408737789400814e-06,
      "loss": 0.0288,
      "mean_token_accuracy": 0.9893993139266968,
      "num_tokens": 5113592.0,
      "step": 11780
    },
    {
      "epoch": 0.9061610645334974,
      "grad_norm": 26.875,
      "learning_rate": 1.6407584031997539e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 5114114.0,
      "step": 11781
    },
    {
      "epoch": 0.9062379816937158,
      "grad_norm": 8.875,
      "learning_rate": 1.6406430274594264e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 5114688.0,
      "step": 11782
    },
    {
      "epoch": 0.9063148988539343,
      "grad_norm": 15.6875,
      "learning_rate": 1.6405276517190986e-06,
      "loss": 0.0163,
      "mean_token_accuracy": 0.9882006049156189,
      "num_tokens": 5115030.0,
      "step": 11783
    },
    {
      "epoch": 0.9063918160141528,
      "grad_norm": 23.375,
      "learning_rate": 1.6404122759787707e-06,
      "loss": 0.0179,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 5115382.0,
      "step": 11784
    },
    {
      "epoch": 0.9064687331743712,
      "grad_norm": 22.125,
      "learning_rate": 1.6402969002384431e-06,
      "loss": 0.0316,
      "mean_token_accuracy": 0.984674334526062,
      "num_tokens": 5115644.0,
      "step": 11785
    },
    {
      "epoch": 0.9065456503345897,
      "grad_norm": 25.375,
      "learning_rate": 1.6401815244981156e-06,
      "loss": 0.0356,
      "mean_token_accuracy": 0.9817517995834351,
      "num_tokens": 5115920.0,
      "step": 11786
    },
    {
      "epoch": 0.9066225674948081,
      "grad_norm": 38.75,
      "learning_rate": 1.6400661487577879e-06,
      "loss": 0.018,
      "mean_token_accuracy": 0.994301974773407,
      "num_tokens": 5116272.0,
      "step": 11787
    },
    {
      "epoch": 0.9066994846550266,
      "grad_norm": 41.5,
      "learning_rate": 1.6399507730174603e-06,
      "loss": 0.0243,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 5116606.0,
      "step": 11788
    },
    {
      "epoch": 0.906776401815245,
      "grad_norm": 50.75,
      "learning_rate": 1.6398353972771328e-06,
      "loss": 0.025,
      "mean_token_accuracy": 0.9937694668769836,
      "num_tokens": 5116928.0,
      "step": 11789
    },
    {
      "epoch": 0.9068533189754634,
      "grad_norm": 50.5,
      "learning_rate": 1.6397200215368048e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 5117528.0,
      "step": 11790
    },
    {
      "epoch": 0.9069302361356819,
      "grad_norm": 32.75,
      "learning_rate": 1.639604645796477e-06,
      "loss": 0.027,
      "mean_token_accuracy": 0.9885057210922241,
      "num_tokens": 5117790.0,
      "step": 11791
    },
    {
      "epoch": 0.9070071532959003,
      "grad_norm": 10.5625,
      "learning_rate": 1.6394892700561496e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 5118330.0,
      "step": 11792
    },
    {
      "epoch": 0.9070840704561187,
      "grad_norm": 14.0,
      "learning_rate": 1.639373894315822e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9950082898139954,
      "num_tokens": 5118932.0,
      "step": 11793
    },
    {
      "epoch": 0.9071609876163372,
      "grad_norm": 8.0625,
      "learning_rate": 1.6392585185754943e-06,
      "loss": 0.0153,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 5119466.0,
      "step": 11794
    },
    {
      "epoch": 0.9072379047765556,
      "grad_norm": 10.0,
      "learning_rate": 1.6391431428351663e-06,
      "loss": 0.0145,
      "mean_token_accuracy": 0.9910873174667358,
      "num_tokens": 5120028.0,
      "step": 11795
    },
    {
      "epoch": 0.907314821936774,
      "grad_norm": 7.25,
      "learning_rate": 1.6390277670948388e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.9975062608718872,
      "num_tokens": 5120832.0,
      "step": 11796
    },
    {
      "epoch": 0.9073917390969926,
      "grad_norm": 16.125,
      "learning_rate": 1.6389123913545113e-06,
      "loss": 0.0107,
      "mean_token_accuracy": 0.9934533834457397,
      "num_tokens": 5121444.0,
      "step": 11797
    },
    {
      "epoch": 0.907468656257211,
      "grad_norm": 27.625,
      "learning_rate": 1.6387970156141835e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9915013909339905,
      "num_tokens": 5121798.0,
      "step": 11798
    },
    {
      "epoch": 0.9075455734174295,
      "grad_norm": 10.0,
      "learning_rate": 1.638681639873856e-06,
      "loss": 0.0095,
      "mean_token_accuracy": 0.9950617551803589,
      "num_tokens": 5122610.0,
      "step": 11799
    },
    {
      "epoch": 0.9076224905776479,
      "grad_norm": 46.0,
      "learning_rate": 1.6385662641335285e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 5122908.0,
      "step": 11800
    },
    {
      "epoch": 0.9076994077378663,
      "grad_norm": 5.0625,
      "learning_rate": 1.6384508883932005e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9987179636955261,
      "num_tokens": 5123690.0,
      "step": 11801
    },
    {
      "epoch": 0.9077763248980848,
      "grad_norm": 76.5,
      "learning_rate": 1.6383355126528728e-06,
      "loss": 0.0339,
      "mean_token_accuracy": 0.9840637445449829,
      "num_tokens": 5123942.0,
      "step": 11802
    },
    {
      "epoch": 0.9078532420583032,
      "grad_norm": 6.90625,
      "learning_rate": 1.6382201369125453e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5124528.0,
      "step": 11803
    },
    {
      "epoch": 0.9079301592185216,
      "grad_norm": 20.375,
      "learning_rate": 1.6381047611722177e-06,
      "loss": 0.0133,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 5125074.0,
      "step": 11804
    },
    {
      "epoch": 0.9080070763787401,
      "grad_norm": 9.6875,
      "learning_rate": 1.63798938543189e-06,
      "loss": 0.0142,
      "mean_token_accuracy": 0.9943925142288208,
      "num_tokens": 5125610.0,
      "step": 11805
    },
    {
      "epoch": 0.9080839935389585,
      "grad_norm": 19.875,
      "learning_rate": 1.637874009691562e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.989830493927002,
      "num_tokens": 5125906.0,
      "step": 11806
    },
    {
      "epoch": 0.908160910699177,
      "grad_norm": 17.0,
      "learning_rate": 1.6377586339512345e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9964476227760315,
      "num_tokens": 5126470.0,
      "step": 11807
    },
    {
      "epoch": 0.9082378278593954,
      "grad_norm": 47.75,
      "learning_rate": 1.637643258210907e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9884058237075806,
