{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.8361314731615037,
  "eval_steps": 150,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14187751241428234,
      "grad_norm": 19.588634490966797,
      "learning_rate": 9.763481551561022e-07,
      "logits/chosen": -2.82487154006958,
      "logits/rejected": -2.916430950164795,
      "logps/chosen": -288.4345397949219,
      "logps/rejected": -191.39353942871094,
      "loss": 0.4926,
      "rewards/accuracies": 0.8491666913032532,
      "rewards/chosen": 0.3584444224834442,
      "rewards/margins": 0.5347671508789062,
      "rewards/rejected": -0.17632275819778442,
      "step": 150
    },
    {
      "epoch": 0.14187751241428234,
      "eval_logits/chosen": -2.941565990447998,
      "eval_logits/rejected": -3.0292787551879883,
      "eval_logps/chosen": -290.8563232421875,
      "eval_logps/rejected": -199.278076171875,
      "eval_loss": 0.2873661518096924,
      "eval_rewards/accuracies": 0.9881852269172668,
      "eval_rewards/chosen": 0.28964486718177795,
      "eval_rewards/margins": 1.2328312397003174,
      "eval_rewards/rejected": -0.9431862831115723,
      "eval_runtime": 562.5858,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.881,
      "step": 150
    },
    {
      "epoch": 0.2837550248285647,
      "grad_norm": 4.974961280822754,
      "learning_rate": 9.526963103122043e-07,
      "logits/chosen": -2.976027727127075,
      "logits/rejected": -3.071249485015869,
      "logps/chosen": -284.5093078613281,
      "logps/rejected": -204.3736572265625,
      "loss": 0.1775,
      "rewards/accuracies": 0.9891666769981384,
      "rewards/chosen": 0.5465984344482422,
      "rewards/margins": 2.021888494491577,
      "rewards/rejected": -1.475290060043335,
      "step": 300
    },
    {
      "epoch": 0.2837550248285647,
      "eval_logits/chosen": -2.971857786178589,
      "eval_logits/rejected": -3.0642826557159424,
      "eval_logps/chosen": -286.28594970703125,
      "eval_logps/rejected": -209.5599822998047,
      "eval_loss": 0.11425645649433136,
      "eval_rewards/accuracies": 0.9910207986831665,
      "eval_rewards/chosen": 0.7466850876808167,
      "eval_rewards/margins": 2.7180631160736084,
      "eval_rewards/rejected": -1.9713778495788574,
      "eval_runtime": 562.6097,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.881,
      "step": 300
    },
    {
      "epoch": 0.425632537242847,
      "grad_norm": 22.92677116394043,
      "learning_rate": 9.290444654683065e-07,
      "logits/chosen": -2.9028356075286865,
      "logits/rejected": -3.013791084289551,
      "logps/chosen": -282.4086608886719,
      "logps/rejected": -209.1995849609375,
      "loss": 0.0941,
      "rewards/accuracies": 0.9916666746139526,
      "rewards/chosen": 0.9992213845252991,
      "rewards/margins": 3.071294069290161,
      "rewards/rejected": -2.0720725059509277,
      "step": 450
    },
    {
      "epoch": 0.425632537242847,
      "eval_logits/chosen": -2.90273118019104,
      "eval_logits/rejected": -3.002680540084839,
      "eval_logps/chosen": -282.59246826171875,
      "eval_logps/rejected": -213.33204650878906,
      "eval_loss": 0.07543039321899414,
      "eval_rewards/accuracies": 0.9933837652206421,
      "eval_rewards/chosen": 1.1160286664962769,
      "eval_rewards/margins": 3.464611768722534,
      "eval_rewards/rejected": -2.348583459854126,
      "eval_runtime": 563.5688,
      "eval_samples_per_second": 3.753,
      "eval_steps_per_second": 1.877,
      "step": 450
    },
    {
      "epoch": 0.5675100496571294,
      "grad_norm": 22.853031158447266,
      "learning_rate": 9.053926206244087e-07,
      "logits/chosen": -2.9159042835235596,
      "logits/rejected": -3.0238444805145264,
      "logps/chosen": -284.7449951171875,
      "logps/rejected": -216.9216766357422,
      "loss": 0.0687,
