OpenRS-GRPO3 / trainer_state.json
xiwenc1's picture
Model save
270a755 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.5714285714285714,
"eval_steps": 500,
"global_step": 500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"completion_length": 3001.9584350585938,
"epoch": 0.001142857142857143,
"grad_norm": 0.18922260403633118,
"kl": 0.0,
"learning_rate": 2e-08,
"loss": -0.0,
"reward": -0.010712452232837677,
"reward_std": 0.48354096710681915,
"rewards/cosine_scaled_reward": -0.1928562317043543,
"rewards/format_reward": 0.37500000558793545,
"step": 1
},
{
"completion_length": 2822.541717529297,
"epoch": 0.002285714285714286,
"grad_norm": 0.28424975275993347,
"kl": 0.0,
"learning_rate": 4e-08,
"loss": 0.0,
"reward": 0.4385625521535985,
"reward_std": 0.8208381980657578,
"rewards/cosine_scaled_reward": -0.009885392151772976,
"rewards/format_reward": 0.4583333432674408,
"step": 2
},
{
"completion_length": 2882.4166870117188,
"epoch": 0.0034285714285714284,
"grad_norm": 0.18410934507846832,
"kl": 3.517171717248857e-05,
"learning_rate": 6e-08,
"loss": 0.0,
"reward": -0.291525443084538,
"reward_std": 0.3761885389685631,
"rewards/cosine_scaled_reward": -0.27076271921396255,
"rewards/format_reward": 0.25,
"step": 3
},
{
"completion_length": 3245.416748046875,
"epoch": 0.004571428571428572,
"grad_norm": 0.16615347564220428,
"kl": 2.9280781745910645e-05,
"learning_rate": 8e-08,
"loss": 0.0,
"reward": -0.25264428183436394,
"reward_std": 0.4561151713132858,
"rewards/cosine_scaled_reward": -0.24090547114610672,
"rewards/format_reward": 0.22916667349636555,
"step": 4
},
{
"completion_length": 2911.3334350585938,
"epoch": 0.005714285714285714,
"grad_norm": 0.21166956424713135,
"kl": 3.331899642944336e-05,
"learning_rate": 1e-07,
"loss": 0.0,
"reward": 0.8400040492415428,
"reward_std": 0.885560505092144,
"rewards/cosine_scaled_reward": 0.18041866831481457,
"rewards/format_reward": 0.479166679084301,
"step": 5
},
{
"completion_length": 2720.89599609375,
"epoch": 0.006857142857142857,
"grad_norm": 0.23326514661312103,
"kl": 4.035234451293945e-05,
"learning_rate": 1.2e-07,
"loss": 0.0,
"reward": 0.2041575275361538,
"reward_std": 0.6658071056008339,
"rewards/cosine_scaled_reward": -0.11667125090025365,
"rewards/format_reward": 0.4375,
"step": 6
},
{
"completion_length": 2360.6458740234375,
"epoch": 0.008,
"grad_norm": 0.2280312329530716,
"kl": 1.850724220275879e-05,
"learning_rate": 1.4e-07,
"loss": 0.0,
"reward": 0.7341702952980995,
"reward_std": 0.44598812609910965,
"rewards/cosine_scaled_reward": 0.09625181555747986,
"rewards/format_reward": 0.5416666865348816,
"step": 7
},
{
"completion_length": 2888.8750610351562,
"epoch": 0.009142857142857144,
"grad_norm": 0.20181676745414734,
"kl": 2.9474496841430664e-05,
"learning_rate": 1.6e-07,
"loss": 0.0,
"reward": -0.007411351427435875,
"reward_std": 0.6588219478726387,
"rewards/cosine_scaled_reward": -0.2016223482787609,
"rewards/format_reward": 0.3958333395421505,
"step": 8
},
{
"completion_length": 3309.8541870117188,
"epoch": 0.010285714285714285,
"grad_norm": 0.18089492619037628,
"kl": 4.0411949157714844e-05,
"learning_rate": 1.8e-07,
"loss": 0.0,
"reward": 0.08656559139490128,
"reward_std": 0.7023323476314545,
"rewards/cosine_scaled_reward": -0.060883864760398865,
"rewards/format_reward": 0.2083333358168602,
"step": 9
},
{
"completion_length": 2354.729217529297,
"epoch": 0.011428571428571429,
"grad_norm": 0.2087497115135193,
"kl": 3.822147846221924e-05,
"learning_rate": 2e-07,
"loss": 0.0,
"reward": 0.43211155757308006,
"reward_std": 0.7549905180931091,
"rewards/cosine_scaled_reward": -0.054777566343545914,
"rewards/format_reward": 0.5416666679084301,
"step": 10
},
{
"completion_length": 2589.041778564453,
"epoch": 0.012571428571428572,
"grad_norm": 0.2392556220293045,
"kl": 4.547834396362305e-05,
"learning_rate": 2.1999999999999998e-07,
"loss": 0.0,
"reward": 0.570576427038759,
"reward_std": 1.0850151628255844,
"rewards/cosine_scaled_reward": -0.006378462538123131,
"rewards/format_reward": 0.5833333432674408,
"step": 11
},
{
"completion_length": 2472.4375610351562,
"epoch": 0.013714285714285714,
"grad_norm": 0.2759428322315216,
"kl": 2.5950372219085693e-05,
"learning_rate": 2.4e-07,
"loss": 0.0,
"reward": 0.7166856527328491,
"reward_std": 0.7806050479412079,
"rewards/cosine_scaled_reward": 0.056259457021951675,
"rewards/format_reward": 0.6041666716337204,
"step": 12
},
{
"completion_length": 2164.854232788086,
"epoch": 0.014857142857142857,
"grad_norm": 0.1908554881811142,
"kl": 2.492964267730713e-05,
"learning_rate": 2.6e-07,
"loss": 0.0,
"reward": 0.7641210779547691,
"reward_std": 0.6774896830320358,
"rewards/cosine_scaled_reward": 0.10081052035093307,
"rewards/format_reward": 0.5625,
"step": 13
},
{
"completion_length": 2820.2501220703125,
"epoch": 0.016,
"grad_norm": 0.18801091611385345,
"kl": 3.5822391510009766e-05,
"learning_rate": 2.8e-07,
"loss": 0.0,
"reward": -0.07261240109801292,
"reward_std": 0.6130082383751869,
"rewards/cosine_scaled_reward": -0.2133895456790924,
"rewards/format_reward": 0.3541666753590107,
"step": 14
},
{
"completion_length": 3089.104248046875,
"epoch": 0.017142857142857144,
"grad_norm": 0.19688484072685242,
"kl": 3.3468008041381836e-05,
"learning_rate": 3e-07,
"loss": 0.0,
"reward": 0.27143352539860643,
"reward_std": 0.8015548288822174,
"rewards/cosine_scaled_reward": 0.0003000907599925995,
"rewards/format_reward": 0.2708333469927311,
"step": 15
},
{
"completion_length": 2362.6876220703125,
"epoch": 0.018285714285714287,
"grad_norm": 0.24537329375743866,
"kl": 2.1159648895263672e-05,
"learning_rate": 3.2e-07,
"loss": 0.0,
"reward": 0.7649998441338539,
"reward_std": 1.021081954240799,
"rewards/cosine_scaled_reward": 0.11166658625006676,
"rewards/format_reward": 0.5416666865348816,
"step": 16
},
{
"completion_length": 3128.25,
"epoch": 0.019428571428571427,
"grad_norm": 0.20609241724014282,
"kl": 4.242360591888428e-05,
"learning_rate": 3.4000000000000003e-07,
"loss": 0.0,
"reward": -0.19378644227981567,
"reward_std": 0.5115947872400284,
"rewards/cosine_scaled_reward": -0.21147656068205833,
"rewards/format_reward": 0.2291666716337204,
"step": 17
},
{
"completion_length": 2980.9793090820312,
"epoch": 0.02057142857142857,
"grad_norm": 0.2714909315109253,
"kl": 3.966689109802246e-05,
"learning_rate": 3.6e-07,
"loss": 0.0,
"reward": -0.022395100444555283,
"reward_std": 0.6723635420203209,
"rewards/cosine_scaled_reward": -0.16744754649698734,
"rewards/format_reward": 0.3125000037252903,
"step": 18
},
{
"completion_length": 3212.604248046875,
"epoch": 0.021714285714285714,
"grad_norm": 0.17160819470882416,
"kl": 3.719329833984375e-05,
"learning_rate": 3.7999999999999996e-07,
"loss": 0.0,
"reward": 0.07102994620800018,
"reward_std": 0.8850104063749313,
"rewards/cosine_scaled_reward": -0.1207350417971611,
"rewards/format_reward": 0.31250001303851604,
"step": 19
},
{
"completion_length": 2613.6041870117188,
"epoch": 0.022857142857142857,
"grad_norm": 0.24837209284305573,
"kl": 3.3915042877197266e-05,
"learning_rate": 4e-07,
"loss": 0.0,
"reward": 0.33390188589692116,
"reward_std": 0.713263601064682,
"rewards/cosine_scaled_reward": -0.08304904773831367,
"rewards/format_reward": 0.5000000149011612,
"step": 20
},
{
"completion_length": 2582.041748046875,
"epoch": 0.024,
"grad_norm": 0.2642858326435089,
"kl": 2.1037645637989044e-05,
"learning_rate": 4.1999999999999995e-07,
"loss": 0.0,
"reward": 0.2965797185897827,
"reward_std": 0.5356749221682549,
"rewards/cosine_scaled_reward": -0.06004347978159785,
"rewards/format_reward": 0.41666667722165585,
"step": 21
},
{
"completion_length": 3307.6459350585938,
"epoch": 0.025142857142857144,
"grad_norm": 0.22419147193431854,
"kl": 4.1961669921875e-05,
"learning_rate": 4.3999999999999997e-07,
"loss": 0.0,
"reward": 0.25835999101400375,
"reward_std": 1.1261206567287445,
"rewards/cosine_scaled_reward": -0.04790334962308407,
"rewards/format_reward": 0.3541666716337204,
"step": 22
},
{
"completion_length": 3249.1876220703125,
"epoch": 0.026285714285714287,
"grad_norm": 0.19173863530158997,
"kl": 4.4405460357666016e-05,
"learning_rate": 4.6e-07,
"loss": 0.0,
"reward": 0.27271851897239685,
"reward_std": 0.7990642189979553,
"rewards/cosine_scaled_reward": -0.04072406329214573,
"rewards/format_reward": 0.3541666828095913,
"step": 23
},
{
"completion_length": 2154.25,
"epoch": 0.027428571428571427,
"grad_norm": 0.259212851524353,
"kl": 1.8768012523651123e-05,
"learning_rate": 4.8e-07,
"loss": 0.0,
"reward": 0.5731075070798397,
"reward_std": 0.8421577215194702,
"rewards/cosine_scaled_reward": -0.02594624925404787,
"rewards/format_reward": 0.6250000074505806,
"step": 24
},
{
"completion_length": 2784.7916870117188,
"epoch": 0.02857142857142857,
"grad_norm": 0.29162946343421936,
"kl": 3.090500831604004e-05,
"learning_rate": 5e-07,
"loss": 0.0,
"reward": 0.1705078724771738,
"reward_std": 0.6685621440410614,
"rewards/cosine_scaled_reward": -0.08141273260116577,
"rewards/format_reward": 0.3333333358168602,
"step": 25
},
{
"completion_length": 3185.729248046875,
"epoch": 0.029714285714285714,
"grad_norm": 0.15754370391368866,
"kl": 2.549588680267334e-05,
"learning_rate": 5.2e-07,
"loss": 0.0,
"reward": 0.13516026688739657,
"reward_std": 0.6272664293646812,
"rewards/cosine_scaled_reward": -0.05741987004876137,
"rewards/format_reward": 0.2500000111758709,
"step": 26
},
{
"completion_length": 3129.2083740234375,
"epoch": 0.030857142857142857,
"grad_norm": 0.16376672685146332,
"kl": 2.86102294921875e-05,
"learning_rate": 5.4e-07,
"loss": 0.0,
"reward": 0.0651654414832592,
"reward_std": 0.5805819556117058,
"rewards/cosine_scaled_reward": -0.08200062438845634,
"rewards/format_reward": 0.22916667722165585,
"step": 27
},
{
"completion_length": 3173.4791870117188,
"epoch": 0.032,
"grad_norm": 0.2187095433473587,
"kl": 3.802776336669922e-05,
"learning_rate": 5.6e-07,
"loss": 0.0,
"reward": 0.1244891807436943,
"reward_std": 0.8137174546718597,
"rewards/cosine_scaled_reward": -0.07317209523171186,
"rewards/format_reward": 0.27083334885537624,
"step": 28
},
{
"completion_length": 3206.0208740234375,
"epoch": 0.03314285714285714,
"grad_norm": 0.15626287460327148,
"kl": 1.7024576663970947e-05,
"learning_rate": 5.8e-07,
"loss": 0.0,
"reward": -0.0882865646854043,
"reward_std": 0.6182056441903114,
"rewards/cosine_scaled_reward": -0.13789328234270215,
"rewards/format_reward": 0.1875000111758709,
"step": 29
},
{
"completion_length": 3293.979248046875,
"epoch": 0.03428571428571429,
"grad_norm": 0.176454558968544,
"kl": 2.8930604457855225e-05,
"learning_rate": 6e-07,
"loss": 0.0,
"reward": 0.12017922103404999,
"reward_std": 0.8002806901931763,
"rewards/cosine_scaled_reward": -0.10657705180346966,
"rewards/format_reward": 0.33333334885537624,
"step": 30
},
{
"completion_length": 2556.0625610351562,
"epoch": 0.03542857142857143,
"grad_norm": 0.2976699471473694,
"kl": 3.6090612411499023e-05,
"learning_rate": 6.2e-07,
"loss": 0.0,
"reward": 0.13020960986614227,
"reward_std": 0.5589020624756813,
"rewards/cosine_scaled_reward": -0.15364519506692886,