      "num_tokens": 5126818.0,
      "step": 11808
    },
    {
      "epoch": 0.9083147450196138,
      "grad_norm": 6.03125,
      "learning_rate": 1.6375278824705792e-06,
      "loss": 0.0079,
      "mean_token_accuracy": 0.9973545074462891,
      "num_tokens": 5127576.0,
      "step": 11809
    },
    {
      "epoch": 0.9083916621798324,
      "grad_norm": 12.375,
      "learning_rate": 1.6374125067302517e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9946714043617249,
      "num_tokens": 5128140.0,
      "step": 11810
    },
    {
      "epoch": 0.9084685793400508,
      "grad_norm": 22.75,
      "learning_rate": 1.6372971309899237e-06,
      "loss": 0.0141,
      "mean_token_accuracy": 0.9947643876075745,
      "num_tokens": 5128714.0,
      "step": 11811
    },
    {
      "epoch": 0.9085454965002692,
      "grad_norm": 17.25,
      "learning_rate": 1.6371817552495962e-06,
      "loss": 0.0172,
      "mean_token_accuracy": 0.9944444298744202,
      "num_tokens": 5129076.0,
      "step": 11812
    },
    {
      "epoch": 0.9086224136604877,
      "grad_norm": 16.5,
      "learning_rate": 1.6370663795092685e-06,
      "loss": 0.0106,
      "mean_token_accuracy": 0.9982425570487976,
      "num_tokens": 5129646.0,
      "step": 11813
    },
    {
      "epoch": 0.9086993308207061,
      "grad_norm": 20.375,
      "learning_rate": 1.636951003768941e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9975550174713135,
      "num_tokens": 5130466.0,
      "step": 11814
    },
    {
      "epoch": 0.9087762479809245,
      "grad_norm": 24.0,
      "learning_rate": 1.6368356280286132e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9900990128517151,
      "num_tokens": 5130770.0,
      "step": 11815
    },
    {
      "epoch": 0.908853165141143,
      "grad_norm": 12.25,
      "learning_rate": 1.6367202522882857e-06,
      "loss": 0.0152,
      "mean_token_accuracy": 0.9916201233863831,
      "num_tokens": 5131130.0,
      "step": 11816
    },
    {
      "epoch": 0.9089300823013614,
      "grad_norm": 39.75,
      "learning_rate": 1.6366048765479577e-06,
      "loss": 0.0247,
      "mean_token_accuracy": 0.9894737005233765,
      "num_tokens": 5131416.0,
      "step": 11817
    },
    {
      "epoch": 0.9090069994615799,
      "grad_norm": 3.625,
      "learning_rate": 1.6364895008076302e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 5131978.0,
      "step": 11818
    },
    {
      "epoch": 0.9090839166217983,
      "grad_norm": 49.0,
      "learning_rate": 1.6363741250673024e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 5132284.0,
      "step": 11819
    },
    {
      "epoch": 0.9091608337820167,
      "grad_norm": 14.9375,
      "learning_rate": 1.636258749326975e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.994727611541748,
      "num_tokens": 5132854.0,
      "step": 11820
    },
    {
      "epoch": 0.9092377509422352,
      "grad_norm": 44.25,
      "learning_rate": 1.6361433735866474e-06,
      "loss": 0.0221,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 5133184.0,
      "step": 11821
    },
    {
      "epoch": 0.9093146681024536,
      "grad_norm": 24.375,
      "learning_rate": 1.6360279978463194e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 5133518.0,
      "step": 11822
    },
    {
      "epoch": 0.909391585262672,
      "grad_norm": 25.75,
      "learning_rate": 1.6359126221059917e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 5134036.0,
      "step": 11823
    },
    {
      "epoch": 0.9094685024228906,
      "grad_norm": 10.0625,
      "learning_rate": 1.6357972463656642e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 5134582.0,
      "step": 11824
    },
    {
      "epoch": 0.909545419583109,
      "grad_norm": 39.25,
      "learning_rate": 1.6356818706253366e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.9914039969444275,
      "num_tokens": 5134932.0,
      "step": 11825
    },
    {
      "epoch": 0.9096223367433275,
      "grad_norm": 53.0,
      "learning_rate": 1.6355664948850089e-06,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9799330830574036,
      "num_tokens": 5135234.0,
      "step": 11826
    },
    {
      "epoch": 0.9096992539035459,
      "grad_norm": 28.625,
      "learning_rate": 1.6354511191446814e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9942638874053955,
      "num_tokens": 5135758.0,
      "step": 11827
    },
    {
      "epoch": 0.9097761710637643,
      "grad_norm": 17.375,
      "learning_rate": 1.6353357434043534e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946949481964111,
      "num_tokens": 5136514.0,
      "step": 11828
    },
    {
      "epoch": 0.9098530882239828,
      "grad_norm": 34.5,
      "learning_rate": 1.6352203676640259e-06,
      "loss": 0.0207,
      "mean_token_accuracy": 0.9881656765937805,
      "num_tokens": 5136854.0,
      "step": 11829
    },
    {
      "epoch": 0.9099300053842012,
      "grad_norm": 9.0,
      "learning_rate": 1.6351049919236981e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 5137422.0,
      "step": 11830
    },
    {
      "epoch": 0.9100069225444196,
      "grad_norm": 28.5,
      "learning_rate": 1.6349896161833706e-06,
      "loss": 0.0228,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 5137744.0,
      "step": 11831
    },
    {
      "epoch": 0.9100838397046381,
      "grad_norm": 18.625,
      "learning_rate": 1.634874240443043e-06,
      "loss": 0.0146,
      "mean_token_accuracy": 0.9945553541183472,
      "num_tokens": 5138296.0,
      "step": 11832
    },
    {
      "epoch": 0.9101607568648565,
      "grad_norm": 63.25,
      "learning_rate": 1.6347588647027151e-06,
      "loss": 0.0441,
      "mean_token_accuracy": 0.9837662577629089,
      "num_tokens": 5138606.0,
      "step": 11833
    },
    {
      "epoch": 0.9102376740250749,
      "grad_norm": 25.625,
      "learning_rate": 1.6346434889623874e-06,
      "loss": 0.0322,
      "mean_token_accuracy": 0.9863481521606445,
      "num_tokens": 5138900.0,
      "step": 11834
    },
    {
      "epoch": 0.9103145911852935,
      "grad_norm": 38.5,
      "learning_rate": 1.6345281132220598e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.9888268113136292,
      "num_tokens": 5139260.0,
      "step": 11835
    },
    {
      "epoch": 0.9103915083455119,
      "grad_norm": 36.25,
      "learning_rate": 1.6344127374817323e-06,
      "loss": 0.0155,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 5139782.0,
      "step": 11836
    },
    {