      "rewards/accuracies": 0.9958333373069763,
      "rewards/chosen": 0.9207213521003723,
      "rewards/margins": 3.7362918853759766,
      "rewards/rejected": -2.815570592880249,
      "step": 600
    },
    {
      "epoch": 0.5675100496571294,
      "eval_logits/chosen": -2.8956897258758545,
      "eval_logits/rejected": -3.010921001434326,
      "eval_logps/chosen": -283.6835632324219,
      "eval_logps/rejected": -220.16329956054688,
      "eval_loss": 0.058039747178554535,
      "eval_rewards/accuracies": 0.9948015213012695,
      "eval_rewards/chosen": 1.0069197416305542,
      "eval_rewards/margins": 4.038629531860352,
      "eval_rewards/rejected": -3.031709671020508,
      "eval_runtime": 563.2115,
      "eval_samples_per_second": 3.755,
      "eval_steps_per_second": 1.879,
      "step": 600
    },
    {
      "epoch": 0.7093875620714116,
      "grad_norm": 4.9486188888549805,
      "learning_rate": 8.817407757805109e-07,
      "logits/chosen": -2.9240283966064453,
      "logits/rejected": -3.0465874671936035,
      "logps/chosen": -282.1627197265625,
      "logps/rejected": -223.1807098388672,
      "loss": 0.0522,
      "rewards/accuracies": 0.9950000047683716,
      "rewards/chosen": 0.8215107321739197,
      "rewards/margins": 4.242582321166992,
      "rewards/rejected": -3.421071767807007,
      "step": 750
    },
    {
      "epoch": 0.7093875620714116,
      "eval_logits/chosen": -2.9431002140045166,
      "eval_logits/rejected": -3.0670864582061768,
      "eval_logps/chosen": -285.1332092285156,
      "eval_logps/rejected": -226.6123046875,
      "eval_loss": 0.0461815781891346,
      "eval_rewards/accuracies": 0.9943289160728455,
      "eval_rewards/chosen": 0.8619542717933655,
      "eval_rewards/margins": 4.538564205169678,
      "eval_rewards/rejected": -3.676610231399536,
      "eval_runtime": 562.549,
      "eval_samples_per_second": 3.76,
      "eval_steps_per_second": 1.881,
      "step": 750
    },
    {
      "epoch": 0.851265074485694,
      "grad_norm": 0.37568044662475586,
      "learning_rate": 8.580889309366131e-07,
      "logits/chosen": -2.9640204906463623,
      "logits/rejected": -3.1055846214294434,
      "logps/chosen": -286.8418273925781,
      "logps/rejected": -232.3475799560547,
      "loss": 0.0416,
      "rewards/accuracies": 0.9958333373069763,
      "rewards/chosen": 0.5082718729972839,
      "rewards/margins": 4.854653358459473,
      "rewards/rejected": -4.346381664276123,
      "step": 900
    },
    {
      "epoch": 0.851265074485694,
      "eval_logits/chosen": -2.946694850921631,
      "eval_logits/rejected": -3.080843210220337,
      "eval_logps/chosen": -286.02020263671875,
      "eval_logps/rejected": -232.0843048095703,
      "eval_loss": 0.039623409509658813,
      "eval_rewards/accuracies": 0.9957466721534729,
      "eval_rewards/chosen": 0.7732585668563843,
      "eval_rewards/margins": 4.997069835662842,
      "eval_rewards/rejected": -4.223811626434326,
      "eval_runtime": 563.1761,
      "eval_samples_per_second": 3.755,
      "eval_steps_per_second": 1.879,
      "step": 900
    },
    {
      "epoch": 0.9931425868999764,
      "grad_norm": 10.330118179321289,
      "learning_rate": 8.344370860927152e-07,
      "logits/chosen": -2.9750888347625732,
      "logits/rejected": -3.1029014587402344,
      "logps/chosen": -288.4482421875,
      "logps/rejected": -237.4921417236328,
      "loss": 0.0382,
      "rewards/accuracies": 0.9941666722297668,
      "rewards/chosen": 0.47679439187049866,
      "rewards/margins": 5.323023796081543,
      "rewards/rejected": -4.846229076385498,
      "step": 1050
    },
    {
      "epoch": 0.9931425868999764,
      "eval_logits/chosen": -2.9789628982543945,
      "eval_logits/rejected": -3.1136224269866943,
      "eval_logps/chosen": -287.87921142578125,