"rewards/format_reward": 0.43750001303851604,
"step": 31
},
{
"completion_length": 3466.125,
"epoch": 0.036571428571428574,
"grad_norm": 0.15761366486549377,
"kl": 3.0994415283203125e-05,
"learning_rate": 6.4e-07,
"loss": 0.0,
"reward": -0.34774322621524334,
"reward_std": 0.4613388404250145,
"rewards/cosine_scaled_reward": -0.22595495358109474,
"rewards/format_reward": 0.1041666679084301,
"step": 32
},
{
"completion_length": 3078.729248046875,
"epoch": 0.037714285714285714,
"grad_norm": 0.17744146287441254,
"kl": 1.9311904907226562e-05,
"learning_rate": 6.6e-07,
"loss": 0.0,
"reward": 0.1523735709488392,
"reward_std": 0.7702403217554092,
"rewards/cosine_scaled_reward": -0.10089654847979546,
"rewards/format_reward": 0.354166679084301,
"step": 33
},
{
"completion_length": 3068.2083740234375,
"epoch": 0.038857142857142854,
"grad_norm": 0.2183830887079239,
"kl": 1.940131187438965e-05,
"learning_rate": 6.800000000000001e-07,
"loss": 0.0,
"reward": 0.029434625059366226,
"reward_std": 0.7817529812455177,
"rewards/cosine_scaled_reward": -0.1415326923597604,
"rewards/format_reward": 0.3125000111758709,
"step": 34
},
{
"completion_length": 3028.916748046875,
"epoch": 0.04,
"grad_norm": 0.1779097616672516,
"kl": 2.0578503608703613e-05,
"learning_rate": 7e-07,
"loss": 0.0,
"reward": 0.2327469252049923,
"reward_std": 0.9538670182228088,
"rewards/cosine_scaled_reward": -0.09195987693965435,
"rewards/format_reward": 0.41666667349636555,
"step": 35
},
{
"completion_length": 2689.3959350585938,
"epoch": 0.04114285714285714,
"grad_norm": 0.16330143809318542,
"kl": 4.976987838745117e-05,
"learning_rate": 7.2e-07,
"loss": 0.0,
"reward": 0.5622920989990234,
"reward_std": 0.39920446276664734,
"rewards/cosine_scaled_reward": 0.07281268946826458,
"rewards/format_reward": 0.4166666716337204,
"step": 36
},
{
"completion_length": 2801.6459045410156,
"epoch": 0.04228571428571429,
"grad_norm": 0.19838035106658936,
"kl": 4.4226646423339844e-05,
"learning_rate": 7.4e-07,
"loss": 0.0,
"reward": 0.41371238604187965,
"reward_std": 0.575165145099163,
"rewards/cosine_scaled_reward": -0.07439382094889879,
"rewards/format_reward": 0.5625000149011612,
"step": 37
},
{
"completion_length": 3009.5000610351562,
"epoch": 0.04342857142857143,
"grad_norm": 0.1800134778022766,
"kl": 6.097555160522461e-05,
"learning_rate": 7.599999999999999e-07,
"loss": 0.0,
"reward": 0.12262389855459332,
"reward_std": 0.6652626991271973,
"rewards/cosine_scaled_reward": -0.15743806213140488,
"rewards/format_reward": 0.4375000149011612,
"step": 38
},
{
"completion_length": 3139.604248046875,
"epoch": 0.044571428571428574,
"grad_norm": 0.23411938548088074,
"kl": 4.2323023080825806e-05,
"learning_rate": 7.799999999999999e-07,
"loss": 0.0,
"reward": 0.24589091911911964,
"reward_std": 0.8911770880222321,
"rewards/cosine_scaled_reward": -0.07497121207416058,
"rewards/format_reward": 0.3958333395421505,
"step": 39
},
{
"completion_length": 3011.8958740234375,
"epoch": 0.045714285714285714,
"grad_norm": 0.1625184565782547,
"kl": 4.693865776062012e-05,
"learning_rate": 8e-07,
"loss": 0.0,
"reward": 0.12772860191762447,
"reward_std": 0.7576778829097748,
"rewards/cosine_scaled_reward": -0.12363571301102638,
"rewards/format_reward": 0.3750000037252903,
"step": 40
},
{
"completion_length": 3124.2500610351562,
"epoch": 0.046857142857142854,
"grad_norm": 0.17387458682060242,
"kl": 1.8164515495300293e-05,
"learning_rate": 8.199999999999999e-07,
"loss": 0.0,
"reward": 0.21644378546625376,
"reward_std": 0.6694483831524849,
"rewards/cosine_scaled_reward": -0.06886144913733006,
"rewards/format_reward": 0.354166679084301,
"step": 41
},
{
"completion_length": 2182.5000610351562,
"epoch": 0.048,
"grad_norm": 0.34351351857185364,
"kl": 0.00022931396961212158,
"learning_rate": 8.399999999999999e-07,
"loss": 0.0,
"reward": 0.6580724753439426,
"reward_std": 0.8123672902584076,
"rewards/cosine_scaled_reward": 0.016536223702132702,
"rewards/format_reward": 0.6250000111758709,
"step": 42
},
{
"completion_length": 2828.7709350585938,
"epoch": 0.04914285714285714,
"grad_norm": 0.19178950786590576,
"kl": 3.0837953090667725e-05,
"learning_rate": 8.599999999999999e-07,
"loss": 0.0,
"reward": 0.31538891326636076,
"reward_std": 0.8877717405557632,
"rewards/cosine_scaled_reward": -0.06105554662644863,
"rewards/format_reward": 0.4375000149011612,
"step": 43
},
{
"completion_length": 3036.6250610351562,
"epoch": 0.05028571428571429,
"grad_norm": 0.22349753975868225,
"kl": 0.0003269314765930176,
"learning_rate": 8.799999999999999e-07,
"loss": 0.0,
"reward": -0.01474527781829238,
"reward_std": 0.5097765326499939,
"rewards/cosine_scaled_reward": -0.15320597402751446,
"rewards/format_reward": 0.291666679084301,
"step": 44
},
{
"completion_length": 2778.5000610351562,
"epoch": 0.05142857142857143,
"grad_norm": 0.18280261754989624,
"kl": 5.6609511375427246e-05,
"learning_rate": 9e-07,
"loss": 0.0,
"reward": 0.8612850233912468,
"reward_std": 1.1412108689546585,
"rewards/cosine_scaled_reward": 0.13897587358951569,
"rewards/format_reward": 0.583333358168602,
"step": 45
},
{
"completion_length": 2998.229248046875,
"epoch": 0.052571428571428575,
"grad_norm": 0.16858145594596863,
"kl": 5.383044481277466e-05,
"learning_rate": 9.2e-07,
"loss": 0.0,
"reward": 0.28540395572781563,
"reward_std": 0.43213801458477974,
"rewards/cosine_scaled_reward": -0.02396468259394169,
"rewards/format_reward": 0.3333333358168602,
"step": 46
},
{
"completion_length": 2551.6458587646484,
"epoch": 0.053714285714285714,
"grad_norm": 0.23799559473991394,
"kl": 9.156018495559692e-05,
"learning_rate": 9.399999999999999e-07,
"loss": 0.0,
"reward": 0.44500812888145447,
"reward_std": 0.783466711640358,
"rewards/cosine_scaled_reward": -0.04832928255200386,
"rewards/format_reward": 0.5416666865348816,
"step": 47
},
{
"completion_length": 2939.1041870117188,
"epoch": 0.054857142857142854,
"grad_norm": 0.18564291298389435,
"kl": 0.00010335445404052734,
"learning_rate": 9.6e-07,
"loss": 0.0,
"reward": 0.34564049541950226,
"reward_std": 0.9494538530707359,
"rewards/cosine_scaled_reward": -0.02509642019867897,
"rewards/format_reward": 0.3958333395421505,
"step": 48
},
{
"completion_length": 2282.895835876465,
"epoch": 0.056,
"grad_norm": 0.23054172098636627,
"kl": 0.00024145841598510742,
"learning_rate": 9.8e-07,
"loss": 0.0,
"reward": 0.44885391741991043,
"reward_std": 0.7631752789020538,
"rewards/cosine_scaled_reward": -0.06723971478641033,
"rewards/format_reward": 0.5833333358168602,
"step": 49
},
{
"completion_length": 2204.1041870117188,
"epoch": 0.05714285714285714,
"grad_norm": 0.29597678780555725,
"kl": 0.0005988925695419312,
"learning_rate": 1e-06,
"loss": 0.0,
"reward": 0.5718545913696289,
"reward_std": 0.6146985068917274,
"rewards/cosine_scaled_reward": 0.02551062125712633,
"rewards/format_reward": 0.5208333432674408,
"step": 50
},
{
"completion_length": 3201.6666870117188,
"epoch": 0.05828571428571429,
"grad_norm": 0.15334689617156982,
"kl": 8.338689804077148e-05,
"learning_rate": 9.999890338174275e-07,
"loss": 0.0,
"reward": 0.14526839554309845,
"reward_std": 0.8655073121190071,
"rewards/cosine_scaled_reward": -0.07319913152605295,
"rewards/format_reward": 0.2916666716337204,
"step": 51
},
{
"completion_length": 3215.0000610351562,
"epoch": 0.05942857142857143,
"grad_norm": 0.17531076073646545,
"kl": 0.0001531541347503662,
"learning_rate": 9.999561358041868e-07,
"loss": 0.0,
"reward": -0.015788130462169647,
"reward_std": 0.7165202274918556,
"rewards/cosine_scaled_reward": -0.13289407594129443,
"rewards/format_reward": 0.2500000074505806,
"step": 52
},
{
"completion_length": 2991.8125610351562,
"epoch": 0.060571428571428575,
"grad_norm": 0.28014782071113586,
"kl": 0.000295490026473999,
"learning_rate": 9.999013075636804e-07,
"loss": 0.0,
"reward": -0.08945630304515362,
"reward_std": 0.6164149194955826,
"rewards/cosine_scaled_reward": -0.1905614770948887,
"rewards/format_reward": 0.2916666828095913,
"step": 53
},
{
"completion_length": 2437.4375,
"epoch": 0.061714285714285715,
"grad_norm": 0.21041618287563324,
"kl": 0.00014019012451171875,
"learning_rate": 9.998245517681593e-07,
"loss": 0.0,
"reward": 0.08023202046751976,
"reward_std": 0.43379058688879013,
"rewards/cosine_scaled_reward": -0.18905067443847656,
"rewards/format_reward": 0.4583333432674408,
"step": 54
},
{
"completion_length": 3115.2501220703125,
"epoch": 0.06285714285714286,
"grad_norm": 0.18965038657188416,
"kl": 0.00021153688430786133,
"learning_rate": 9.997258721585931e-07,
"loss": 0.0,
"reward": 0.14397013932466507,
"reward_std": 0.7335802316665649,
"rewards/cosine_scaled_reward": -0.07384827360510826,
"rewards/format_reward": 0.291666679084301,
"step": 55
},
{
"completion_length": 3097.0625,
"epoch": 0.064,
"grad_norm": 0.1861303448677063,
"kl": 0.0006959438323974609,
"learning_rate": 9.996052735444862e-07,
"loss": 0.0,
"reward": 0.21384139358997345,
"reward_std": 0.7854617610573769,
"rewards/cosine_scaled_reward": -0.038912639021873474,
"rewards/format_reward": 0.29166667722165585,
"step": 56
},
{
"completion_length": 2959.479248046875,
"epoch": 0.06514285714285714,
"grad_norm": 0.17092932760715485,
"kl": 0.0004864931106567383,
"learning_rate": 9.994627618036452e-07,
"loss": 0.0,
"reward": -0.0896548442542553,
"reward_std": 0.597286906093359,
"rewards/cosine_scaled_reward": -0.22191076539456844,
"rewards/format_reward": 0.3541666716337204,
"step": 57
},
{
"completion_length": 3173.9583740234375,
"epoch": 0.06628571428571428,
"grad_norm": 0.16764329373836517,
"kl": 0.0009320974349975586,
"learning_rate": 9.992983438818915e-07,
"loss": 0.0,
"reward": -0.259409268386662,
"reward_std": 0.41252440214157104,
"rewards/cosine_scaled_reward": -0.22345462441444397,
"rewards/format_reward": 0.1875,
"step": 58
},
{
"completion_length": 2903.4375610351562,
"epoch": 0.06742857142857143,
"grad_norm": 0.2438689023256302,
"kl": 0.0004966259002685547,
"learning_rate": 9.991120277927223e-07,
"loss": 0.0,
"reward": 0.5037799216806889,
"reward_std": 0.6180380508303642,
"rewards/cosine_scaled_reward": 0.05397331342101097,
"rewards/format_reward": 0.39583334513008595,
"step": 59
},
{
"completion_length": 2846.9375610351562,
"epoch": 0.06857142857142857,
"grad_norm": 0.19560052454471588,
"kl": 0.0006766319274902344,
"learning_rate": 9.989038226169207e-07,
"loss": 0.0,
"reward": 0.619435504078865,
"reward_std": 0.54927659034729,
"rewards/cosine_scaled_reward": 0.12221772782504559,
"rewards/format_reward": 0.3750000149011612,
"step": 60
},
{
"completion_length": 2432.729248046875,
"epoch": 0.06971428571428571,
"grad_norm": 0.18966424465179443,
"kl": 0.00019347667694091797,
"learning_rate": 9.98673738502114e-07,
"loss": 0.0,
"reward": 0.8035399168729782,
"reward_std": 0.6529746800661087,
"rewards/cosine_scaled_reward": 0.12051995098590851,
"rewards/format_reward": 0.5625000149011612,
"step": 61
},
{
"completion_length": 3356.4583740234375,
"epoch": 0.07085714285714285,
"grad_norm": 0.1480625420808792,
"kl": 0.0005307793617248535,
"learning_rate": 9.98421786662277e-07,
"loss": 0.0,