      "epoch": 0.9104684255057304,
      "grad_norm": 76.0,
      "learning_rate": 1.6342973617414046e-06,
      "loss": 0.0413,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 5140084.0,
      "step": 11837
    },
    {
      "epoch": 0.9105453426659488,
      "grad_norm": 47.0,
      "learning_rate": 1.634181986001077e-06,
      "loss": 0.0189,
      "mean_token_accuracy": 0.9941691160202026,
      "num_tokens": 5140428.0,
      "step": 11838
    },
    {
      "epoch": 0.9106222598261672,
      "grad_norm": 28.375,
      "learning_rate": 1.634066610260749e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.996497392654419,
      "num_tokens": 5141000.0,
      "step": 11839
    },
    {
      "epoch": 0.9106991769863857,
      "grad_norm": 4.5625,
      "learning_rate": 1.6339512345204216e-06,
      "loss": 0.0111,
      "mean_token_accuracy": 0.9962894320487976,
      "num_tokens": 5141540.0,
      "step": 11840
    },
    {
      "epoch": 0.9107760941466041,
      "grad_norm": 56.25,
      "learning_rate": 1.6338358587800938e-06,
      "loss": 0.029,
      "mean_token_accuracy": 0.9870129823684692,
      "num_tokens": 5141850.0,
      "step": 11841
    },
    {
      "epoch": 0.9108530113068225,
      "grad_norm": 12.8125,
      "learning_rate": 1.6337204830397663e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 5142396.0,
      "step": 11842
    },
    {
      "epoch": 0.910929928467041,
      "grad_norm": 17.5,
      "learning_rate": 1.6336051072994388e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9961165189743042,
      "num_tokens": 5142912.0,
      "step": 11843
    },
    {
      "epoch": 0.9110068456272594,
      "grad_norm": 13.5,
      "learning_rate": 1.6334897315591108e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 5143166.0,
      "step": 11844
    },
    {
      "epoch": 0.9110837627874779,
      "grad_norm": 64.0,
      "learning_rate": 1.633374355818783e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.9966611266136169,
      "num_tokens": 5143766.0,
      "step": 11845
    },
    {
      "epoch": 0.9111606799476963,
      "grad_norm": 51.25,
      "learning_rate": 1.6332589800784555e-06,
      "loss": 0.024,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 5144082.0,
      "step": 11846
    },
    {
      "epoch": 0.9112375971079147,
      "grad_norm": 23.5,
      "learning_rate": 1.633143604338128e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 5144604.0,
      "step": 11847
    },
    {
      "epoch": 0.9113145142681333,
      "grad_norm": 10.3125,
      "learning_rate": 1.6330282285978003e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 5145188.0,
      "step": 11848
    },
    {
      "epoch": 0.9113914314283517,
      "grad_norm": 38.0,
      "learning_rate": 1.6329128528574723e-06,
      "loss": 0.0279,
      "mean_token_accuracy": 0.9841269850730896,
      "num_tokens": 5145504.0,
      "step": 11849
    },
    {
      "epoch": 0.9114683485885701,
      "grad_norm": 61.25,
      "learning_rate": 1.6327974771171448e-06,
      "loss": 0.034,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 5145786.0,
      "step": 11850
    },
    {
      "epoch": 0.9115452657487886,
      "grad_norm": 28.375,
      "learning_rate": 1.6326821013768172e-06,
      "loss": 0.0276,
      "mean_token_accuracy": 0.9887005686759949,
      "num_tokens": 5146142.0,
      "step": 11851
    },
    {
      "epoch": 0.911622182909007,
      "grad_norm": 5.375,
      "learning_rate": 1.6325667256364895e-06,
      "loss": 0.0126,
      "mean_token_accuracy": 0.9945155382156372,
      "num_tokens": 5146690.0,
      "step": 11852
    },
    {
      "epoch": 0.9116991000692255,
      "grad_norm": 19.875,
      "learning_rate": 1.632451349896162e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 5147236.0,
      "step": 11853
    },
    {
      "epoch": 0.9117760172294439,
      "grad_norm": 4.0,
      "learning_rate": 1.6323359741558344e-06,
      "loss": 0.0127,
      "mean_token_accuracy": 0.994575023651123,
      "num_tokens": 5147790.0,
      "step": 11854
    },
    {
      "epoch": 0.9118529343896623,
      "grad_norm": 40.75,
      "learning_rate": 1.6322205984155065e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9942362904548645,
      "num_tokens": 5148138.0,
      "step": 11855
    },
    {
      "epoch": 0.9119298515498808,
      "grad_norm": 20.25,
      "learning_rate": 1.6321052226751788e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9858657121658325,
      "num_tokens": 5148422.0,
      "step": 11856
    },
    {
      "epoch": 0.9120067687100992,
      "grad_norm": 16.875,
      "learning_rate": 1.6319898469348512e-06,
      "loss": 0.0135,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 5148942.0,
      "step": 11857
    },
    {
      "epoch": 0.9120836858703176,
      "grad_norm": 69.5,
      "learning_rate": 1.6318744711945237e-06,
      "loss": 0.017,
      "mean_token_accuracy": 0.9948365092277527,
      "num_tokens": 5149524.0,
      "step": 11858
    },
    {
      "epoch": 0.9121606030305361,
      "grad_norm": 12.75,
      "learning_rate": 1.631759095454196e-06,
      "loss": 0.0205,
      "mean_token_accuracy": 0.991150438785553,
      "num_tokens": 5149866.0,
      "step": 11859
    },
    {
      "epoch": 0.9122375201907545,
      "grad_norm": 7.59375,
      "learning_rate": 1.631643719713868e-06,
      "loss": 0.0105,
      "mean_token_accuracy": 0.9963302612304688,
      "num_tokens": 5150412.0,
      "step": 11860
    },
    {
      "epoch": 0.912314437350973,
      "grad_norm": 11.375,
      "learning_rate": 1.6315283439735405e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.996277928352356,
      "num_tokens": 5151220.0,
      "step": 11861
    },
    {
      "epoch": 0.9123913545111915,
      "grad_norm": 89.0,
      "learning_rate": 1.631412968233213e-06,
      "loss": 0.0294,
      "mean_token_accuracy": 0.9878419637680054,
      "num_tokens": 5151550.0,
      "step": 11862
    },
    {
      "epoch": 0.9124682716714099,
      "grad_norm": 8.0,
      "learning_rate": 1.6312975924928852e-06,
      "loss": 0.019,
      "mean_token_accuracy": 0.997032642364502,
      "num_tokens": 5151888.0,
      "step": 11863
    },
    {
      "epoch": 0.9125451888316284,
      "grad_norm": 30.0,
      "learning_rate": 1.6311822167525577e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9847095012664795,
      "num_tokens": 5152216.0,
      "step": 11864
    },
    {
      "epoch": 0.9126221059918468,
      "grad_norm": 47.75,