      "eval_logps/rejected": -238.6780548095703,
      "eval_loss": 0.0340222492814064,
      "eval_rewards/accuracies": 0.9948015213012695,
      "eval_rewards/chosen": 0.587354838848114,
      "eval_rewards/margins": 5.470540523529053,
      "eval_rewards/rejected": -4.883185863494873,
      "eval_runtime": 563.0622,
      "eval_samples_per_second": 3.756,
      "eval_steps_per_second": 1.879,
      "step": 1050
    },
    {
      "epoch": 1.1343107117521873,
      "grad_norm": 2.3795278072357178,
      "learning_rate": 8.107852412488174e-07,
      "logits/chosen": -3.019599676132202,
      "logits/rejected": -3.1642258167266846,
      "logps/chosen": -289.4803466796875,
      "logps/rejected": -247.00460815429688,
      "loss": 0.0232,
      "rewards/accuracies": 0.999162495136261,
      "rewards/chosen": 0.016025420278310776,
      "rewards/margins": 5.90906286239624,
      "rewards/rejected": -5.893036842346191,
      "step": 1200
    },
    {
      "epoch": 1.1343107117521873,
      "eval_logits/chosen": -3.02445650100708,
      "eval_logits/rejected": -3.166768789291382,
      "eval_logps/chosen": -292.52386474609375,
      "eval_logps/rejected": -247.54971313476562,
      "eval_loss": 0.0312405526638031,
      "eval_rewards/accuracies": 0.9938563108444214,
      "eval_rewards/chosen": 0.12289439141750336,
      "eval_rewards/margins": 5.893245220184326,
      "eval_rewards/rejected": -5.770349979400635,
      "eval_runtime": 562.6683,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.88,
      "step": 1200
    },
    {
      "epoch": 1.2761882241664697,
      "grad_norm": 0.9838615655899048,
      "learning_rate": 7.871333964049196e-07,
      "logits/chosen": -3.0383057594299316,
      "logits/rejected": -3.189190149307251,
      "logps/chosen": -291.999267578125,
      "logps/rejected": -253.01220703125,
      "loss": 0.0322,
      "rewards/accuracies": 0.9950000047683716,
      "rewards/chosen": 0.02593408152461052,
      "rewards/margins": 6.261542797088623,
      "rewards/rejected": -6.2356085777282715,
      "step": 1350
    },
    {
      "epoch": 1.2761882241664697,
      "eval_logits/chosen": -3.048924446105957,
      "eval_logits/rejected": -3.1801936626434326,
      "eval_logps/chosen": -296.5158996582031,
      "eval_logps/rejected": -256.31951904296875,
      "eval_loss": 0.0295143760740757,
      "eval_rewards/accuracies": 0.9914934039115906,
      "eval_rewards/chosen": -0.2763078212738037,
      "eval_rewards/margins": 6.371026039123535,
      "eval_rewards/rejected": -6.647334098815918,
      "eval_runtime": 563.5207,
      "eval_samples_per_second": 3.753,
      "eval_steps_per_second": 1.877,
      "step": 1350
    },
    {
      "epoch": 1.4180657365807519,
      "grad_norm": 0.705359697341919,
      "learning_rate": 7.634815515610217e-07,
      "logits/chosen": -3.047039270401001,
      "logits/rejected": -3.1857874393463135,
      "logps/chosen": -298.34417724609375,
      "logps/rejected": -261.52630615234375,
      "loss": 0.0185,
      "rewards/accuracies": 0.9983333349227905,
      "rewards/chosen": -0.4116480052471161,
      "rewards/margins": 6.801379203796387,
      "rewards/rejected": -7.213026523590088,
      "step": 1500
    },
    {
      "epoch": 1.4180657365807519,
      "eval_logits/chosen": -3.0422451496124268,
      "eval_logits/rejected": -3.189257860183716,
      "eval_logps/chosen": -293.9241027832031,
      "eval_logps/rejected": -256.4940185546875,
      "eval_loss": 0.025707272812724113,
      "eval_rewards/accuracies": 0.9938563108444214,
      "eval_rewards/chosen": -0.017133429646492004,
      "eval_rewards/margins": 6.64764928817749,
      "eval_rewards/rejected": -6.664782524108887,
      "eval_runtime": 562.6032,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.881,
      "step": 1500
    },
    {
      "epoch": 1.5599432489950344,