"reward": 0.33844682574272156,
"reward_std": 0.7905219346284866,
"rewards/cosine_scaled_reward": 0.023390088230371475,
"rewards/format_reward": 0.29166667722165585,
"step": 62
},
{
"completion_length": 1956.7084350585938,
"epoch": 0.072,
"grad_norm": 0.26979929208755493,
"kl": 0.006456255912780762,
"learning_rate": 9.981479793771866e-07,
"loss": 0.0003,
"reward": 0.7974750846624374,
"reward_std": 0.7315621674060822,
"rewards/cosine_scaled_reward": 0.05498753860592842,
"rewards/format_reward": 0.6875000149011612,
"step": 63
},
{
"completion_length": 2895.8958740234375,
"epoch": 0.07314285714285715,
"grad_norm": 0.14358438551425934,
"kl": 0.00026297569274902344,
"learning_rate": 9.97852329991824e-07,
"loss": 0.0,
"reward": 0.6833263337612152,
"reward_std": 0.5810272544622421,
"rewards/cosine_scaled_reward": 0.12291315197944641,
"rewards/format_reward": 0.4375000149011612,
"step": 64
},
{
"completion_length": 2582.1875610351562,
"epoch": 0.07428571428571429,
"grad_norm": 0.1719641089439392,
"kl": 0.0006394386291503906,
"learning_rate": 9.975348529157229e-07,
"loss": 0.0,
"reward": 0.4943835213780403,
"reward_std": 0.9102021306753159,
"rewards/cosine_scaled_reward": -0.002808244898915291,
"rewards/format_reward": 0.5000000074505806,
"step": 65
},
{
"completion_length": 3066.2709350585938,
"epoch": 0.07542857142857143,
"grad_norm": 0.15914808213710785,
"kl": 0.00035455822944641113,
"learning_rate": 9.971955636222684e-07,
"loss": 0.0,
"reward": 0.270260289311409,
"reward_std": 0.6933658719062805,
"rewards/cosine_scaled_reward": -0.10445320140570402,
"rewards/format_reward": 0.47916667722165585,
"step": 66
},
{
"completion_length": 3021.1666870117188,
"epoch": 0.07657142857142857,
"grad_norm": 0.15889614820480347,
"kl": 0.0007028579711914062,
"learning_rate": 9.968344786479415e-07,
"loss": 0.0,
"reward": 0.2754221335053444,
"reward_std": 0.6702268719673157,
"rewards/cosine_scaled_reward": -0.028955606278032064,
"rewards/format_reward": 0.33333334140479565,
"step": 67
},
{
"completion_length": 2520.7291870117188,
"epoch": 0.07771428571428571,
"grad_norm": 0.25743117928504944,
"kl": 0.001796722412109375,
"learning_rate": 9.964516155915151e-07,
"loss": 0.0001,
"reward": 0.2481890171766281,
"reward_std": 0.4365886226296425,
"rewards/cosine_scaled_reward": -0.06340551376342773,
"rewards/format_reward": 0.375,
"step": 68
},
{
"completion_length": 2519.2500610351562,
"epoch": 0.07885714285714286,
"grad_norm": 0.1728557050228119,
"kl": 0.0004031658172607422,
"learning_rate": 9.960469931131936e-07,
"loss": 0.0,
"reward": 0.4316702373325825,
"reward_std": 0.455346904695034,
"rewards/cosine_scaled_reward": -0.02374822273850441,
"rewards/format_reward": 0.4791666716337204,
"step": 69
},
{
"completion_length": 3165.7708740234375,
"epoch": 0.08,
"grad_norm": 0.13548843562602997,
"kl": 0.0004048347473144531,
"learning_rate": 9.956206309337066e-07,
"loss": 0.0,
"reward": 0.46206507831811905,
"reward_std": 0.8794394135475159,
"rewards/cosine_scaled_reward": -0.008550799917429686,
"rewards/format_reward": 0.4791666716337204,
"step": 70
},
{
"completion_length": 2529.8751220703125,
"epoch": 0.08114285714285714,
"grad_norm": 0.22490063309669495,
"kl": 0.003349781036376953,
"learning_rate": 9.951725498333448e-07,
"loss": 0.0001,
"reward": 0.5197920426726341,
"reward_std": 0.823016032576561,
"rewards/cosine_scaled_reward": -0.010937327519059181,
"rewards/format_reward": 0.5416666865348816,
"step": 71
},
{
"completion_length": 2654.4791870117188,
"epoch": 0.08228571428571428,
"grad_norm": 0.20955628156661987,
"kl": 0.0012530684471130371,
"learning_rate": 9.947027716509488e-07,
"loss": 0.0001,
"reward": 0.7414695173501968,
"reward_std": 0.5663128644227982,
"rewards/cosine_scaled_reward": 0.05823474656790495,
"rewards/format_reward": 0.625,
"step": 72
},
{
"completion_length": 1640.041748046875,
"epoch": 0.08342857142857144,
"grad_norm": 0.2790416479110718,
"kl": 0.005096435546875,
"learning_rate": 9.942113192828444e-07,
"loss": 0.0002,
"reward": 0.91811203956604,
"reward_std": 0.8141122311353683,
"rewards/cosine_scaled_reward": 0.08405601512640715,
"rewards/format_reward": 0.75,
"step": 73
},
{
"completion_length": 2470.5833740234375,
"epoch": 0.08457142857142858,
"grad_norm": 0.1816757470369339,
"kl": 0.001283407211303711,
"learning_rate": 9.93698216681727e-07,
"loss": 0.0001,
"reward": 0.595103541854769,
"reward_std": 0.6585821881890297,
"rewards/cosine_scaled_reward": 0.01630176231265068,
"rewards/format_reward": 0.5625000074505806,
"step": 74
},
{
"completion_length": 2660.8750915527344,
"epoch": 0.08571428571428572,
"grad_norm": 0.2641614079475403,
"kl": 0.0019817352294921875,
"learning_rate": 9.931634888554935e-07,
"loss": 0.0001,
"reward": 0.2930721901357174,
"reward_std": 0.7745417281985283,
"rewards/cosine_scaled_reward": -0.06179725006222725,
"rewards/format_reward": 0.4166666679084301,
"step": 75
},
{
"completion_length": 2939.604248046875,
"epoch": 0.08685714285714285,
"grad_norm": 0.3027961552143097,
"kl": 0.0008873939514160156,
"learning_rate": 9.926071618660237e-07,
"loss": 0.0,
"reward": 0.3248658664524555,
"reward_std": 0.909210205078125,
"rewards/cosine_scaled_reward": -0.014650408178567886,
"rewards/format_reward": 0.3541666828095913,
"step": 76
},
{
"completion_length": 3152.1666870117188,
"epoch": 0.088,
"grad_norm": 0.16062307357788086,
"kl": 0.001007080078125,
"learning_rate": 9.9202926282791e-07,
"loss": 0.0,
"reward": 0.11907588690519333,
"reward_std": 0.613786868751049,
"rewards/cosine_scaled_reward": -0.06546205282211304,
"rewards/format_reward": 0.25000000558793545,
"step": 77
},
{
"completion_length": 2635.041717529297,
"epoch": 0.08914285714285715,
"grad_norm": 0.19166067242622375,
"kl": 0.0008903741836547852,
"learning_rate": 9.91429819907136e-07,
"loss": 0.0,
"reward": 0.5825019255280495,
"reward_std": 0.7261854261159897,
"rewards/cosine_scaled_reward": 0.06208430230617523,
"rewards/format_reward": 0.4583333507180214,
"step": 78
},
{
"completion_length": 2755.5209350585938,
"epoch": 0.09028571428571429,
"grad_norm": 0.17263904213905334,
"kl": 0.0004019737243652344,
"learning_rate": 9.908088623197048e-07,
"loss": 0.0,
"reward": 0.3458161875605583,
"reward_std": 0.717200756072998,
"rewards/cosine_scaled_reward": -0.056258589029312134,
"rewards/format_reward": 0.4583333432674408,
"step": 79
},
{
"completion_length": 3565.4375,
"epoch": 0.09142857142857143,
"grad_norm": 0.1458665281534195,
"kl": 0.00030177831649780273,
"learning_rate": 9.901664203302124e-07,
"loss": 0.0,
"reward": -0.254756236448884,
"reward_std": 0.5783224925398827,
"rewards/cosine_scaled_reward": -0.17946145310997963,
"rewards/format_reward": 0.10416666977107525,
"step": 80
},
{
"completion_length": 3124.291748046875,
"epoch": 0.09257142857142857,
"grad_norm": 0.17081034183502197,
"kl": 0.0008420944213867188,
"learning_rate": 9.895025252503755e-07,
"loss": 0.0,
"reward": 0.06758889555931091,
"reward_std": 0.7439121454954147,
"rewards/cosine_scaled_reward": -0.10162222757935524,
"rewards/format_reward": 0.27083333767950535,
"step": 81
},
{
"completion_length": 3004.0833740234375,
"epoch": 0.09371428571428571,
"grad_norm": 0.17911851406097412,
"kl": 0.0006622076034545898,
"learning_rate": 9.888172094375033e-07,
"loss": 0.0,
"reward": 0.3707614615559578,
"reward_std": 0.8215866684913635,
"rewards/cosine_scaled_reward": -0.0437859346420737,
"rewards/format_reward": 0.45833334885537624,
"step": 82
},
{
"completion_length": 3440.3125,
"epoch": 0.09485714285714286,
"grad_norm": 0.15295840799808502,
"kl": 0.00017774105072021484,
"learning_rate": 9.881105062929221e-07,
"loss": 0.0,
"reward": -0.39509791135787964,
"reward_std": 0.5668186843395233,
"rewards/cosine_scaled_reward": -0.260048970580101,
"rewards/format_reward": 0.125,
"step": 83
},
{
"completion_length": 2530.7083740234375,
"epoch": 0.096,
"grad_norm": 0.2006414830684662,
"kl": 0.0003807544708251953,
"learning_rate": 9.873824502603459e-07,
"loss": 0.0,
"reward": 0.7480560662224889,
"reward_std": 1.0157663226127625,
"rewards/cosine_scaled_reward": 0.0927780270576477,
"rewards/format_reward": 0.5625000149011612,
"step": 84
},
{
"completion_length": 3185.8541870117188,
"epoch": 0.09714285714285714,
"grad_norm": 0.17822831869125366,
"kl": 0.0009975433349609375,
"learning_rate": 9.866330768241983e-07,
"loss": 0.0,
"reward": 0.07128806412220001,
"reward_std": 0.8152596428990364,
"rewards/cosine_scaled_reward": -0.08935598330572248,
"rewards/format_reward": 0.2500000111758709,
"step": 85
},
{
"completion_length": 2827.5625610351562,
"epoch": 0.09828571428571428,
"grad_norm": 0.1663668006658554,
"kl": 0.0010325908660888672,
"learning_rate": 9.85862422507884e-07,
"loss": 0.0,
"reward": 0.19359283335506916,
"reward_std": 0.644717700779438,
"rewards/cosine_scaled_reward": -0.10112026333808899,
"rewards/format_reward": 0.3958333544433117,
"step": 86
},
{
"completion_length": 2942.5,
"epoch": 0.09942857142857142,
"grad_norm": 0.1894853264093399,
"kl": 0.0014657974243164062,
"learning_rate": 9.850705248720068e-07,
"loss": 0.0001,
"reward": 0.1782783716917038,
"reward_std": 0.7724725604057312,
"rewards/cosine_scaled_reward": -0.08794412622228265,
"rewards/format_reward": 0.3541666716337204,
"step": 87
},
{
"completion_length": 2836.3959350585938,
"epoch": 0.10057142857142858,
"grad_norm": 0.1908150315284729,
"kl": 0.0014390945434570312,
"learning_rate": 9.8425742251254e-07,
"loss": 0.0001,
"reward": 0.3470733240246773,
"reward_std": 0.8534664362668991,
"rewards/cosine_scaled_reward": -0.024379996582865715,
"rewards/format_reward": 0.3958333432674408,
"step": 88
},
{
"completion_length": 3278.6041870117188,
"epoch": 0.10171428571428572,
"grad_norm": 0.1539601981639862,
"kl": 0.0004895925521850586,
"learning_rate": 9.83423155058946e-07,
"loss": 0.0,
"reward": 0.30410441011190414,
"reward_std": 0.6761599257588387,
"rewards/cosine_scaled_reward": 0.006218895316123962,
"rewards/format_reward": 0.29166667722165585,
"step": 89
},
{
"completion_length": 2985.3958740234375,
"epoch": 0.10285714285714286,
"grad_norm": 0.17722909152507782,
"kl": 0.00044274330139160156,
"learning_rate": 9.825677631722435e-07,
"loss": 0.0,
"reward": 0.2143753319978714,
"reward_std": 0.6936175674200058,
"rewards/cosine_scaled_reward": -0.0594789981842041,
"rewards/format_reward": 0.3333333358168602,
"step": 90
},
{
"completion_length": 3058.5000610351562,
"epoch": 0.104,
"grad_norm": 0.19192735850811005,
"kl": 0.0004374980926513672,
"learning_rate": 9.816912885430258e-07,
"loss": 0.0,
"reward": 0.302869388833642,
"reward_std": 0.5636695921421051,
"rewards/cosine_scaled_reward": -0.04648197069764137,
"rewards/format_reward": 0.3958333432674408,
"step": 91
},
{
"completion_length": 2699.6250610351562,
"epoch": 0.10514285714285715,
"grad_norm": 0.17412729561328888,
"kl": 0.0012140274047851562,
"learning_rate": 9.807937738894303e-07,
"loss": 0.0,
"reward": 0.5564358681440353,
"reward_std": 0.717531181871891,
"rewards/cosine_scaled_reward": 0.059467924758791924,
"rewards/format_reward": 0.4375000111758709,
"step": 92
},
{