      "learning_rate": 1.63106684101223e-06,
      "loss": 0.0235,
      "mean_token_accuracy": 0.9910179376602173,
      "num_tokens": 5152552.0,
      "step": 11865
    },
    {
      "epoch": 0.9126990231520652,
      "grad_norm": 34.0,
      "learning_rate": 1.6309514652719022e-06,
      "loss": 0.0258,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 5152870.0,
      "step": 11866
    },
    {
      "epoch": 0.9127759403122837,
      "grad_norm": 18.25,
      "learning_rate": 1.6308360895315744e-06,
      "loss": 0.0242,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5153164.0,
      "step": 11867
    },
    {
      "epoch": 0.9128528574725021,
      "grad_norm": 15.1875,
      "learning_rate": 1.630720713791247e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9942196607589722,
      "num_tokens": 5153684.0,
      "step": 11868
    },
    {
      "epoch": 0.9129297746327205,
      "grad_norm": 7.5,
      "learning_rate": 1.6306053380509192e-06,
      "loss": 0.0077,
      "mean_token_accuracy": 0.9982237815856934,
      "num_tokens": 5154248.0,
      "step": 11869
    },
    {
      "epoch": 0.913006691792939,
      "grad_norm": 20.25,
      "learning_rate": 1.6304899623105916e-06,
      "loss": 0.0121,
      "mean_token_accuracy": 0.9948186278343201,
      "num_tokens": 5154828.0,
      "step": 11870
    },
    {
      "epoch": 0.9130836089531574,
      "grad_norm": 29.625,
      "learning_rate": 1.6303745865702637e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9963592290878296,
      "num_tokens": 5155654.0,
      "step": 11871
    },
    {
      "epoch": 0.913160526113376,
      "grad_norm": 4.5,
      "learning_rate": 1.6302592108299362e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.9946140050888062,
      "num_tokens": 5156212.0,
      "step": 11872
    },
    {
      "epoch": 0.9132374432735944,
      "grad_norm": 4.1875,
      "learning_rate": 1.6301438350896084e-06,
      "loss": 0.0066,
      "mean_token_accuracy": 0.9987593293190002,
      "num_tokens": 5157020.0,
      "step": 11873
    },
    {
      "epoch": 0.9133143604338128,
      "grad_norm": 61.75,
      "learning_rate": 1.6300284593492809e-06,
      "loss": 0.0342,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 5157280.0,
      "step": 11874
    },
    {
      "epoch": 0.9133912775940313,
      "grad_norm": 24.625,
      "learning_rate": 1.6299130836089533e-06,
      "loss": 0.0368,
      "mean_token_accuracy": 0.9841897487640381,
      "num_tokens": 5157534.0,
      "step": 11875
    },
    {
      "epoch": 0.9134681947542497,
      "grad_norm": 11.6875,
      "learning_rate": 1.6297977078686256e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9982486963272095,
      "num_tokens": 5158106.0,
      "step": 11876
    },
    {
      "epoch": 0.9135451119144681,
      "grad_norm": 12.3125,
      "learning_rate": 1.6296823321282977e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9982608556747437,
      "num_tokens": 5158682.0,
      "step": 11877
    },
    {
      "epoch": 0.9136220290746866,
      "grad_norm": 17.375,
      "learning_rate": 1.6295669563879701e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9852398633956909,
      "num_tokens": 5158954.0,
      "step": 11878
    },
    {
      "epoch": 0.913698946234905,
      "grad_norm": 30.875,
      "learning_rate": 1.6294515806476426e-06,
      "loss": 0.0237,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 5159278.0,
      "step": 11879
    },
    {
      "epoch": 0.9137758633951234,
      "grad_norm": 26.5,
      "learning_rate": 1.6293362049073149e-06,
      "loss": 0.0204,
      "mean_token_accuracy": 0.9876922965049744,
      "num_tokens": 5159604.0,
      "step": 11880
    },
    {
      "epoch": 0.9138527805553419,
      "grad_norm": 4.71875,
      "learning_rate": 1.6292208291669873e-06,
      "loss": 0.0098,
      "mean_token_accuracy": 0.9981916546821594,
      "num_tokens": 5160158.0,
      "step": 11881
    },
    {
      "epoch": 0.9139296977155603,
      "grad_norm": 18.375,
      "learning_rate": 1.6291054534266594e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9829351305961609,
      "num_tokens": 5160452.0,
      "step": 11882
    },
    {
      "epoch": 0.9140066148757788,
      "grad_norm": 47.75,
      "learning_rate": 1.6289900776863318e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9861591458320618,
      "num_tokens": 5160742.0,
      "step": 11883
    },
    {
      "epoch": 0.9140835320359972,
      "grad_norm": 23.875,
      "learning_rate": 1.628874701946004e-06,
      "loss": 0.0252,
      "mean_token_accuracy": 0.9865319728851318,
      "num_tokens": 5161040.0,
      "step": 11884
    },
    {
      "epoch": 0.9141604491962156,
      "grad_norm": 8.3125,
      "learning_rate": 1.6287593262056766e-06,
      "loss": 0.0162,
      "mean_token_accuracy": 0.994535505771637,
      "num_tokens": 5161408.0,
      "step": 11885
    },
    {
      "epoch": 0.9142373663564342,
      "grad_norm": 10.5,
      "learning_rate": 1.628643950465349e-06,
      "loss": 0.0148,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 5161980.0,
      "step": 11886
    },
    {
      "epoch": 0.9143142835166526,
      "grad_norm": 13.0625,
      "learning_rate": 1.628528574725021e-06,
      "loss": 0.0233,
      "mean_token_accuracy": 0.9881305694580078,
      "num_tokens": 5162318.0,
      "step": 11887
    },
    {
      "epoch": 0.914391200676871,
      "grad_norm": 21.125,
      "learning_rate": 1.6284131989846933e-06,
      "loss": 0.0147,
      "mean_token_accuracy": 0.994434118270874,
      "num_tokens": 5162858.0,
      "step": 11888
    },
    {
      "epoch": 0.9144681178370895,
      "grad_norm": 10.875,
      "learning_rate": 1.6282978232443658e-06,
      "loss": 0.0116,
      "mean_token_accuracy": 0.994991660118103,
      "num_tokens": 5163458.0,
      "step": 11889
    },
    {
      "epoch": 0.9145450349973079,
      "grad_norm": 50.0,
      "learning_rate": 1.6281824475040383e-06,
      "loss": 0.0253,
      "mean_token_accuracy": 0.9901639223098755,
      "num_tokens": 5163764.0,
      "step": 11890
    },
    {
      "epoch": 0.9146219521575264,
      "grad_norm": 10.375,
      "learning_rate": 1.6280670717637105e-06,
      "loss": 0.0177,
      "mean_token_accuracy": 0.9938080310821533,
      "num_tokens": 5164088.0,
      "step": 11891
    },
    {
      "epoch": 0.9146988693177448,
      "grad_norm": 41.25,
      "learning_rate": 1.627951696023383e-06,
      "loss": 0.0115,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 5164674.0,