      "grad_norm": 0.3177180290222168,
      "learning_rate": 7.398297067171239e-07,
      "logits/chosen": -3.0570619106292725,
      "logits/rejected": -3.196061611175537,
      "logps/chosen": -296.9585266113281,
      "logps/rejected": -262.5743103027344,
      "loss": 0.0336,
      "rewards/accuracies": 0.9958333373069763,
      "rewards/chosen": -0.3190596401691437,
      "rewards/margins": 7.057786464691162,
      "rewards/rejected": -7.376845836639404,
      "step": 1650
    },
    {
      "epoch": 1.5599432489950344,
      "eval_logits/chosen": -3.0481271743774414,
      "eval_logits/rejected": -3.193105459213257,
      "eval_logps/chosen": -294.719482421875,
      "eval_logps/rejected": -261.1886291503906,
      "eval_loss": 0.024315308779478073,
      "eval_rewards/accuracies": 0.9938563108444214,
      "eval_rewards/chosen": -0.09667125344276428,
      "eval_rewards/margins": 7.037571430206299,
      "eval_rewards/rejected": -7.134242534637451,
      "eval_runtime": 562.6021,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.881,
      "step": 1650
    },
    {
      "epoch": 1.7018207614093166,
      "grad_norm": 1.28083336353302,
      "learning_rate": 7.16177861873226e-07,
      "logits/chosen": -3.0715489387512207,
      "logits/rejected": -3.202450752258301,
      "logps/chosen": -296.61590576171875,
      "logps/rejected": -270.3304443359375,
      "loss": 0.0175,
      "rewards/accuracies": 0.9950000047683716,
      "rewards/chosen": -0.5375873446464539,
      "rewards/margins": 7.566472053527832,
      "rewards/rejected": -8.104060173034668,
      "step": 1800
    },
    {
      "epoch": 1.7018207614093166,
      "eval_logits/chosen": -3.0735578536987305,
      "eval_logits/rejected": -3.224771022796631,
      "eval_logps/chosen": -295.2926025390625,
      "eval_logps/rejected": -264.1260681152344,
      "eval_loss": 0.02150208130478859,
      "eval_rewards/accuracies": 0.9957466721534729,
      "eval_rewards/chosen": -0.15398064255714417,
      "eval_rewards/margins": 7.2740068435668945,
      "eval_rewards/rejected": -7.427987098693848,
      "eval_runtime": 562.9492,
      "eval_samples_per_second": 3.757,
      "eval_steps_per_second": 1.879,
      "step": 1800
    },
    {
      "epoch": 1.843698273823599,
      "grad_norm": 0.21602575480937958,
      "learning_rate": 6.925260170293283e-07,
      "logits/chosen": -3.103926181793213,
      "logits/rejected": -3.251049280166626,
      "logps/chosen": -296.3301086425781,
      "logps/rejected": -273.81640625,
      "loss": 0.0155,
      "rewards/accuracies": 0.9975000023841858,
      "rewards/chosen": -0.617112398147583,
      "rewards/margins": 7.931570529937744,
      "rewards/rejected": -8.54868221282959,
      "step": 1950
    },
    {
      "epoch": 1.843698273823599,
      "eval_logits/chosen": -3.1030561923980713,
      "eval_logits/rejected": -3.2201902866363525,
      "eval_logps/chosen": -301.2184143066406,
      "eval_logps/rejected": -281.40008544921875,
      "eval_loss": 0.022298738360404968,
      "eval_rewards/accuracies": 0.9943289160728455,
      "eval_rewards/chosen": -0.746565043926239,
      "eval_rewards/margins": 8.40882396697998,
      "eval_rewards/rejected": -9.155388832092285,
      "eval_runtime": 563.115,
      "eval_samples_per_second": 3.756,
      "eval_steps_per_second": 1.879,
      "step": 1950
    },
    {
      "epoch": 1.9855757862378813,
      "grad_norm": 0.49206382036209106,
      "learning_rate": 6.688741721854304e-07,
      "logits/chosen": -3.116490125656128,
      "logits/rejected": -3.230647087097168,
      "logps/chosen": -301.8218078613281,
      "logps/rejected": -286.8536376953125,
      "loss": 0.0169,
      "rewards/accuracies": 0.9975000023841858,
      "rewards/chosen": -1.0506435632705688,
      "rewards/margins": 8.755069732666016,