"completion_length": 2346.1250610351562,
"epoch": 0.10628571428571429,
"grad_norm": 0.2216739058494568,
"kl": 0.0015277862548828125,
"learning_rate": 9.798752629550546e-07,
"loss": 0.0001,
"reward": 0.5152877140790224,
"reward_std": 0.6053595095872879,
"rewards/cosine_scaled_reward": -0.0027728192508220673,
"rewards/format_reward": 0.5208333358168602,
"step": 93
},
{
"completion_length": 3211.8333740234375,
"epoch": 0.10742857142857143,
"grad_norm": 0.18879717588424683,
"kl": 0.000827789306640625,
"learning_rate": 9.78935800506826e-07,
"loss": 0.0,
"reward": -0.015797210857272148,
"reward_std": 0.735307015478611,
"rewards/cosine_scaled_reward": -0.13289860635995865,
"rewards/format_reward": 0.2500000037252903,
"step": 94
},
{
"completion_length": 3198.1459350585938,
"epoch": 0.10857142857142857,
"grad_norm": 0.15773996710777283,
"kl": 0.00040841102600097656,
"learning_rate": 9.779754323328192e-07,
"loss": 0.0,
"reward": -0.016605263575911522,
"reward_std": 0.7409057542681694,
"rewards/cosine_scaled_reward": -0.16455264016985893,
"rewards/format_reward": 0.3125000037252903,
"step": 95
},
{
"completion_length": 2382.2084350585938,
"epoch": 0.10971428571428571,
"grad_norm": 0.2195434868335724,
"kl": 0.0015625953674316406,
"learning_rate": 9.769942052400235e-07,
"loss": 0.0001,
"reward": 0.3714839336462319,
"reward_std": 0.5286353975534439,
"rewards/cosine_scaled_reward": -0.08509137481451035,
"rewards/format_reward": 0.5416666865348816,
"step": 96
},
{
"completion_length": 2969.479248046875,
"epoch": 0.11085714285714286,
"grad_norm": 0.22521458566188812,
"kl": 0.0011968612670898438,
"learning_rate": 9.759921670520634e-07,
"loss": 0.0,
"reward": 0.023297425359487534,
"reward_std": 0.6409126222133636,
"rewards/cosine_scaled_reward": -0.14460130035877228,
"rewards/format_reward": 0.3125000111758709,
"step": 97
},
{
"completion_length": 2782.479248046875,
"epoch": 0.112,
"grad_norm": 0.48463404178619385,
"kl": 0.01540231704711914,
"learning_rate": 9.749693666068663e-07,
"loss": 0.0006,
"reward": 0.2789543569087982,
"reward_std": 0.6075774282217026,
"rewards/cosine_scaled_reward": -0.05843949131667614,
"rewards/format_reward": 0.39583333395421505,
"step": 98
},
{
"completion_length": 2909.1458435058594,
"epoch": 0.11314285714285714,
"grad_norm": 0.19729964435100555,
"kl": 0.000751495361328125,
"learning_rate": 9.739258537542835e-07,
"loss": 0.0,
"reward": 0.3117349073290825,
"reward_std": 0.5036360248923302,
"rewards/cosine_scaled_reward": -0.010799217969179153,
"rewards/format_reward": 0.33333333395421505,
"step": 99
},
{
"completion_length": 3006.9375610351562,
"epoch": 0.11428571428571428,
"grad_norm": 0.1744341254234314,
"kl": 0.0009255409240722656,
"learning_rate": 9.728616793536587e-07,
"loss": 0.0,
"reward": 0.4609271613880992,
"reward_std": 0.859523817896843,
"rewards/cosine_scaled_reward": 0.022130253724753857,
"rewards/format_reward": 0.416666679084301,
"step": 100
},
{
"completion_length": 2650.979217529297,
"epoch": 0.11542857142857142,
"grad_norm": 0.18743358552455902,
"kl": 0.0010256767272949219,
"learning_rate": 9.717768952713511e-07,
"loss": 0.0,
"reward": 0.21766437217593193,
"reward_std": 0.6801646202802658,
"rewards/cosine_scaled_reward": -0.09950115904211998,
"rewards/format_reward": 0.41666667722165585,
"step": 101
},
{
"completion_length": 2554.1875610351562,
"epoch": 0.11657142857142858,
"grad_norm": 0.18099477887153625,
"kl": 0.001209259033203125,
"learning_rate": 9.706715543782064e-07,
"loss": 0.0,
"reward": 0.29797927755862474,
"reward_std": 0.4223637208342552,
"rewards/cosine_scaled_reward": -0.10101036727428436,
"rewards/format_reward": 0.5,
"step": 102
},
{
"completion_length": 2658.979248046875,
"epoch": 0.11771428571428572,
"grad_norm": 0.15931963920593262,
"kl": 0.0010652542114257812,
"learning_rate": 9.695457105469804e-07,
"loss": 0.0,
"reward": 0.23173093050718307,
"reward_std": 0.6561538353562355,
"rewards/cosine_scaled_reward": -0.10288454219698906,
"rewards/format_reward": 0.4375,
"step": 103
},
{
"completion_length": 2505.5833740234375,
"epoch": 0.11885714285714286,
"grad_norm": 0.18099236488342285,
"kl": 0.0005369186401367188,
"learning_rate": 9.683994186497132e-07,
"loss": 0.0,
"reward": 0.9495851993560791,
"reward_std": 0.7366478592157364,
"rewards/cosine_scaled_reward": 0.19354257080703974,
"rewards/format_reward": 0.5625000149011612,
"step": 104
},
{
"completion_length": 2778.3125,
"epoch": 0.12,
"grad_norm": 0.20704622566699982,
"kl": 0.0013036727905273438,
"learning_rate": 9.672327345550543e-07,
"loss": 0.0001,
"reward": 0.24978191778063774,
"reward_std": 0.7690765783190727,
"rewards/cosine_scaled_reward": -0.09385904669761658,
"rewards/format_reward": 0.4375000074505806,
"step": 105
},
{
"completion_length": 2914.6250610351562,
"epoch": 0.12114285714285715,
"grad_norm": 0.1921072155237198,
"kl": 0.001560211181640625,
"learning_rate": 9.66045715125541e-07,
"loss": 0.0001,
"reward": 0.3622821723110974,
"reward_std": 0.912909746170044,
"rewards/cosine_scaled_reward": -0.01677557732909918,
"rewards/format_reward": 0.3958333395421505,
"step": 106
},
{
"completion_length": 2109.354248046875,
"epoch": 0.12228571428571429,
"grad_norm": 0.2092333436012268,
"kl": 0.0012311935424804688,
"learning_rate": 9.648384182148252e-07,
"loss": 0.0,
"reward": 0.5422526616603136,
"reward_std": 0.8620414137840271,
"rewards/cosine_scaled_reward": -0.062207008711993694,
"rewards/format_reward": 0.6666666865348816,
"step": 107
},
{
"completion_length": 2214.9583587646484,
"epoch": 0.12342857142857143,
"grad_norm": 0.2060472071170807,
"kl": 0.0030279159545898438,
"learning_rate": 9.636109026648554e-07,
"loss": 0.0001,
"reward": 0.6110497042536736,
"reward_std": 0.7519760131835938,
"rewards/cosine_scaled_reward": 0.03469152469187975,
"rewards/format_reward": 0.5416666716337204,
"step": 108
},
{
"completion_length": 2803.2083740234375,
"epoch": 0.12457142857142857,
"grad_norm": 0.20012950897216797,
"kl": 0.0010724067687988281,
"learning_rate": 9.623632283030077e-07,
"loss": 0.0,
"reward": 0.08136957883834839,
"reward_std": 0.6037605702877045,
"rewards/cosine_scaled_reward": -0.15723188465926796,
"rewards/format_reward": 0.3958333395421505,
"step": 109
},
{
"completion_length": 2464.7500610351562,
"epoch": 0.12571428571428572,
"grad_norm": 0.22272928059101105,
"kl": 0.0024871826171875,
"learning_rate": 9.610954559391704e-07,
"loss": 0.0001,
"reward": 0.9461969807744026,
"reward_std": 0.842521145939827,
"rewards/cosine_scaled_reward": 0.1710151496808976,
"rewards/format_reward": 0.604166679084301,
"step": 110
},
{
"completion_length": 1819.3750610351562,
"epoch": 0.12685714285714286,
"grad_norm": 0.24292264878749847,
"kl": 0.0017080307006835938,
"learning_rate": 9.598076473627796e-07,
"loss": 0.0001,
"reward": 0.7363898158073425,
"reward_std": 0.7160477414727211,
"rewards/cosine_scaled_reward": 0.003611571155488491,
"rewards/format_reward": 0.7291666716337204,
"step": 111
},
{
"completion_length": 2938.8333740234375,
"epoch": 0.128,
"grad_norm": 0.20110748708248138,
"kl": 0.0015575885772705078,
"learning_rate": 9.58499865339809e-07,
"loss": 0.0001,
"reward": 0.47850653529167175,
"reward_std": 0.8684659749269485,
"rewards/cosine_scaled_reward": 0.020503249019384384,
"rewards/format_reward": 0.4375000149011612,
"step": 112
},
{
"completion_length": 2170.791717529297,
"epoch": 0.12914285714285714,
"grad_norm": 0.22352100908756256,
"kl": 0.0018672943115234375,
"learning_rate": 9.571721736097088e-07,
"loss": 0.0001,
"reward": 0.8810203373432159,
"reward_std": 0.6223750859498978,
"rewards/cosine_scaled_reward": 0.09676016308367252,
"rewards/format_reward": 0.6875000074505806,
"step": 113
},
{
"completion_length": 1761.1250457763672,
"epoch": 0.13028571428571428,
"grad_norm": 0.22250227630138397,
"kl": 0.0011968612670898438,
"learning_rate": 9.55824636882301e-07,
"loss": 0.0,
"reward": 0.8257871624082327,
"reward_std": 0.5129944495856762,
"rewards/cosine_scaled_reward": 0.08997690677642822,
"rewards/format_reward": 0.6458333432674408,
"step": 114
},
{
"completion_length": 2406.6875610351562,
"epoch": 0.13142857142857142,
"grad_norm": 0.18063588440418243,
"kl": 0.0014677047729492188,
"learning_rate": 9.54457320834625e-07,
"loss": 0.0001,
"reward": 0.7925823777914047,
"reward_std": 1.05247762799263,
"rewards/cosine_scaled_reward": 0.08379119075834751,
"rewards/format_reward": 0.6250000149011612,
"step": 115
},
{
"completion_length": 2417.8958435058594,
"epoch": 0.13257142857142856,
"grad_norm": 0.23391500115394592,
"kl": 0.004772186279296875,
"learning_rate": 9.530702921077358e-07,
"loss": 0.0002,
"reward": 0.15583854354918003,
"reward_std": 0.6483651623129845,
"rewards/cosine_scaled_reward": -0.14083073096117005,
"rewards/format_reward": 0.4375,
"step": 116
},
{
"completion_length": 3263.666748046875,
"epoch": 0.1337142857142857,
"grad_norm": 0.1551298350095749,
"kl": 0.0015048980712890625,
"learning_rate": 9.516636183034564e-07,
"loss": 0.0001,
"reward": -0.017866918817162514,
"reward_std": 0.6443519741296768,
"rewards/cosine_scaled_reward": -0.10268345987424254,
"rewards/format_reward": 0.18750000558793545,
"step": 117
},
{
"completion_length": 2010.8750305175781,
"epoch": 0.13485714285714287,
"grad_norm": 0.21352525055408478,
"kl": 0.001308441162109375,
"learning_rate": 9.502373679810839e-07,
"loss": 0.0001,
"reward": 0.8165969103574753,
"reward_std": 0.8080126643180847,
"rewards/cosine_scaled_reward": 0.012465095147490501,
"rewards/format_reward": 0.7916666865348816,
"step": 118
},
{
"completion_length": 1884.0625610351562,
"epoch": 0.136,
"grad_norm": 0.27621325850486755,
"kl": 0.0025424957275390625,
"learning_rate": 9.487916106540465e-07,
"loss": 0.0001,
"reward": 0.5952838063240051,
"reward_std": 0.5625797361135483,
"rewards/cosine_scaled_reward": -0.046108097303658724,
"rewards/format_reward": 0.6875,
"step": 119
},
{
"completion_length": 2755.3959350585938,
"epoch": 0.13714285714285715,
"grad_norm": 0.23236262798309326,
"kl": 0.0017757415771484375,
"learning_rate": 9.473264167865171e-07,
"loss": 0.0001,
"reward": 0.063610197044909,
"reward_std": 0.8038829490542412,
"rewards/cosine_scaled_reward": -0.18694491172209382,
"rewards/format_reward": 0.4375000074505806,
"step": 120
},
{
"completion_length": 2782.5833740234375,
"epoch": 0.1382857142857143,
"grad_norm": 0.186203733086586,
"kl": 0.0014486312866210938,
"learning_rate": 9.458418577899774e-07,
"loss": 0.0001,
"reward": 0.2508644163608551,
"reward_std": 0.5808881223201752,
"rewards/cosine_scaled_reward": -0.09331781789660454,
"rewards/format_reward": 0.4375000111758709,
"step": 121
},
{
"completion_length": 2584.0625915527344,
"epoch": 0.13942857142857143,
"grad_norm": 0.2748485803604126,
"kl": 0.0027008056640625,
"learning_rate": 9.443380060197385e-07,
"loss": 0.0001,
"reward": 0.24610598012804985,
"reward_std": 0.4119979292154312,
"rewards/cosine_scaled_reward": -0.11653035134077072,
"rewards/format_reward": 0.47916667722165585,
"step": 122
},
{
"completion_length": 2821.416748046875,
"epoch": 0.14057142857142857,
"grad_norm": 0.27934524416923523,