      "step": 11892
    },
    {
      "epoch": 0.9147757864779632,
      "grad_norm": 44.5,
      "learning_rate": 1.627836320283055e-06,
      "loss": 0.0215,
      "mean_token_accuracy": 0.9968454241752625,
      "num_tokens": 5164994.0,
      "step": 11893
    },
    {
      "epoch": 0.9148527036381817,
      "grad_norm": 12.625,
      "learning_rate": 1.6277209445427275e-06,
      "loss": 0.0132,
      "mean_token_accuracy": 0.9970501661300659,
      "num_tokens": 5165334.0,
      "step": 11894
    },
    {
      "epoch": 0.9149296207984001,
      "grad_norm": 25.75,
      "learning_rate": 1.6276055688023998e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9952038526535034,
      "num_tokens": 5166170.0,
      "step": 11895
    },
    {
      "epoch": 0.9150065379586185,
      "grad_norm": 22.5,
      "learning_rate": 1.6274901930620723e-06,
      "loss": 0.0305,
      "mean_token_accuracy": 0.9857650995254517,
      "num_tokens": 5166452.0,
      "step": 11896
    },
    {
      "epoch": 0.915083455118837,
      "grad_norm": 7.9375,
      "learning_rate": 1.6273748173217447e-06,
      "loss": 0.0174,
      "mean_token_accuracy": 0.9924670457839966,
      "num_tokens": 5166984.0,
      "step": 11897
    },
    {
      "epoch": 0.9151603722790554,
      "grad_norm": 17.875,
      "learning_rate": 1.6272594415814168e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 5167556.0,
      "step": 11898
    },
    {
      "epoch": 0.9152372894392738,
      "grad_norm": 65.5,
      "learning_rate": 1.627144065841089e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.989159882068634,
      "num_tokens": 5167928.0,
      "step": 11899
    },
    {
      "epoch": 0.9153142065994924,
      "grad_norm": 38.25,
      "learning_rate": 1.6270286901007615e-06,
      "loss": 0.0437,
      "mean_token_accuracy": 0.9805447459220886,
      "num_tokens": 5168186.0,
      "step": 11900
    },
    {
      "epoch": 0.9153911237597108,
      "grad_norm": 35.25,
      "learning_rate": 1.626913314360434e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9873816967010498,
      "num_tokens": 5168504.0,
      "step": 11901
    },
    {
      "epoch": 0.9154680409199293,
      "grad_norm": 26.625,
      "learning_rate": 1.6267979386201062e-06,
      "loss": 0.0124,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 5169050.0,
      "step": 11902
    },
    {
      "epoch": 0.9155449580801477,
      "grad_norm": 43.0,
      "learning_rate": 1.6266825628797787e-06,
      "loss": 0.0379,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 5169332.0,
      "step": 11903
    },
    {
      "epoch": 0.9156218752403661,
      "grad_norm": 27.875,
      "learning_rate": 1.6265671871394507e-06,
      "loss": 0.0139,
      "mean_token_accuracy": 0.9928698539733887,
      "num_tokens": 5169894.0,
      "step": 11904
    },
    {
      "epoch": 0.9156987924005846,
      "grad_norm": 20.25,
      "learning_rate": 1.6264518113991232e-06,
      "loss": 0.0196,
      "mean_token_accuracy": 0.990963876247406,
      "num_tokens": 5170228.0,
      "step": 11905
    },
    {
      "epoch": 0.915775709560803,
      "grad_norm": 10.375,
      "learning_rate": 1.6263364356587955e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 5170744.0,
      "step": 11906
    },
    {
      "epoch": 0.9158526267210214,
      "grad_norm": 30.75,
      "learning_rate": 1.626221059918468e-06,
      "loss": 0.0206,
      "mean_token_accuracy": 0.9936507940292358,
      "num_tokens": 5171060.0,
      "step": 11907
    },
    {
      "epoch": 0.9159295438812399,
      "grad_norm": 76.0,
      "learning_rate": 1.6261056841781404e-06,
      "loss": 0.0419,
      "mean_token_accuracy": 0.9765886068344116,
      "num_tokens": 5171362.0,
      "step": 11908
    },
    {
      "epoch": 0.9160064610414583,
      "grad_norm": 17.5,
      "learning_rate": 1.6259903084378125e-06,
      "loss": 0.0085,
      "mean_token_accuracy": 0.9951456189155579,
      "num_tokens": 5172188.0,
      "step": 11909
    },
    {
      "epoch": 0.9160833782016768,
      "grad_norm": 5.71875,
      "learning_rate": 1.6258749326974847e-06,
      "loss": 0.0081,
      "mean_token_accuracy": 0.9961439371109009,
      "num_tokens": 5172968.0,
      "step": 11910
    },
    {
      "epoch": 0.9161602953618952,
      "grad_norm": 30.75,
      "learning_rate": 1.6257595569571572e-06,
      "loss": 0.0248,
      "mean_token_accuracy": 0.9871794581413269,
      "num_tokens": 5173282.0,
      "step": 11911
    },
    {
      "epoch": 0.9162372125221137,
      "grad_norm": 19.625,
      "learning_rate": 1.6256441812168297e-06,
      "loss": 0.0194,
      "mean_token_accuracy": 0.9898989796638489,
      "num_tokens": 5173580.0,
      "step": 11912
    },
    {
      "epoch": 0.9163141296823322,
      "grad_norm": 6.65625,
      "learning_rate": 1.625528805476502e-06,
      "loss": 0.0089,
      "mean_token_accuracy": 0.9950739145278931,
      "num_tokens": 5174394.0,
      "step": 11913
    },
    {
      "epoch": 0.9163910468425506,
      "grad_norm": 11.4375,
      "learning_rate": 1.625413429736174e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 5174962.0,
      "step": 11914
    },
    {
      "epoch": 0.916467964002769,
      "grad_norm": 18.625,
      "learning_rate": 1.6252980539958464e-06,
      "loss": 0.022,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5175256.0,
      "step": 11915
    },
    {
      "epoch": 0.9165448811629875,
      "grad_norm": 25.125,
      "learning_rate": 1.625182678255519e-06,
      "loss": 0.0202,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 5175568.0,
      "step": 11916
    },
    {
      "epoch": 0.9166217983232059,
      "grad_norm": 43.25,
      "learning_rate": 1.6250673025151912e-06,
      "loss": 0.0403,
      "mean_token_accuracy": 0.9797979593276978,
      "num_tokens": 5175868.0,
      "step": 11917
    },
    {
      "epoch": 0.9166987154834243,
      "grad_norm": 83.0,
      "learning_rate": 1.6249519267748636e-06,
      "loss": 0.0227,
      "mean_token_accuracy": 0.9939209818840027,
      "num_tokens": 5176198.0,
      "step": 11918
    },
    {
      "epoch": 0.9167756326436428,
      "grad_norm": 14.6875,
      "learning_rate": 1.624836551034536e-06,
      "loss": 0.0128,
      "mean_token_accuracy": 0.9942418336868286,
      "num_tokens": 5176720.0,
      "step": 11919
    },
    {
      "epoch": 0.9168525498038612,
      "grad_norm": 11.625,
      "learning_rate": 1.6247211752942081e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.9905362725257874,