      "rewards/rejected": -9.80571460723877,
      "step": 2100
    },
    {
      "epoch": 1.9855757862378813,
      "eval_logits/chosen": -3.11275053024292,
      "eval_logits/rejected": -3.2121145725250244,
      "eval_logps/chosen": -302.8241271972656,
      "eval_logps/rejected": -287.6995544433594,
      "eval_loss": 0.021507343277335167,
      "eval_rewards/accuracies": 0.9938563108444214,
      "eval_rewards/chosen": -0.9071367383003235,
      "eval_rewards/margins": 8.878198623657227,
      "eval_rewards/rejected": -9.785334587097168,
      "eval_runtime": 562.6464,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.88,
      "step": 2100
    },
    {
      "epoch": 2.126743911090092,
      "grad_norm": 0.9229658246040344,
      "learning_rate": 6.452223273415326e-07,
      "logits/chosen": -3.1183557510375977,
      "logits/rejected": -3.2241806983947754,
      "logps/chosen": -295.1089172363281,
      "logps/rejected": -285.6842346191406,
      "loss": 0.0114,
      "rewards/accuracies": 0.9966499209403992,
      "rewards/chosen": -0.7439343929290771,
      "rewards/margins": 9.040164947509766,
      "rewards/rejected": -9.784099578857422,
      "step": 2250
    },
    {
      "epoch": 2.126743911090092,
      "eval_logits/chosen": -3.11570405960083,
      "eval_logits/rejected": -3.198913335800171,
      "eval_logps/chosen": -303.6178894042969,
      "eval_logps/rejected": -293.61297607421875,
      "eval_loss": 0.02255001664161682,
      "eval_rewards/accuracies": 0.9933837652206421,
      "eval_rewards/chosen": -0.9865113496780396,
      "eval_rewards/margins": 9.390168190002441,
      "eval_rewards/rejected": -10.376679420471191,
      "eval_runtime": 562.5858,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 1.881,
      "step": 2250
    },
    {
      "epoch": 2.2686214235043747,
      "grad_norm": 1.9584112167358398,
      "learning_rate": 6.215704824976348e-07,
      "logits/chosen": -3.115262985229492,
      "logits/rejected": -3.222729206085205,
      "logps/chosen": -299.4010009765625,
      "logps/rejected": -296.6014404296875,
      "loss": 0.0315,
      "rewards/accuracies": 0.9975000023841858,
      "rewards/chosen": -1.011648416519165,
      "rewards/margins": 9.797198295593262,
      "rewards/rejected": -10.808846473693848,
      "step": 2400
    },
    {
      "epoch": 2.2686214235043747,
      "eval_logits/chosen": -3.1395390033721924,
      "eval_logits/rejected": -3.2528908252716064,
      "eval_logps/chosen": -303.15643310546875,
      "eval_logps/rejected": -292.42431640625,
      "eval_loss": 0.019048307090997696,
      "eval_rewards/accuracies": 0.9948015213012695,
      "eval_rewards/chosen": -0.9403614401817322,
      "eval_rewards/margins": 9.317451477050781,
      "eval_rewards/rejected": -10.2578125,
      "eval_runtime": 563.9817,
      "eval_samples_per_second": 3.75,
      "eval_steps_per_second": 1.876,
      "step": 2400
    },
    {
      "epoch": 2.410498935918657,
      "grad_norm": 1.9372230768203735,
      "learning_rate": 5.979186376537369e-07,
      "logits/chosen": -3.1377475261688232,
      "logits/rejected": -3.264430046081543,
      "logps/chosen": -303.7994384765625,
      "logps/rejected": -298.91485595703125,
      "loss": 0.012,
      "rewards/accuracies": 0.9983333349227905,
      "rewards/chosen": -1.0909898281097412,
      "rewards/margins": 9.924967765808105,
      "rewards/rejected": -11.015958786010742,
      "step": 2550
    },
    {
      "epoch": 2.410498935918657,
      "eval_logits/chosen": -3.118379592895508,
      "eval_logits/rejected": -3.2569830417633057,
      "eval_logps/chosen": -298.316162109375,
      "eval_logps/rejected": -283.5343322753906,
      "eval_loss": 0.017369672656059265,
      "eval_rewards/accuracies": 0.996219277381897,
      "eval_rewards/chosen": -0.4563392996788025,