"kl": 0.0057964324951171875,
"learning_rate": 9.428149347714143e-07,
"loss": 0.0002,
"reward": 0.10106497257947922,
"reward_std": 0.7212768346071243,
"rewards/cosine_scaled_reward": -0.10571751650422812,
"rewards/format_reward": 0.31250000186264515,
"step": 123
},
{
"completion_length": 2463.0000915527344,
"epoch": 0.1417142857142857,
"grad_norm": 0.22744229435920715,
"kl": 0.0014429092407226562,
"learning_rate": 9.412727182773486e-07,
"loss": 0.0001,
"reward": 0.8311970978975296,
"reward_std": 0.8409435376524925,
"rewards/cosine_scaled_reward": 0.11351519823074341,
"rewards/format_reward": 0.6041666716337204,
"step": 124
},
{
"completion_length": 3028.6459350585938,
"epoch": 0.14285714285714285,
"grad_norm": 0.230963334441185,
"kl": 0.0016689300537109375,
"learning_rate": 9.397114317029974e-07,
"loss": 0.0001,
"reward": 0.08886492438614368,
"reward_std": 0.5733988359570503,
"rewards/cosine_scaled_reward": -0.13265088573098183,
"rewards/format_reward": 0.35416666977107525,
"step": 125
},
{
"completion_length": 1943.6667175292969,
"epoch": 0.144,
"grad_norm": 0.26326608657836914,
"kl": 0.0029306411743164062,
"learning_rate": 9.381311511432658e-07,
"loss": 0.0001,
"reward": 0.6195674315094948,
"reward_std": 0.7148094028234482,
"rewards/cosine_scaled_reward": -0.03396627772599459,
"rewards/format_reward": 0.6875000149011612,
"step": 126
},
{
"completion_length": 2635.9166870117188,
"epoch": 0.14514285714285713,
"grad_norm": 0.2009022980928421,
"kl": 0.0012750625610351562,
"learning_rate": 9.36531953618799e-07,
"loss": 0.0001,
"reward": 0.41267674416303635,
"reward_std": 0.8958253264427185,
"rewards/cosine_scaled_reward": -0.022828295826911926,
"rewards/format_reward": 0.4583333395421505,
"step": 127
},
{
"completion_length": 2321.2500610351562,
"epoch": 0.1462857142857143,
"grad_norm": 0.19144511222839355,
"kl": 0.00170135498046875,
"learning_rate": 9.34913917072228e-07,
"loss": 0.0001,
"reward": 0.5039072521030903,
"reward_std": 0.8606824576854706,
"rewards/cosine_scaled_reward": -0.029296381399035454,
"rewards/format_reward": 0.5625000149011612,
"step": 128
},
{
"completion_length": 1837.7917175292969,
"epoch": 0.14742857142857144,
"grad_norm": 0.22352828085422516,
"kl": 0.0069904327392578125,
"learning_rate": 9.332771203643714e-07,
"loss": 0.0003,
"reward": 0.5292131304740906,
"reward_std": 0.6835447549819946,
"rewards/cosine_scaled_reward": -0.09997677942737937,
"rewards/format_reward": 0.7291666716337204,
"step": 129
},
{
"completion_length": 2737.7918090820312,
"epoch": 0.14857142857142858,
"grad_norm": 0.21162550151348114,
"kl": 0.00159454345703125,
"learning_rate": 9.316216432703916e-07,
"loss": 0.0001,
"reward": 0.1000329963862896,
"reward_std": 0.6349897980690002,
"rewards/cosine_scaled_reward": -0.15831685485318303,
"rewards/format_reward": 0.416666679084301,
"step": 130
},
{
"completion_length": 2639.3750610351562,
"epoch": 0.14971428571428572,
"grad_norm": 0.21849536895751953,
"kl": 0.001689910888671875,
"learning_rate": 9.299475664759068e-07,
"loss": 0.0001,
"reward": 0.3575108479708433,
"reward_std": 0.7335042506456375,
"rewards/cosine_scaled_reward": -0.07124457694590092,
"rewards/format_reward": 0.5000000055879354,
"step": 131
},
{
"completion_length": 2595.791748046875,
"epoch": 0.15085714285714286,
"grad_norm": 0.2819630801677704,
"kl": 0.0028476715087890625,
"learning_rate": 9.282549715730579e-07,
"loss": 0.0001,
"reward": -0.14514993596822023,
"reward_std": 0.4842909276485443,
"rewards/cosine_scaled_reward": -0.2704916410148144,
"rewards/format_reward": 0.3958333432674408,
"step": 132
},
{
"completion_length": 2752.2084350585938,
"epoch": 0.152,
"grad_norm": 0.20234017074108124,
"kl": 0.002986907958984375,
"learning_rate": 9.265439410565328e-07,
"loss": 0.0001,
"reward": 0.15741661936044693,
"reward_std": 0.6936222016811371,
"rewards/cosine_scaled_reward": -0.15045835822820663,
"rewards/format_reward": 0.4583333507180214,
"step": 133
},
{
"completion_length": 1920.0000610351562,
"epoch": 0.15314285714285714,
"grad_norm": 0.30035078525543213,
"kl": 0.00372314453125,
"learning_rate": 9.248145583195447e-07,
"loss": 0.0001,
"reward": 0.7273320555686951,
"reward_std": 0.7046244740486145,
"rewards/cosine_scaled_reward": 0.01991601102054119,
"rewards/format_reward": 0.6875,
"step": 134
},
{
"completion_length": 2167.8334350585938,
"epoch": 0.15428571428571428,
"grad_norm": 0.2100658118724823,
"kl": 0.0017566680908203125,
"learning_rate": 9.230669076497687e-07,
"loss": 0.0001,
"reward": 0.4464118145406246,
"reward_std": 0.3928603231906891,
"rewards/cosine_scaled_reward": -0.03721076436340809,
"rewards/format_reward": 0.520833333954215,
"step": 135
},
{
"completion_length": 2357.2084350585938,
"epoch": 0.15542857142857142,
"grad_norm": 0.24143747985363007,
"kl": 0.00250244140625,
"learning_rate": 9.213010742252327e-07,
"loss": 0.0001,
"reward": 0.6343938559293747,
"reward_std": 0.7614049315452576,
"rewards/cosine_scaled_reward": 0.004696924239397049,
"rewards/format_reward": 0.6250000149011612,
"step": 136
},
{
"completion_length": 2482.1875610351562,
"epoch": 0.15657142857142858,
"grad_norm": 0.22769631445407867,
"kl": 0.003353118896484375,
"learning_rate": 9.195171441101668e-07,
"loss": 0.0001,
"reward": 0.09337181597948074,
"reward_std": 0.6429625153541565,
"rewards/cosine_scaled_reward": -0.18248076736927032,
"rewards/format_reward": 0.4583333395421505,
"step": 137
},
{
"completion_length": 2043.5000610351562,
"epoch": 0.15771428571428572,
"grad_norm": 0.2516387403011322,
"kl": 0.003204345703125,
"learning_rate": 9.177152042508077e-07,
"loss": 0.0001,
"reward": 0.9434101283550262,
"reward_std": 0.8629068434238434,
"rewards/cosine_scaled_reward": 0.13837172836065292,
"rewards/format_reward": 0.6666666865348816,
"step": 138
},
{
"completion_length": 2047.5625,
"epoch": 0.15885714285714286,
"grad_norm": 0.2453654259443283,
"kl": 0.0018482208251953125,
"learning_rate": 9.158953424711624e-07,
"loss": 0.0001,
"reward": 0.7086100317537785,
"reward_std": 0.5908297449350357,
"rewards/cosine_scaled_reward": 0.00013833213597536087,
"rewards/format_reward": 0.7083333507180214,
"step": 139
},
{
"completion_length": 2375.7084350585938,
"epoch": 0.16,
"grad_norm": 0.1979781836271286,
"kl": 0.0024309158325195312,
"learning_rate": 9.140576474687263e-07,
"loss": 0.0001,
"reward": 0.619663898833096,
"reward_std": 0.41604873538017273,
"rewards/cosine_scaled_reward": 0.007748600095510483,
"rewards/format_reward": 0.6041666716337204,
"step": 140
},
{
"completion_length": 2381.7708740234375,
"epoch": 0.16114285714285714,
"grad_norm": 0.23859204351902008,
"kl": 0.003826141357421875,
"learning_rate": 9.122022088101613e-07,
"loss": 0.0002,
"reward": 0.3829444032162428,
"reward_std": 0.5856426432728767,
"rewards/cosine_scaled_reward": -0.05852780118584633,
"rewards/format_reward": 0.5000000149011612,
"step": 141
},
{
"completion_length": 2265.4584045410156,
"epoch": 0.16228571428571428,
"grad_norm": 0.26838216185569763,
"kl": 0.0041828155517578125,
"learning_rate": 9.103291169269299e-07,
"loss": 0.0002,
"reward": 0.459966566413641,
"reward_std": 0.5846913754940033,
"rewards/cosine_scaled_reward": -0.030433382838964462,
"rewards/format_reward": 0.5208333358168602,
"step": 142
},
{
"completion_length": 2539.479248046875,
"epoch": 0.16342857142857142,
"grad_norm": 0.18913578987121582,
"kl": 0.0029649734497070312,
"learning_rate": 9.084384631108882e-07,
"loss": 0.0001,
"reward": 0.42852520011365414,
"reward_std": 0.6579816788434982,
"rewards/cosine_scaled_reward": -0.05657072924077511,
"rewards/format_reward": 0.5416666865348816,
"step": 143
},
{
"completion_length": 2627.8125610351562,
"epoch": 0.16457142857142856,
"grad_norm": 0.1790352761745453,
"kl": 0.0033435821533203125,
"learning_rate": 9.065303395098358e-07,
"loss": 0.0001,
"reward": 0.7180662602186203,
"reward_std": 0.9851722121238708,
"rewards/cosine_scaled_reward": 0.05694979056715965,
"rewards/format_reward": 0.6041666939854622,
"step": 144
},
{
"completion_length": 2526.541778564453,
"epoch": 0.1657142857142857,
"grad_norm": 0.22108972072601318,
"kl": 0.0050811767578125,
"learning_rate": 9.046048391230247e-07,
"loss": 0.0002,
"reward": 0.40608268324285746,
"reward_std": 0.8329223841428757,
"rewards/cosine_scaled_reward": -0.0573753397911787,
"rewards/format_reward": 0.5208333432674408,
"step": 145
},
{
"completion_length": 2217.0625610351562,
"epoch": 0.16685714285714287,
"grad_norm": 0.37226402759552,
"kl": 0.00292205810546875,
"learning_rate": 9.026620557966279e-07,
"loss": 0.0001,
"reward": 0.29790709912776947,
"reward_std": 0.6913768872618675,
"rewards/cosine_scaled_reward": -0.19479646161198616,
"rewards/format_reward": 0.6875000149011612,
"step": 146
},
{
"completion_length": 2100.500030517578,
"epoch": 0.168,
"grad_norm": 0.1791330724954605,
"kl": 0.0022411346435546875,
"learning_rate": 9.007020842191634e-07,
"loss": 0.0001,
"reward": 0.6166809126734734,
"reward_std": 0.7499666661024094,
"rewards/cosine_scaled_reward": -0.014576207846403122,
"rewards/format_reward": 0.6458333432674408,
"step": 147
},
{
"completion_length": 1953.1667022705078,
"epoch": 0.16914285714285715,
"grad_norm": 0.26837047934532166,
"kl": 0.0032806396484375,
"learning_rate": 8.987250199168808e-07,
"loss": 0.0001,
"reward": 0.22773092985153198,
"reward_std": 0.5684618726372719,
"rewards/cosine_scaled_reward": -0.21946788486093283,
"rewards/format_reward": 0.6666666679084301,
"step": 148
},
{
"completion_length": 2537.8958740234375,
"epoch": 0.1702857142857143,
"grad_norm": 0.19118516147136688,
"kl": 0.0032138824462890625,
"learning_rate": 8.967309592491052e-07,
"loss": 0.0001,
"reward": 0.9459018707275391,
"reward_std": 0.6409400217235088,
"rewards/cosine_scaled_reward": 0.18128425255417824,
"rewards/format_reward": 0.5833333488553762,
"step": 149
},
{
"completion_length": 2327.9375610351562,
"epoch": 0.17142857142857143,
"grad_norm": 0.22041891515254974,
"kl": 0.004886627197265625,
"learning_rate": 8.9471999940354e-07,
"loss": 0.0002,
"reward": 0.2033998966217041,
"reward_std": 0.6696746721863747,
"rewards/cosine_scaled_reward": -0.1587167321704328,
"rewards/format_reward": 0.5208333488553762,
"step": 150
},
{
"completion_length": 2248.9375610351562,
"epoch": 0.17257142857142857,
"grad_norm": 0.20176808536052704,
"kl": 0.0022144317626953125,
"learning_rate": 8.926922383915315e-07,
"loss": 0.0001,
"reward": 0.2744840234518051,
"reward_std": 0.46313488483428955,
"rewards/cosine_scaled_reward": -0.11275799572467804,
"rewards/format_reward": 0.5000000149011612,
"step": 151
},
{
"completion_length": 2933.0833740234375,
"epoch": 0.1737142857142857,
"grad_norm": 0.26598456501960754,
"kl": 0.005481719970703125,
"learning_rate": 8.906477750432903e-07,
"loss": 0.0002,
"reward": 0.04549443535506725,
"reward_std": 0.48727741837501526,
"rewards/cosine_scaled_reward": -0.13350277952849865,
"rewards/format_reward": 0.3125000074505806,
"step": 152
},
{
"completion_length": 2075.729278564453,
"epoch": 0.17485714285714285,
"grad_norm": 0.19798634946346283,
"kl": 0.0023260116577148438,