      "num_tokens": 5177038.0,
      "step": 11920
    },
    {
      "epoch": 0.9169294669640797,
      "grad_norm": 16.75,
      "learning_rate": 1.6246057995538804e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9965811967849731,
      "num_tokens": 5177624.0,
      "step": 11921
    },
    {
      "epoch": 0.9170063841242981,
      "grad_norm": 16.875,
      "learning_rate": 1.6244904238135529e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9945945739746094,
      "num_tokens": 5178180.0,
      "step": 11922
    },
    {
      "epoch": 0.9170833012845165,
      "grad_norm": 8.3125,
      "learning_rate": 1.6243750480732253e-06,
      "loss": 0.0108,
      "mean_token_accuracy": 0.9944751262664795,
      "num_tokens": 5178724.0,
      "step": 11923
    },
    {
      "epoch": 0.917160218444735,
      "grad_norm": 39.0,
      "learning_rate": 1.6242596723328976e-06,
      "loss": 0.0224,
      "mean_token_accuracy": 0.9870967864990234,
      "num_tokens": 5179036.0,
      "step": 11924
    },
    {
      "epoch": 0.9172371356049535,
      "grad_norm": 18.75,
      "learning_rate": 1.6241442965925696e-06,
      "loss": 0.014,
      "mean_token_accuracy": 1.0,
      "num_tokens": 5179412.0,
      "step": 11925
    },
    {
      "epoch": 0.9173140527651719,
      "grad_norm": 4.4375,
      "learning_rate": 1.6240289208522421e-06,
      "loss": 0.0071,
      "mean_token_accuracy": 0.9964850544929504,
      "num_tokens": 5179982.0,
      "step": 11926
    },
    {
      "epoch": 0.9173909699253904,
      "grad_norm": 57.25,
      "learning_rate": 1.6239135451119146e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9871382713317871,
      "num_tokens": 5180294.0,
      "step": 11927
    },
    {
      "epoch": 0.9174678870856088,
      "grad_norm": 52.0,
      "learning_rate": 1.6237981693715868e-06,
      "loss": 0.0178,
      "mean_token_accuracy": 0.9943661689758301,
      "num_tokens": 5180650.0,
      "step": 11928
    },
    {
      "epoch": 0.9175448042458273,
      "grad_norm": 28.375,
      "learning_rate": 1.6236827936312593e-06,
      "loss": 0.0214,
      "mean_token_accuracy": 0.9903225898742676,
      "num_tokens": 5180962.0,
      "step": 11929
    },
    {
      "epoch": 0.9176217214060457,
      "grad_norm": 7.25,
      "learning_rate": 1.6235674178909316e-06,
      "loss": 0.0099,
      "mean_token_accuracy": 0.9935897588729858,
      "num_tokens": 5181744.0,
      "step": 11930
    },
    {
      "epoch": 0.9176986385662641,
      "grad_norm": 18.875,
      "learning_rate": 1.6234520421506038e-06,
      "loss": 0.0244,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 5182090.0,
      "step": 11931
    },
    {
      "epoch": 0.9177755557264826,
      "grad_norm": 13.75,
      "learning_rate": 1.623336666410276e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9926874041557312,
      "num_tokens": 5182638.0,
      "step": 11932
    },
    {
      "epoch": 0.917852472886701,
      "grad_norm": 60.5,
      "learning_rate": 1.6232212906699486e-06,
      "loss": 0.0282,
      "mean_token_accuracy": 0.9868852496147156,
      "num_tokens": 5182944.0,
      "step": 11933
    },
    {
      "epoch": 0.9179293900469194,
      "grad_norm": 30.125,
      "learning_rate": 1.6231059149296208e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 5183264.0,
      "step": 11934
    },
    {
      "epoch": 0.9180063072071379,
      "grad_norm": 26.75,
      "learning_rate": 1.6229905391892933e-06,
      "loss": 0.0236,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 5183586.0,
      "step": 11935
    },
    {
      "epoch": 0.9180832243673563,
      "grad_norm": 59.25,
      "learning_rate": 1.6228751634489653e-06,
      "loss": 0.0191,
      "mean_token_accuracy": 0.991919219493866,
      "num_tokens": 5184082.0,
      "step": 11936
    },
    {
      "epoch": 0.9181601415275747,
      "grad_norm": 8.4375,
      "learning_rate": 1.6227597877086378e-06,
      "loss": 0.012,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 5184666.0,
      "step": 11937
    },
    {
      "epoch": 0.9182370586877933,
      "grad_norm": 14.8125,
      "learning_rate": 1.62264441196831e-06,
      "loss": 0.023,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 5184988.0,
      "step": 11938
    },
    {
      "epoch": 0.9183139758480117,
      "grad_norm": 47.25,
      "learning_rate": 1.6225290362279825e-06,
      "loss": 0.014,
      "mean_token_accuracy": 0.9932088255882263,
      "num_tokens": 5185578.0,
      "step": 11939
    },
    {
      "epoch": 0.9183908930082302,
      "grad_norm": 22.875,
      "learning_rate": 1.622413660487655e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9875389337539673,
      "num_tokens": 5185900.0,
      "step": 11940
    },
    {
      "epoch": 0.9184678101684486,
      "grad_norm": 9.75,
      "learning_rate": 1.6222982847473273e-06,
      "loss": 0.0103,
      "mean_token_accuracy": 0.9963436722755432,
      "num_tokens": 5186448.0,
      "step": 11941
    },
    {
      "epoch": 0.918544727328667,
      "grad_norm": 16.0,
      "learning_rate": 1.6221829090069993e-06,
      "loss": 0.0193,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 5186798.0,
      "step": 11942
    },
    {
      "epoch": 0.9186216444888855,
      "grad_norm": 56.5,
      "learning_rate": 1.6220675332666718e-06,
      "loss": 0.0158,
      "mean_token_accuracy": 0.9924952983856201,
      "num_tokens": 5187332.0,
      "step": 11943
    },
    {
      "epoch": 0.9186985616491039,
      "grad_norm": 16.125,
      "learning_rate": 1.6219521575263442e-06,
      "loss": 0.0267,
      "mean_token_accuracy": 0.9895470142364502,
      "num_tokens": 5187620.0,
      "step": 11944
    },
    {
      "epoch": 0.9187754788093223,
      "grad_norm": 22.0,
      "learning_rate": 1.6218367817860165e-06,
      "loss": 0.0164,
      "mean_token_accuracy": 0.9941747784614563,
      "num_tokens": 5188136.0,
      "step": 11945
    },
    {
      "epoch": 0.9188523959695408,
      "grad_norm": 58.75,
      "learning_rate": 1.621721406045689e-06,
      "loss": 0.0398,
      "mean_token_accuracy": 0.983660101890564,
      "num_tokens": 5188444.0,
      "step": 11946
    },
    {
      "epoch": 0.9189293131297592,
      "grad_norm": 132.0,
      "learning_rate": 1.621606030305361e-06,
      "loss": 0.032,
      "mean_token_accuracy": 0.987500011920929,
      "num_tokens": 5188768.0,
      "step": 11947
    },
    {
      "epoch": 0.9190062302899777,