      "eval_rewards/margins": 8.912469863891602,
      "eval_rewards/rejected": -9.368810653686523,
      "eval_runtime": 563.7804,
      "eval_samples_per_second": 3.751,
      "eval_steps_per_second": 1.877,
      "step": 2550
    },
    {
      "epoch": 2.5523764483329394,
      "grad_norm": 2.599371910095215,
      "learning_rate": 5.742667928098392e-07,
      "logits/chosen": -3.1383397579193115,
      "logits/rejected": -3.241206169128418,
      "logps/chosen": -307.1145324707031,
      "logps/rejected": -304.4930114746094,
      "loss": 0.012,
      "rewards/accuracies": 0.9983333349227905,
      "rewards/chosen": -1.1029903888702393,
      "rewards/margins": 10.304451942443848,
      "rewards/rejected": -11.407442092895508,
      "step": 2700
    },
    {
      "epoch": 2.5523764483329394,
      "eval_logits/chosen": -3.127473831176758,
      "eval_logits/rejected": -3.245652198791504,
      "eval_logps/chosen": -301.93988037109375,
      "eval_logps/rejected": -291.9409484863281,
      "eval_loss": 0.01718134619295597,
      "eval_rewards/accuracies": 0.996219277381897,
      "eval_rewards/chosen": -0.8187108635902405,
      "eval_rewards/margins": 9.390762329101562,
      "eval_rewards/rejected": -10.209474563598633,
      "eval_runtime": 564.551,
      "eval_samples_per_second": 3.746,
      "eval_steps_per_second": 1.874,
      "step": 2700
    },
    {
      "epoch": 2.6942539607472216,
      "grad_norm": 0.546609103679657,
      "learning_rate": 5.506149479659413e-07,
      "logits/chosen": -3.126167058944702,
      "logits/rejected": -3.2429487705230713,
      "logps/chosen": -300.4310302734375,
      "logps/rejected": -300.19415283203125,
      "loss": 0.0325,
      "rewards/accuracies": 0.9975000023841858,
      "rewards/chosen": -0.9882186651229858,
      "rewards/margins": 10.136972427368164,
      "rewards/rejected": -11.125190734863281,
      "step": 2850
    },
    {
      "epoch": 2.6942539607472216,
      "eval_logits/chosen": -3.134120225906372,
      "eval_logits/rejected": -3.2371535301208496,
      "eval_logps/chosen": -302.8757629394531,
      "eval_logps/rejected": -298.766357421875,
      "eval_loss": 0.01783796213567257,
      "eval_rewards/accuracies": 0.9952741265296936,
      "eval_rewards/chosen": -0.9122979044914246,
      "eval_rewards/margins": 9.979716300964355,
      "eval_rewards/rejected": -10.892014503479004,
      "eval_runtime": 563.1261,
      "eval_samples_per_second": 3.756,
      "eval_steps_per_second": 1.879,
      "step": 2850
    },
    {
      "epoch": 2.8361314731615037,
      "grad_norm": 0.13827018439769745,
      "learning_rate": 5.269631031220436e-07,
      "logits/chosen": -3.1727616786956787,
      "logits/rejected": -3.2297608852386475,
      "logps/chosen": -307.07745361328125,
      "logps/rejected": -316.3017272949219,
      "loss": 0.0076,
      "rewards/accuracies": 0.9983333349227905,
      "rewards/chosen": -1.8074291944503784,
      "rewards/margins": 11.044861793518066,
      "rewards/rejected": -12.852290153503418,
      "step": 3000
    },
    {
      "epoch": 2.8361314731615037,
      "eval_logits/chosen": -3.1847798824310303,
      "eval_logits/rejected": -3.0981080532073975,
      "eval_logps/chosen": -336.5445861816406,
      "eval_logps/rejected": -362.6168212890625,
      "eval_loss": 0.037136226892471313,
      "eval_rewards/accuracies": 0.9867674708366394,
      "eval_rewards/chosen": -4.279180526733398,
      "eval_rewards/margins": 12.997879981994629,
      "eval_rewards/rejected": -17.277057647705078,
      "eval_runtime": 765.7084,
      "eval_samples_per_second": 2.762,
      "eval_steps_per_second": 1.382,
      "step": 3000
    }
  ],
  "logging_steps": 150,
  "max_steps": 6342,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 150,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}