"learning_rate": 8.88586709003076e-07,
"loss": 0.0001,
"reward": 0.9074295610189438,
"reward_std": 0.5949664637446404,
"rewards/cosine_scaled_reward": 0.09954808466136456,
"rewards/format_reward": 0.7083333432674408,
"step": 153
},
{
"completion_length": 2754.791748046875,
"epoch": 0.176,
"grad_norm": 0.19806884229183197,
"kl": 0.0027751922607421875,
"learning_rate": 8.865091407243394e-07,
"loss": 0.0001,
"reward": 0.6114984676241875,
"reward_std": 0.6458263993263245,
"rewards/cosine_scaled_reward": 0.08699923381209373,
"rewards/format_reward": 0.4375000149011612,
"step": 154
},
{
"completion_length": 3066.354248046875,
"epoch": 0.17714285714285713,
"grad_norm": 0.29352447390556335,
"kl": 0.00366973876953125,
"learning_rate": 8.844151714648274e-07,
"loss": 0.0001,
"reward": 0.017834719270467758,
"reward_std": 0.8173775672912598,
"rewards/cosine_scaled_reward": -0.14733264222741127,
"rewards/format_reward": 0.31250000558793545,
"step": 155
},
{
"completion_length": 2754.4375610351562,
"epoch": 0.1782857142857143,
"grad_norm": 0.19924919307231903,
"kl": 0.005084991455078125,
"learning_rate": 8.823049032816478e-07,
"loss": 0.0002,
"reward": 0.06147514842450619,
"reward_std": 0.5927468463778496,
"rewards/cosine_scaled_reward": -0.14634575322270393,
"rewards/format_reward": 0.3541666828095913,
"step": 156
},
{
"completion_length": 2206.250030517578,
"epoch": 0.17942857142857144,
"grad_norm": 0.23515692353248596,
"kl": 0.0025005340576171875,
"learning_rate": 8.801784390262943e-07,
"loss": 0.0001,
"reward": 0.9370372518897057,
"reward_std": 0.8283505141735077,
"rewards/cosine_scaled_reward": 0.1456019375473261,
"rewards/format_reward": 0.6458333507180214,
"step": 157
},
{
"completion_length": 2862.7708740234375,
"epoch": 0.18057142857142858,
"grad_norm": 0.19533918797969818,
"kl": 0.004573822021484375,
"learning_rate": 8.780358823396352e-07,
"loss": 0.0002,
"reward": -0.053055196069180965,
"reward_std": 0.6198497340083122,
"rewards/cosine_scaled_reward": -0.23486093431711197,
"rewards/format_reward": 0.41666667349636555,
"step": 158
},
{
"completion_length": 1894.8542175292969,
"epoch": 0.18171428571428572,
"grad_norm": 0.22211147844791412,
"kl": 0.002826690673828125,
"learning_rate": 8.758773376468604e-07,
"loss": 0.0001,
"reward": 1.2009564340114594,
"reward_std": 0.7813936918973923,
"rewards/cosine_scaled_reward": 0.1942282197996974,
"rewards/format_reward": 0.8125000149011612,
"step": 159
},
{
"completion_length": 1964.9792175292969,
"epoch": 0.18285714285714286,
"grad_norm": 0.21944580972194672,
"kl": 0.003116607666015625,
"learning_rate": 8.737029101523929e-07,
"loss": 0.0001,
"reward": 0.9964812844991684,
"reward_std": 0.7849611788988113,
"rewards/cosine_scaled_reward": 0.10240732878446579,
"rewards/format_reward": 0.7916666716337204,
"step": 160
},
{
"completion_length": 2853.4583740234375,
"epoch": 0.184,
"grad_norm": 0.1943131983280182,
"kl": 0.00357818603515625,
"learning_rate": 8.715127058347614e-07,
"loss": 0.0001,
"reward": 0.10683573782444,
"reward_std": 0.6206659823656082,
"rewards/cosine_scaled_reward": -0.1549154706299305,
"rewards/format_reward": 0.41666667722165585,
"step": 161
},
{
"completion_length": 2725.604248046875,
"epoch": 0.18514285714285714,
"grad_norm": 0.18736310303211212,
"kl": 0.00328826904296875,
"learning_rate": 8.693068314414344e-07,
"loss": 0.0001,
"reward": 0.16440774500370026,
"reward_std": 0.7531605362892151,
"rewards/cosine_scaled_reward": -0.1782128056511283,
"rewards/format_reward": 0.5208333507180214,
"step": 162
},
{
"completion_length": 2378.604248046875,
"epoch": 0.18628571428571428,
"grad_norm": 0.27985262870788574,
"kl": 0.00499725341796875,
"learning_rate": 8.670853944836176e-07,
"loss": 0.0002,
"reward": 0.3438632491452154,
"reward_std": 0.8545256406068802,
"rewards/cosine_scaled_reward": -0.06765171512961388,
"rewards/format_reward": 0.479166679084301,
"step": 163
},
{
"completion_length": 2041.791748046875,
"epoch": 0.18742857142857142,
"grad_norm": 0.2726307511329651,
"kl": 0.00522613525390625,
"learning_rate": 8.648485032310144e-07,
"loss": 0.0002,
"reward": 0.1412298008799553,
"reward_std": 0.45817675441503525,
"rewards/cosine_scaled_reward": -0.18980177072808146,
"rewards/format_reward": 0.520833333954215,
"step": 164
},
{
"completion_length": 2056.8958740234375,
"epoch": 0.18857142857142858,
"grad_norm": 0.24950121343135834,
"kl": 0.0033512115478515625,
"learning_rate": 8.625962667065487e-07,
"loss": 0.0001,
"reward": 0.4896044433116913,
"reward_std": 0.6808345168828964,
"rewards/cosine_scaled_reward": -0.057281110901385546,
"rewards/format_reward": 0.6041666828095913,
"step": 165
},
{
"completion_length": 1701.5625,
"epoch": 0.18971428571428572,
"grad_norm": 0.20629195868968964,
"kl": 0.0029201507568359375,
"learning_rate": 8.603287946810513e-07,
"loss": 0.0001,
"reward": 0.4639076357707381,
"reward_std": 0.4746507927775383,
"rewards/cosine_scaled_reward": -0.13262954354286194,
"rewards/format_reward": 0.7291666716337204,
"step": 166
},
{
"completion_length": 2674.4375610351562,
"epoch": 0.19085714285714286,
"grad_norm": 0.19312238693237305,
"kl": 0.004119873046875,
"learning_rate": 8.580461976679099e-07,
"loss": 0.0002,
"reward": 0.26991652697324753,
"reward_std": 0.8362310528755188,
"rewards/cosine_scaled_reward": -0.11504174256697297,
"rewards/format_reward": 0.5000000111758709,
"step": 167
},
{
"completion_length": 1723.5209045410156,
"epoch": 0.192,
"grad_norm": 0.19439440965652466,
"kl": 0.002704620361328125,
"learning_rate": 8.557485869176825e-07,
"loss": 0.0001,
"reward": 0.7088751941919327,
"reward_std": 0.7652025148272514,
"rewards/cosine_scaled_reward": -0.051812431775033474,
"rewards/format_reward": 0.8125000149011612,
"step": 168
},
{
"completion_length": 2231.979248046875,
"epoch": 0.19314285714285714,
"grad_norm": 0.2904442250728607,
"kl": 0.004784584045410156,
"learning_rate": 8.534360744126753e-07,
"loss": 0.0002,
"reward": 0.9261031150817871,
"reward_std": 0.9859992563724518,
"rewards/cosine_scaled_reward": 0.12971824035048485,
"rewards/format_reward": 0.6666666939854622,
"step": 169
},
{
"completion_length": 2114.0208740234375,
"epoch": 0.19428571428571428,
"grad_norm": 0.19766280055046082,
"kl": 0.003971099853515625,
"learning_rate": 8.511087728614862e-07,
"loss": 0.0002,
"reward": 0.3019937239587307,
"reward_std": 0.6615323200821877,
"rewards/cosine_scaled_reward": -0.1510864682495594,
"rewards/format_reward": 0.6041666865348816,
"step": 170
},
{
"completion_length": 1614.7708435058594,
"epoch": 0.19542857142857142,
"grad_norm": 0.23038722574710846,
"kl": 0.0032196044921875,
"learning_rate": 8.487667956935087e-07,
"loss": 0.0001,
"reward": 0.5918858665972948,
"reward_std": 0.47077811881899834,
"rewards/cosine_scaled_reward": -0.09989039599895477,
"rewards/format_reward": 0.7916666716337204,
"step": 171
},
{
"completion_length": 1724.5833740234375,
"epoch": 0.19657142857142856,
"grad_norm": 0.2515551447868347,
"kl": 0.00432586669921875,
"learning_rate": 8.464102570534061e-07,
"loss": 0.0002,
"reward": 0.8076295026112348,
"reward_std": 0.5722271054983139,
"rewards/cosine_scaled_reward": 0.04964808002114296,
"rewards/format_reward": 0.7083333432674408,
"step": 172
},
{
"completion_length": 2433.4375610351562,
"epoch": 0.1977142857142857,
"grad_norm": 0.3010346591472626,
"kl": 0.004909515380859375,
"learning_rate": 8.440392717955475e-07,
"loss": 0.0002,
"reward": 0.34551432851003483,
"reward_std": 0.7096427381038666,
"rewards/cosine_scaled_reward": -0.08765951948589645,
"rewards/format_reward": 0.5208333488553762,
"step": 173
},
{
"completion_length": 2419.979217529297,
"epoch": 0.19885714285714284,
"grad_norm": 0.19969363510608673,
"kl": 0.005977630615234375,
"learning_rate": 8.416539554784089e-07,
"loss": 0.0002,
"reward": 0.27888998575508595,
"reward_std": 0.5231342613697052,
"rewards/cosine_scaled_reward": -0.1001383513212204,
"rewards/format_reward": 0.4791666716337204,
"step": 174
},
{
"completion_length": 2777.1250610351562,
"epoch": 0.2,
"grad_norm": 0.17995108664035797,
"kl": 0.0071868896484375,
"learning_rate": 8.392544243589427e-07,
"loss": 0.0003,
"reward": 0.6402685008943081,
"reward_std": 0.7186409756541252,
"rewards/cosine_scaled_reward": 0.0909675620496273,
"rewards/format_reward": 0.4583333395421505,
"step": 175
},
{
"completion_length": 2093.25,
"epoch": 0.20114285714285715,
"grad_norm": 0.20400448143482208,
"kl": 0.00519561767578125,
"learning_rate": 8.368407953869103e-07,
"loss": 0.0002,
"reward": 0.5985848978161812,
"reward_std": 0.7769260108470917,
"rewards/cosine_scaled_reward": -0.03404088690876961,
"rewards/format_reward": 0.6666666716337204,
"step": 176
},
{
"completion_length": 2018.1875915527344,
"epoch": 0.2022857142857143,
"grad_norm": 0.21771669387817383,
"kl": 0.004276275634765625,
"learning_rate": 8.344131861991828e-07,
"loss": 0.0002,
"reward": 0.5329161509871483,
"reward_std": 0.5947398841381073,
"rewards/cosine_scaled_reward": -0.0772919338196516,
"rewards/format_reward": 0.6875000149011612,
"step": 177
},
{
"completion_length": 1553.2083892822266,
"epoch": 0.20342857142857143,
"grad_norm": 0.2806382477283478,
"kl": 0.00424957275390625,
"learning_rate": 8.319717151140072e-07,
"loss": 0.0002,
"reward": 0.9667995385825634,
"reward_std": 0.4322159215807915,
"rewards/cosine_scaled_reward": 0.09798309206962585,
"rewards/format_reward": 0.770833358168602,
"step": 178
},
{
"completion_length": 2260.9584350585938,
"epoch": 0.20457142857142857,
"grad_norm": 0.19229480624198914,
"kl": 0.00655364990234375,
"learning_rate": 8.295165011252396e-07,
"loss": 0.0003,
"reward": 0.6450915709137917,
"reward_std": 0.6193302199244499,
"rewards/cosine_scaled_reward": 0.030879119411110878,
"rewards/format_reward": 0.583333358168602,
"step": 179
},
{
"completion_length": 2204.104248046875,
"epoch": 0.2057142857142857,
"grad_norm": 0.34143543243408203,
"kl": 0.00551605224609375,
"learning_rate": 8.270476638965461e-07,
"loss": 0.0002,
"reward": 0.49652543663978577,
"reward_std": 0.9920015186071396,
"rewards/cosine_scaled_reward": -0.06423728261142969,
"rewards/format_reward": 0.6250000111758709,
"step": 180
},
{
"completion_length": 2376.3333740234375,
"epoch": 0.20685714285714285,
"grad_norm": 0.1844940185546875,
"kl": 0.002460479736328125,
"learning_rate": 8.245653237555705e-07,
"loss": 0.0001,
"reward": 0.7340436186641455,
"reward_std": 0.7672436386346817,
"rewards/cosine_scaled_reward": 0.033688463270664215,
"rewards/format_reward": 0.6666666865348816,
"step": 181
},
{
"completion_length": 1781.0208435058594,
"epoch": 0.208,
"grad_norm": 0.27145451307296753,
"kl": 0.005458831787109375,
"learning_rate": 8.220696016880687e-07,
"loss": 0.0002,
"reward": 0.6259329319000244,
"reward_std": 0.7968147397041321,
"rewards/cosine_scaled_reward": -0.062033540569245815,
"rewards/format_reward": 0.7500000111758709,
"step": 182
},
{
"completion_length": 2343.8959045410156,
"epoch": 0.20914285714285713,
"grad_norm": 0.2297639399766922,
"kl": 0.00566864013671875,
"learning_rate": 8.195606193320136e-07,
"loss": 0.0002,