      "grad_norm": 26.875,
      "learning_rate": 1.6214906545650335e-06,
      "loss": 0.0187,
      "mean_token_accuracy": 0.99071204662323,
      "num_tokens": 5189092.0,
      "step": 11948
    },
    {
      "epoch": 0.9190831474501961,
      "grad_norm": 19.375,
      "learning_rate": 1.6213752788247057e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9947090148925781,
      "num_tokens": 5189660.0,
      "step": 11949
    },
    {
      "epoch": 0.9191600646104146,
      "grad_norm": 42.0,
      "learning_rate": 1.6212599030843782e-06,
      "loss": 0.0149,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 5190222.0,
      "step": 11950
    },
    {
      "epoch": 0.9192369817706331,
      "grad_norm": 43.0,
      "learning_rate": 1.6211445273440507e-06,
      "loss": 0.0211,
      "mean_token_accuracy": 0.9905956387519836,
      "num_tokens": 5190542.0,
      "step": 11951
    },
    {
      "epoch": 0.9193138989308515,
      "grad_norm": 36.25,
      "learning_rate": 1.6210291516037227e-06,
      "loss": 0.0222,
      "mean_token_accuracy": 0.9935691356658936,
      "num_tokens": 5190854.0,
      "step": 11952
    },
    {
      "epoch": 0.9193908160910699,
      "grad_norm": 55.0,
      "learning_rate": 1.620913775863395e-06,
      "loss": 0.0343,
      "mean_token_accuracy": 0.9845559597015381,
      "num_tokens": 5191114.0,
      "step": 11953
    },
    {
      "epoch": 0.9194677332512884,
      "grad_norm": 22.875,
      "learning_rate": 1.6207984001230675e-06,
      "loss": 0.0359,
      "mean_token_accuracy": 0.9880239367485046,
      "num_tokens": 5191450.0,
      "step": 11954
    },
    {
      "epoch": 0.9195446504115068,
      "grad_norm": 19.125,
      "learning_rate": 1.62068302438274e-06,
      "loss": 0.0144,
      "mean_token_accuracy": 0.9944953918457031,
      "num_tokens": 5191996.0,
      "step": 11955
    },
    {
      "epoch": 0.9196215675717253,
      "grad_norm": 11.1875,
      "learning_rate": 1.6205676486424122e-06,
      "loss": 0.0213,
      "mean_token_accuracy": 0.9864406585693359,
      "num_tokens": 5192292.0,
      "step": 11956
    },
    {
      "epoch": 0.9196984847319437,
      "grad_norm": 44.0,
      "learning_rate": 1.6204522729020847e-06,
      "loss": 0.0136,
      "mean_token_accuracy": 0.9946523904800415,
      "num_tokens": 5192854.0,
      "step": 11957
    },
    {
      "epoch": 0.9197754018921621,
      "grad_norm": 25.75,
      "learning_rate": 1.6203368971617567e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9939939975738525,
      "num_tokens": 5193188.0,
      "step": 11958
    },
    {
      "epoch": 0.9198523190523806,
      "grad_norm": 10.6875,
      "learning_rate": 1.6202215214214292e-06,
      "loss": 0.0131,
      "mean_token_accuracy": 0.9964221715927124,
      "num_tokens": 5193748.0,
      "step": 11959
    },
    {
      "epoch": 0.919929236212599,
      "grad_norm": 12.375,
      "learning_rate": 1.6201061456811014e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5194042.0,
      "step": 11960
    },
    {
      "epoch": 0.9200061533728174,
      "grad_norm": 9.125,
      "learning_rate": 1.619990769940774e-06,
      "loss": 0.0201,
      "mean_token_accuracy": 0.9896907210350037,
      "num_tokens": 5194334.0,
      "step": 11961
    },
    {
      "epoch": 0.920083070533036,
      "grad_norm": 29.125,
      "learning_rate": 1.6198753942004464e-06,
      "loss": 0.0073,
      "mean_token_accuracy": 0.9960668683052063,
      "num_tokens": 5195354.0,
      "step": 11962
    },
    {
      "epoch": 0.9201599876932544,
      "grad_norm": 13.0625,
      "learning_rate": 1.6197600184601184e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 5195922.0,
      "step": 11963
    },
    {
      "epoch": 0.9202369048534728,
      "grad_norm": 48.0,
      "learning_rate": 1.6196446427197907e-06,
      "loss": 0.0334,
      "mean_token_accuracy": 0.985401451587677,
      "num_tokens": 5196198.0,
      "step": 11964
    },
    {
      "epoch": 0.9203138220136913,
      "grad_norm": 7.96875,
      "learning_rate": 1.6195292669794631e-06,
      "loss": 0.0171,
      "mean_token_accuracy": 0.9923224449157715,
      "num_tokens": 5196720.0,
      "step": 11965
    },
    {
      "epoch": 0.9203907391739097,
      "grad_norm": 8.0,
      "learning_rate": 1.6194138912391356e-06,
      "loss": 0.0262,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5197014.0,
      "step": 11966
    },
    {
      "epoch": 0.9204676563341282,
      "grad_norm": 29.75,
      "learning_rate": 1.6192985154988079e-06,
      "loss": 0.0364,
      "mean_token_accuracy": 0.9843137264251709,
      "num_tokens": 5197270.0,
      "step": 11967
    },
    {
      "epoch": 0.9205445734943466,
      "grad_norm": 9.3125,
      "learning_rate": 1.6191831397584803e-06,
      "loss": 0.0122,
      "mean_token_accuracy": 0.9941973090171814,
      "num_tokens": 5197788.0,
      "step": 11968
    },
    {
      "epoch": 0.920621490654565,
      "grad_norm": 13.3125,
      "learning_rate": 1.6190677640181524e-06,
      "loss": 0.0118,
      "mean_token_accuracy": 0.996303141117096,
      "num_tokens": 5198330.0,
      "step": 11969
    },
    {
      "epoch": 0.9206984078147835,
      "grad_norm": 16.375,
      "learning_rate": 1.6189523882778249e-06,
      "loss": 0.0219,
      "mean_token_accuracy": 0.9931740760803223,
      "num_tokens": 5198624.0,
      "step": 11970
    },
    {
      "epoch": 0.9207753249750019,
      "grad_norm": 26.5,
      "learning_rate": 1.6188370125374971e-06,
      "loss": 0.0197,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 5198968.0,
      "step": 11971
    },
    {
      "epoch": 0.9208522421352203,
      "grad_norm": 10.875,
      "learning_rate": 1.6187216367971696e-06,
      "loss": 0.0181,
      "mean_token_accuracy": 0.994413435459137,
      "num_tokens": 5199328.0,
      "step": 11972
    },
    {
      "epoch": 0.9209291592954388,
      "grad_norm": 42.25,
      "learning_rate": 1.618606261056842e-06,
      "loss": 0.0231,
      "mean_token_accuracy": 0.991253674030304,
      "num_tokens": 5199672.0,
      "step": 11973
    },
    {
      "epoch": 0.9210060764556572,
      "grad_norm": 27.25,
      "learning_rate": 1.618490885316514e-06,
      "loss": 0.0198,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 5199990.0,
      "step": 11974
    },
    {
      "epoch": 0.9210829936158758,
      "grad_norm": 12.1875,
      "learning_rate": 1.6183755095761864e-06,
      "loss": 0.0129,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 5200536.0,