"reward": 0.11410272493958473,
"reward_std": 0.5572097525000572,
"rewards/cosine_scaled_reward": -0.2658653110265732,
"rewards/format_reward": 0.6458333432674408,
"step": 183
},
{
"completion_length": 1841.2292175292969,
"epoch": 0.2102857142857143,
"grad_norm": 0.2628481388092041,
"kl": 0.003875732421875,
"learning_rate": 8.170384989716657e-07,
"loss": 0.0002,
"reward": 1.0498279109597206,
"reward_std": 0.812163732945919,
"rewards/cosine_scaled_reward": 0.1394973020069301,
"rewards/format_reward": 0.770833358168602,
"step": 184
},
{
"completion_length": 2690.666748046875,
"epoch": 0.21142857142857144,
"grad_norm": 0.2707008719444275,
"kl": 0.006504058837890625,
"learning_rate": 8.145033635316128e-07,
"loss": 0.0003,
"reward": 0.12935106456279755,
"reward_std": 0.6062737256288528,
"rewards/cosine_scaled_reward": -0.15407447703182697,
"rewards/format_reward": 0.4375000149011612,
"step": 185
},
{
"completion_length": 2648.7083740234375,
"epoch": 0.21257142857142858,
"grad_norm": 0.277004599571228,
"kl": 0.00493621826171875,
"learning_rate": 8.119553365707802e-07,
"loss": 0.0002,
"reward": 0.3933283071964979,
"reward_std": 0.6029615625739098,
"rewards/cosine_scaled_reward": -0.07416917383670807,
"rewards/format_reward": 0.5416666865348816,
"step": 186
},
{
"completion_length": 2165.416778564453,
"epoch": 0.21371428571428572,
"grad_norm": 0.2298295795917511,
"kl": 0.005840301513671875,
"learning_rate": 8.093945422764069e-07,
"loss": 0.0002,
"reward": 0.7806095313280821,
"reward_std": 0.7954358160495758,
"rewards/cosine_scaled_reward": 0.025721419602632523,
"rewards/format_reward": 0.7291666865348816,
"step": 187
},
{
"completion_length": 1475.4167022705078,
"epoch": 0.21485714285714286,
"grad_norm": 0.24691948294639587,
"kl": 0.005809783935546875,
"learning_rate": 8.068211054579943e-07,
"loss": 0.0002,
"reward": 1.0945345759391785,
"reward_std": 0.8786479085683823,
"rewards/cosine_scaled_reward": 0.10976729169487953,
"rewards/format_reward": 0.8750000149011612,
"step": 188
},
{
"completion_length": 1539.8333740234375,
"epoch": 0.216,
"grad_norm": 0.27775290608406067,
"kl": 0.0064697265625,
"learning_rate": 8.04235151541222e-07,
"loss": 0.0003,
"reward": 0.6156105473637581,
"reward_std": 0.7454669773578644,
"rewards/cosine_scaled_reward": -0.119278060272336,
"rewards/format_reward": 0.8541666865348816,
"step": 189
},
{
"completion_length": 2440.979278564453,
"epoch": 0.21714285714285714,
"grad_norm": 0.22604604065418243,
"kl": 0.00592041015625,
"learning_rate": 8.01636806561836e-07,
"loss": 0.0002,
"reward": 0.4239200847223401,
"reward_std": 0.8845669329166412,
"rewards/cosine_scaled_reward": -0.038039978593587875,
"rewards/format_reward": 0.500000013038516,
"step": 190
},
{
"completion_length": 1904.729248046875,
"epoch": 0.21828571428571428,
"grad_norm": 0.2662159204483032,
"kl": 0.00586700439453125,
"learning_rate": 7.990261971595048e-07,
"loss": 0.0002,
"reward": 0.4985937252640724,
"reward_std": 0.7315638810396194,
"rewards/cosine_scaled_reward": -0.09445315971970558,
"rewards/format_reward": 0.6875000074505806,
"step": 191
},
{
"completion_length": 2417.1875610351562,
"epoch": 0.21942857142857142,
"grad_norm": 0.27427810430526733,
"kl": 0.009246826171875,
"learning_rate": 7.964034505716476e-07,
"loss": 0.0004,
"reward": 0.3124541025608778,
"reward_std": 0.6425688564777374,
"rewards/cosine_scaled_reward": -0.12502295151352882,
"rewards/format_reward": 0.5625000074505806,
"step": 192
},
{
"completion_length": 1865.0208740234375,
"epoch": 0.22057142857142858,
"grad_norm": 0.2562018930912018,
"kl": 0.01080322265625,
"learning_rate": 7.93768694627233e-07,
"loss": 0.0004,
"reward": 0.5529625415802002,
"reward_std": 0.5897716134786606,
"rewards/cosine_scaled_reward": -0.0985187292098999,
"rewards/format_reward": 0.7500000149011612,
"step": 193
},
{
"completion_length": 1062.4375305175781,
"epoch": 0.22171428571428572,
"grad_norm": 0.22855594754219055,
"kl": 0.0037994384765625,
"learning_rate": 7.911220577405484e-07,
"loss": 0.0002,
"reward": 1.7262530326843262,
"reward_std": 0.826399639248848,
"rewards/cosine_scaled_reward": 0.3631264716386795,
"rewards/format_reward": 1.0,
"step": 194
},
{
"completion_length": 1257.5625305175781,
"epoch": 0.22285714285714286,
"grad_norm": 0.26371893286705017,
"kl": 0.00643157958984375,
"learning_rate": 7.884636689049422e-07,
"loss": 0.0003,
"reward": 1.151860922574997,
"reward_std": 0.6702793166041374,
"rewards/cosine_scaled_reward": 0.12801377475261688,
"rewards/format_reward": 0.8958333432674408,
"step": 195
},
{
"completion_length": 2729.5000610351562,
"epoch": 0.224,
"grad_norm": 0.2403058111667633,
"kl": 0.00676727294921875,
"learning_rate": 7.857936576865356e-07,
"loss": 0.0003,
"reward": 0.26913030445575714,
"reward_std": 0.6797884181141853,
"rewards/cosine_scaled_reward": -0.08418486639857292,
"rewards/format_reward": 0.4375,
"step": 196
},
{
"completion_length": 2123.791717529297,
"epoch": 0.22514285714285714,
"grad_norm": 0.22864989936351776,
"kl": 0.006134033203125,
"learning_rate": 7.831121542179086e-07,
"loss": 0.0002,
"reward": 0.06347193196415901,
"reward_std": 0.40899810940027237,
"rewards/cosine_scaled_reward": -0.24951404333114624,
"rewards/format_reward": 0.5625000149011612,
"step": 197
},
{
"completion_length": 2316.791748046875,
"epoch": 0.22628571428571428,
"grad_norm": 0.2166266292333603,
"kl": 0.00603485107421875,
"learning_rate": 7.804192891917571e-07,
"loss": 0.0002,
"reward": 0.4236091636121273,
"reward_std": 0.794644683599472,
"rewards/cosine_scaled_reward": -0.1215287372469902,
"rewards/format_reward": 0.6666666865348816,
"step": 198
},
{
"completion_length": 2175.5416870117188,
"epoch": 0.22742857142857142,
"grad_norm": 0.2332044243812561,
"kl": 0.005603790283203125,
"learning_rate": 7.777151938545235e-07,
"loss": 0.0002,
"reward": 1.2629163265228271,
"reward_std": 0.7567542046308517,
"rewards/cosine_scaled_reward": 0.2252081297338009,
"rewards/format_reward": 0.8125000298023224,
"step": 199
},
{
"completion_length": 1545.8125305175781,
"epoch": 0.22857142857142856,
"grad_norm": 0.3451651632785797,
"kl": 0.006610870361328125,
"learning_rate": 7.75e-07,
"loss": 0.0003,
"reward": 1.1884014122188091,
"reward_std": 0.868816927075386,
"rewards/cosine_scaled_reward": 0.1983673730865121,
"rewards/format_reward": 0.7916666716337204,
"step": 200
},
{
"completion_length": 1524.0625610351562,
"epoch": 0.2297142857142857,
"grad_norm": 0.21861064434051514,
"kl": 0.0051116943359375,
"learning_rate": 7.72273839962904e-07,
"loss": 0.0002,
"reward": 0.675473814830184,
"reward_std": 0.6859661787748337,
"rewards/cosine_scaled_reward": -0.07892975211143494,
"rewards/format_reward": 0.8333333432674408,
"step": 201
},
{
"completion_length": 1321.3959045410156,
"epoch": 0.23085714285714284,
"grad_norm": 0.24629908800125122,
"kl": 0.007568359375,
"learning_rate": 7.695368466124296e-07,
"loss": 0.0003,
"reward": 0.9415311962366104,
"reward_std": 0.7775374501943588,
"rewards/cosine_scaled_reward": 0.02284892648458481,
"rewards/format_reward": 0.8958333432674408,
"step": 202
},
{
"completion_length": 1379.7709045410156,
"epoch": 0.232,
"grad_norm": 0.27627113461494446,
"kl": 0.00753021240234375,
"learning_rate": 7.667891533457718e-07,
"loss": 0.0003,
"reward": 1.1194992661476135,
"reward_std": 0.7730197608470917,
"rewards/cosine_scaled_reward": 0.09099959582090378,
"rewards/format_reward": 0.9375000149011612,
"step": 203
},
{
"completion_length": 1120.1667022705078,
"epoch": 0.23314285714285715,
"grad_norm": 0.26729604601860046,
"kl": 0.00640869140625,
"learning_rate": 7.640308940816239e-07,
"loss": 0.0003,
"reward": 1.2363095879554749,
"reward_std": 0.8477204591035843,
"rewards/cosine_scaled_reward": 0.15982142463326454,
"rewards/format_reward": 0.9166666716337204,
"step": 204
},
{
"completion_length": 1971.2291870117188,
"epoch": 0.2342857142857143,
"grad_norm": 0.2195984125137329,
"kl": 0.00698089599609375,
"learning_rate": 7.612622032536507e-07,
"loss": 0.0003,
"reward": 0.8974205702543259,
"reward_std": 0.8895229697227478,
"rewards/cosine_scaled_reward": 0.052876945585012436,
"rewards/format_reward": 0.7916666716337204,
"step": 205
},
{
"completion_length": 2061.3334350585938,
"epoch": 0.23542857142857143,
"grad_norm": 0.268889844417572,
"kl": 0.007232666015625,
"learning_rate": 7.584832158039378e-07,
"loss": 0.0003,
"reward": 0.2433365173637867,
"reward_std": 0.5611164793372154,
"rewards/cosine_scaled_reward": -0.2324984148144722,
"rewards/format_reward": 0.708333358168602,
"step": 206
},
{
"completion_length": 1904.5000305175781,
"epoch": 0.23657142857142857,
"grad_norm": 0.19144296646118164,
"kl": 0.00501251220703125,
"learning_rate": 7.556940671764124e-07,
"loss": 0.0002,
"reward": 0.9468748420476913,
"reward_std": 0.6613385528326035,
"rewards/cosine_scaled_reward": 0.05677075684070587,
"rewards/format_reward": 0.8333333432674408,
"step": 207
},
{
"completion_length": 1996.8125610351562,
"epoch": 0.2377142857142857,
"grad_norm": 0.29941245913505554,
"kl": 0.00930023193359375,
"learning_rate": 7.528948933102438e-07,
"loss": 0.0004,
"reward": 0.29106441140174866,
"reward_std": 0.6116437911987305,
"rewards/cosine_scaled_reward": -0.17738447710871696,
"rewards/format_reward": 0.6458333544433117,
"step": 208
},
{
"completion_length": 1285.9166870117188,
"epoch": 0.23885714285714285,
"grad_norm": 0.3566973805427551,
"kl": 0.00695037841796875,
"learning_rate": 7.500858306332172e-07,
"loss": 0.0003,
"reward": 0.5402148813009262,
"reward_std": 0.7111145555973053,
"rewards/cosine_scaled_reward": -0.14655922167003155,
"rewards/format_reward": 0.8333333432674408,
"step": 209
},
{
"completion_length": 1275.5000305175781,
"epoch": 0.24,
"grad_norm": 0.2605917155742645,
"kl": 0.006221771240234375,
"learning_rate": 7.472670160550848e-07,
"loss": 0.0002,
"reward": 1.3301078528165817,
"reward_std": 0.8438884019851685,
"rewards/cosine_scaled_reward": 0.18588725943118334,
"rewards/format_reward": 0.9583333432674408,
"step": 210
},
{
"completion_length": 2328.6251220703125,
"epoch": 0.24114285714285713,
"grad_norm": 0.1887713074684143,
"kl": 0.0092010498046875,
"learning_rate": 7.444385869608921e-07,
"loss": 0.0004,
"reward": 0.6505604535341263,
"reward_std": 0.5875601321458817,
"rewards/cosine_scaled_reward": -0.008053132332861423,
"rewards/format_reward": 0.6666666865348816,
"step": 211
},
{
"completion_length": 1282.166732788086,
"epoch": 0.2422857142857143,
"grad_norm": 0.2532815635204315,
"kl": 0.0076446533203125,
"learning_rate": 7.416006812042827e-07,
"loss": 0.0003,
"reward": 1.1198171079158783,
"reward_std": 0.8018105626106262,
"rewards/cosine_scaled_reward": 0.11199186649173498,
"rewards/format_reward": 0.8958333432674408,
"step": 212
},
{
"completion_length": 1855.0834045410156,
"epoch": 0.24342857142857144,
"grad_norm": 0.2252466082572937,
"kl": 0.0081024169921875,
"learning_rate": 7.387534371007797e-07,
"loss": 0.0003,
"reward": 0.2633536756038666,
"reward_std": 0.42707760632038116,
"rewards/cosine_scaled_reward": -0.2224898338317871,
"rewards/format_reward": 0.7083333432674408,