      "step": 11975
    },
    {
      "epoch": 0.9211599107760942,
      "grad_norm": 14.75,
      "learning_rate": 1.6182601338358588e-06,
      "loss": 0.0151,
      "mean_token_accuracy": 0.9926605224609375,
      "num_tokens": 5201082.0,
      "step": 11976
    },
    {
      "epoch": 0.9212368279363126,
      "grad_norm": 22.125,
      "learning_rate": 1.6181447580955313e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9936908483505249,
      "num_tokens": 5201400.0,
      "step": 11977
    },
    {
      "epoch": 0.9213137450965311,
      "grad_norm": 19.0,
      "learning_rate": 1.6180293823552036e-06,
      "loss": 0.013,
      "mean_token_accuracy": 0.9931389093399048,
      "num_tokens": 5201984.0,
      "step": 11978
    },
    {
      "epoch": 0.9213906622567495,
      "grad_norm": 12.75,
      "learning_rate": 1.617914006614876e-06,
      "loss": 0.0234,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 5202300.0,
      "step": 11979
    },
    {
      "epoch": 0.9214675794169679,
      "grad_norm": 38.75,
      "learning_rate": 1.617798630874548e-06,
      "loss": 0.0119,
      "mean_token_accuracy": 0.996259331703186,
      "num_tokens": 5203104.0,
      "step": 11980
    },
    {
      "epoch": 0.9215444965771864,
      "grad_norm": 23.75,
      "learning_rate": 1.6176832551342205e-06,
      "loss": 0.0331,
      "mean_token_accuracy": 0.9854227304458618,
      "num_tokens": 5203448.0,
      "step": 11981
    },
    {
      "epoch": 0.9216214137374048,
      "grad_norm": 70.5,
      "learning_rate": 1.6175678793938928e-06,
      "loss": 0.0301,
      "mean_token_accuracy": 0.9866220951080322,
      "num_tokens": 5203748.0,
      "step": 11982
    },
    {
      "epoch": 0.9216983308976232,
      "grad_norm": 45.25,
      "learning_rate": 1.6174525036535653e-06,
      "loss": 0.0274,
      "mean_token_accuracy": 0.9942693114280701,
      "num_tokens": 5204098.0,
      "step": 11983
    },
    {
      "epoch": 0.9217752480578417,
      "grad_norm": 14.875,
      "learning_rate": 1.6173371279132375e-06,
      "loss": 0.0184,
      "mean_token_accuracy": 0.9968454241752625,
      "num_tokens": 5204416.0,
      "step": 11984
    },
    {
      "epoch": 0.9218521652180601,
      "grad_norm": 29.25,
      "learning_rate": 1.6172217521729098e-06,
      "loss": 0.0268,
      "mean_token_accuracy": 0.991304337978363,
      "num_tokens": 5204762.0,
      "step": 11985
    },
    {
      "epoch": 0.9219290823782786,
      "grad_norm": 16.25,
      "learning_rate": 1.617106376432582e-06,
      "loss": 0.0117,
      "mean_token_accuracy": 0.9929453134536743,
      "num_tokens": 5205330.0,
      "step": 11986
    },
    {
      "epoch": 0.922005999538497,
      "grad_norm": 67.5,
      "learning_rate": 1.6169910006922545e-06,
      "loss": 0.0376,
      "mean_token_accuracy": 0.982206404209137,
      "num_tokens": 5205612.0,
      "step": 11987
    },
    {
      "epoch": 0.9220829166987154,
      "grad_norm": 10.9375,
      "learning_rate": 1.6168756249519268e-06,
      "loss": 0.011,
      "mean_token_accuracy": 0.9948717951774597,
      "num_tokens": 5206198.0,
      "step": 11988
    },
    {
      "epoch": 0.922159833858934,
      "grad_norm": 24.875,
      "learning_rate": 1.6167602492115992e-06,
      "loss": 0.0256,
      "mean_token_accuracy": 0.9906542301177979,
      "num_tokens": 5206520.0,
      "step": 11989
    },
    {
      "epoch": 0.9222367510191524,
      "grad_norm": 7.71875,
      "learning_rate": 1.6166448734712713e-06,
      "loss": 0.0058,
      "mean_token_accuracy": 0.9971618056297302,
      "num_tokens": 5207580.0,
      "step": 11990
    },
    {
      "epoch": 0.9223136681793708,
      "grad_norm": 50.5,
      "learning_rate": 1.6165294977309438e-06,
      "loss": 0.0217,
      "mean_token_accuracy": 0.9904761910438538,
      "num_tokens": 5207896.0,
      "step": 11991
    },
    {
      "epoch": 0.9223905853395893,
      "grad_norm": 4.78125,
      "learning_rate": 1.616414121990616e-06,
      "loss": 0.0086,
      "mean_token_accuracy": 0.9962871074676514,
      "num_tokens": 5208706.0,
      "step": 11992
    },
    {
      "epoch": 0.9224675024998077,
      "grad_norm": 3.734375,
      "learning_rate": 1.6162987462502885e-06,
      "loss": 0.005,
      "mean_token_accuracy": 0.996874988079071,
      "num_tokens": 5209990.0,
      "step": 11993
    },
    {
      "epoch": 0.9225444196600262,
      "grad_norm": 68.0,
      "learning_rate": 1.616183370509961e-06,
      "loss": 0.0265,
      "mean_token_accuracy": 0.9888476133346558,
      "num_tokens": 5210260.0,
      "step": 11994
    },
    {
      "epoch": 0.9226213368202446,
      "grad_norm": 23.125,
      "learning_rate": 1.6160679947696332e-06,
      "loss": 0.0269,
      "mean_token_accuracy": 0.9885386824607849,
      "num_tokens": 5210610.0,
      "step": 11995
    },
    {
      "epoch": 0.922698253980463,
      "grad_norm": 12.9375,
      "learning_rate": 1.6159526190293053e-06,
      "loss": 0.0109,
      "mean_token_accuracy": 0.994854211807251,
      "num_tokens": 5211194.0,
      "step": 11996
    },
    {
      "epoch": 0.9227751711406815,
      "grad_norm": 6.0625,
      "learning_rate": 1.6158372432889777e-06,
      "loss": 0.007,
      "mean_token_accuracy": 0.9966386556625366,
      "num_tokens": 5211790.0,
      "step": 11997
    },
    {
      "epoch": 0.9228520883008999,
      "grad_norm": 14.875,
      "learning_rate": 1.6157218675486502e-06,
      "loss": 0.0138,
      "mean_token_accuracy": 0.9929947257041931,
      "num_tokens": 5212362.0,
      "step": 11998
    },
    {
      "epoch": 0.9229290054611183,
      "grad_norm": 49.25,
      "learning_rate": 1.6156064918083225e-06,
      "loss": 0.0308,
      "mean_token_accuracy": 0.9897611141204834,
      "num_tokens": 5212656.0,
      "step": 11999
    },
    {
      "epoch": 0.9230059226213368,
      "grad_norm": 42.0,
      "learning_rate": 1.615491116067995e-06,
      "loss": 0.0203,
      "step": 12000
    },
    {
      "epoch": 0.9230059226213368,
      "eval_loss": 0.018283504992723465,
      "eval_mean_token_accuracy": 0.9921770977973938,
      "eval_num_tokens": 5212984.0,
      "eval_runtime": 21.0631,
      "eval_samples_per_second": 9.495,
      "eval_steps_per_second": 1.187,
      "step": 12000
    }
  ],
  "logging_steps": 1,
  "max_steps": 26002,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 4000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3.744127519198618e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}