"step": 213
},
{
"completion_length": 1971.8959045410156,
"epoch": 0.24457142857142858,
"grad_norm": 0.1935078501701355,
"kl": 0.004627227783203125,
"learning_rate": 7.358969934210438e-07,
"loss": 0.0002,
"reward": 0.7197396508418024,
"reward_std": 0.6319501101970673,
"rewards/cosine_scaled_reward": -0.056796859949827194,
"rewards/format_reward": 0.8333333432674408,
"step": 214
},
{
"completion_length": 1413.3541870117188,
"epoch": 0.24571428571428572,
"grad_norm": 0.2069859504699707,
"kl": 0.00742340087890625,
"learning_rate": 7.330314893841101e-07,
"loss": 0.0003,
"reward": 1.0066201090812683,
"reward_std": 0.9675450921058655,
"rewards/cosine_scaled_reward": 0.03456003498286009,
"rewards/format_reward": 0.9375000149011612,
"step": 215
},
{
"completion_length": 2199.2083435058594,
"epoch": 0.24685714285714286,
"grad_norm": 0.2034756988286972,
"kl": 0.00646209716796875,
"learning_rate": 7.301570646506027e-07,
"loss": 0.0003,
"reward": 0.47810695320367813,
"reward_std": 0.7864377945661545,
"rewards/cosine_scaled_reward": -0.12552986666560173,
"rewards/format_reward": 0.7291666865348816,
"step": 216
},
{
"completion_length": 1527.8958740234375,
"epoch": 0.248,
"grad_norm": 0.2671460807323456,
"kl": 0.006622314453125,
"learning_rate": 7.27273859315928e-07,
"loss": 0.0003,
"reward": 0.7772237807512283,
"reward_std": 0.6489354968070984,
"rewards/cosine_scaled_reward": -0.038471437990665436,
"rewards/format_reward": 0.8541666865348816,
"step": 217
},
{
"completion_length": 1419.3542175292969,
"epoch": 0.24914285714285714,
"grad_norm": 0.2513315677642822,
"kl": 0.00665283203125,
"learning_rate": 7.243820139034464e-07,
"loss": 0.0003,
"reward": 1.1047292775474489,
"reward_std": 0.6393595859408379,
"rewards/cosine_scaled_reward": 0.1252813059836626,
"rewards/format_reward": 0.8541666716337204,
"step": 218
},
{
"completion_length": 1275.3125305175781,
"epoch": 0.2502857142857143,
"grad_norm": 0.2648639380931854,
"kl": 0.00737762451171875,
"learning_rate": 7.214816693576234e-07,
"loss": 0.0003,
"reward": 0.6924525499343872,
"reward_std": 0.6107815653085709,
"rewards/cosine_scaled_reward": -0.13294040283653885,
"rewards/format_reward": 0.9583333432674408,
"step": 219
},
{
"completion_length": 1247.9791717529297,
"epoch": 0.25142857142857145,
"grad_norm": 0.22622907161712646,
"kl": 0.006031036376953125,
"learning_rate": 7.185729670371604e-07,
"loss": 0.0002,
"reward": 0.9419594034552574,
"reward_std": 0.675844706594944,
"rewards/cosine_scaled_reward": 0.023063029162585735,
"rewards/format_reward": 0.8958333432674408,
"step": 220
},
{
"completion_length": 2002.0833740234375,
"epoch": 0.25257142857142856,
"grad_norm": 0.218685120344162,
"kl": 0.0069580078125,
"learning_rate": 7.156560487081051e-07,
"loss": 0.0003,
"reward": 0.6575891096144915,
"reward_std": 0.6497488841414452,
"rewards/cosine_scaled_reward": 0.026711229234933853,
"rewards/format_reward": 0.6041666716337204,
"step": 221
},
{
"completion_length": 2015.6250915527344,
"epoch": 0.2537142857142857,
"grad_norm": 0.20831483602523804,
"kl": 0.0071258544921875,
"learning_rate": 7.127310565369415e-07,
"loss": 0.0003,
"reward": 0.14067217335104942,
"reward_std": 0.48574624210596085,
"rewards/cosine_scaled_reward": -0.28383059799671173,
"rewards/format_reward": 0.7083333432674408,
"step": 222
},
{
"completion_length": 1517.1875305175781,
"epoch": 0.25485714285714284,
"grad_norm": 0.24125142395496368,
"kl": 0.005817413330078125,
"learning_rate": 7.097981330836616e-07,
"loss": 0.0002,
"reward": 0.6348569616675377,
"reward_std": 0.5405807122588158,
"rewards/cosine_scaled_reward": -0.09923820104449987,
"rewards/format_reward": 0.8333333432674408,
"step": 223
},
{
"completion_length": 1914.0000305175781,
"epoch": 0.256,
"grad_norm": 0.2622263431549072,
"kl": 0.00728607177734375,
"learning_rate": 7.068574212948169e-07,
"loss": 0.0003,
"reward": 0.5525996647775173,
"reward_std": 0.5521951243281364,
"rewards/cosine_scaled_reward": -0.04661682341247797,
"rewards/format_reward": 0.6458333432674408,
"step": 224
},
{
"completion_length": 1250.0000305175781,
"epoch": 0.2571428571428571,
"grad_norm": 0.2181866317987442,
"kl": 0.00460052490234375,
"learning_rate": 7.039090644965509e-07,
"loss": 0.0002,
"reward": 1.2948355674743652,
"reward_std": 0.6228364408016205,
"rewards/cosine_scaled_reward": 0.16825110744684935,
"rewards/format_reward": 0.9583333432674408,
"step": 225
},
{
"completion_length": 1461.6666870117188,
"epoch": 0.2582857142857143,
"grad_norm": 0.36098772287368774,
"kl": 0.00760650634765625,
"learning_rate": 7.009532063876148e-07,
"loss": 0.0003,
"reward": 0.3821214698255062,
"reward_std": 0.5764878466725349,
"rewards/cosine_scaled_reward": -0.20477261394262314,
"rewards/format_reward": 0.7916666716337204,
"step": 226
},
{
"completion_length": 1328.5416870117188,
"epoch": 0.25942857142857145,
"grad_norm": 0.27139514684677124,
"kl": 0.00934600830078125,
"learning_rate": 6.979899910323624e-07,
"loss": 0.0004,
"reward": 0.7815765663981438,
"reward_std": 0.7309335023164749,
"rewards/cosine_scaled_reward": -0.03629505028948188,
"rewards/format_reward": 0.8541666865348816,
"step": 227
},
{
"completion_length": 1626.4167175292969,
"epoch": 0.26057142857142856,
"grad_norm": 0.23888561129570007,
"kl": 0.00714874267578125,
"learning_rate": 6.950195628537299e-07,
"loss": 0.0003,
"reward": 0.6099164858460426,
"reward_std": 0.7778853923082352,
"rewards/cosine_scaled_reward": -0.09087510220706463,
"rewards/format_reward": 0.7916666716337204,
"step": 228
},
{
"completion_length": 1440.6875457763672,
"epoch": 0.26171428571428573,
"grad_norm": 0.2864842116832733,
"kl": 0.0119171142578125,
"learning_rate": 6.920420666261961e-07,
"loss": 0.0005,
"reward": 0.7616857700049877,
"reward_std": 0.7498719990253448,
"rewards/cosine_scaled_reward": -0.025407111272215843,
"rewards/format_reward": 0.8125000149011612,
"step": 229
},
{
"completion_length": 1527.8958740234375,
"epoch": 0.26285714285714284,
"grad_norm": 0.2429640144109726,
"kl": 0.00768280029296875,
"learning_rate": 6.890576474687263e-07,
"loss": 0.0003,
"reward": 0.9651975035667419,
"reward_std": 0.824803501367569,
"rewards/cosine_scaled_reward": 0.06593206711113453,
"rewards/format_reward": 0.8333333432674408,
"step": 230
},
{
"completion_length": 1934.2084045410156,
"epoch": 0.264,
"grad_norm": 0.3963330090045929,
"kl": 0.0092926025390625,
"learning_rate": 6.860664508377001e-07,
"loss": 0.0004,
"reward": 0.6242162762209773,
"reward_std": 0.8598367348313332,
"rewards/cosine_scaled_reward": -0.03164188005030155,
"rewards/format_reward": 0.6875000149011612,
"step": 231
},
{
"completion_length": 1455.8333740234375,
"epoch": 0.2651428571428571,
"grad_norm": 0.23361533880233765,
"kl": 0.00798797607421875,
"learning_rate": 6.83068622519821e-07,
"loss": 0.0003,
"reward": 0.7301613166928291,
"reward_std": 0.7596315294504166,
"rewards/cosine_scaled_reward": -0.07241935143247247,
"rewards/format_reward": 0.8750000149011612,
"step": 232
},
{
"completion_length": 2010.8333740234375,
"epoch": 0.2662857142857143,
"grad_norm": 0.2177191823720932,
"kl": 0.0087738037109375,
"learning_rate": 6.800643086250121e-07,
"loss": 0.0004,
"reward": 0.6955921053886414,
"reward_std": 0.8746853768825531,
"rewards/cosine_scaled_reward": -0.0480372947640717,
"rewards/format_reward": 0.7916666865348816,
"step": 233
},
{
"completion_length": 1696.8541870117188,
"epoch": 0.2674285714285714,
"grad_norm": 0.2090214192867279,
"kl": 0.008090972900390625,
"learning_rate": 6.770536555792944e-07,
"loss": 0.0003,
"reward": 0.6760512292385101,
"reward_std": 0.6585969775915146,
"rewards/cosine_scaled_reward": -0.0578077242244035,
"rewards/format_reward": 0.7916666716337204,
"step": 234
},
{
"completion_length": 1389.6667175292969,
"epoch": 0.26857142857142857,
"grad_norm": 0.2434709221124649,
"kl": 0.006389617919921875,
"learning_rate": 6.740368101176495e-07,
"loss": 0.0003,
"reward": 0.6067800773307681,
"reward_std": 0.41319192945957184,
"rewards/cosine_scaled_reward": -0.1028599888086319,
"rewards/format_reward": 0.8125,
"step": 235
},
{
"completion_length": 1396.0625305175781,
"epoch": 0.26971428571428574,
"grad_norm": 0.23188619315624237,
"kl": 0.0070953369140625,
"learning_rate": 6.710139192768694e-07,
"loss": 0.0003,
"reward": 0.5102774500846863,
"reward_std": 0.4952257424592972,
"rewards/cosine_scaled_reward": -0.21361128613352776,
"rewards/format_reward": 0.9375,
"step": 236
},
{
"completion_length": 1236.3958892822266,
"epoch": 0.27085714285714285,
"grad_norm": 0.20340608060359955,
"kl": 0.00693511962890625,
"learning_rate": 6.679851303883891e-07,
"loss": 0.0003,
"reward": 0.9460461437702179,
"reward_std": 0.5520051866769791,
"rewards/cosine_scaled_reward": -0.01656026765704155,
"rewards/format_reward": 0.9791666716337204,
"step": 237
},
{
"completion_length": 1293.0208435058594,
"epoch": 0.272,
"grad_norm": 0.28505584597587585,
"kl": 0.00867462158203125,
"learning_rate": 6.649505910711058e-07,
"loss": 0.0003,
"reward": 0.6806494742631912,
"reward_std": 0.7478103041648865,
"rewards/cosine_scaled_reward": -0.11800861544907093,
"rewards/format_reward": 0.9166666716337204,
"step": 238
},
{
"completion_length": 1973.0000915527344,
"epoch": 0.27314285714285713,
"grad_norm": 0.29544979333877563,
"kl": 0.0121917724609375,
"learning_rate": 6.619104492241847e-07,
"loss": 0.0005,
"reward": 0.5413463786244392,
"reward_std": 0.7216374576091766,
"rewards/cosine_scaled_reward": -0.10432682058308274,
"rewards/format_reward": 0.7500000149011612,
"step": 239
},
{
"completion_length": 1372.6250305175781,
"epoch": 0.2742857142857143,
"grad_norm": 0.2549286484718323,
"kl": 0.00878143310546875,
"learning_rate": 6.588648530198504e-07,
"loss": 0.0004,
"reward": 1.4046210646629333,
"reward_std": 0.5067310631275177,
"rewards/cosine_scaled_reward": 0.2543938383460045,
"rewards/format_reward": 0.8958333432674408,
"step": 240
},
{
"completion_length": 1410.0417175292969,
"epoch": 0.2754285714285714,
"grad_norm": 0.32747364044189453,
"kl": 0.009124755859375,
"learning_rate": 6.558139508961654e-07,
"loss": 0.0004,
"reward": 0.8229547590017319,
"reward_std": 0.6901429891586304,
"rewards/cosine_scaled_reward": -0.026022649370133877,
"rewards/format_reward": 0.8750000149011612,
"step": 241
},
{
"completion_length": 1473.6875228881836,
"epoch": 0.2765714285714286,
"grad_norm": 0.3592205345630646,
"kl": 0.012908935546875,
"learning_rate": 6.527578915497951e-07,
"loss": 0.0005,
"reward": 0.7551293671131134,
"reward_std": 0.663729339838028,
"rewards/cosine_scaled_reward": -0.0599353089928627,
"rewards/format_reward": 0.8750000149011612,
"step": 242
},
{
"completion_length": 1187.0625610351562,
"epoch": 0.2777142857142857,
"grad_norm": 0.22532759606838226,
"kl": 0.0076751708984375,
"learning_rate": 6.496968239287603e-07,
"loss": 0.0003,
"reward": 0.8520705178380013,
"reward_std": 0.45504674315452576,
"rewards/cosine_scaled_reward": -0.05313139781355858,
"rewards/format_reward": 0.9583333432674408,
"step": 243
},
{
"completion_length": 1377.7291870117188,
"epoch": 0.27885714285714286,
"grad_norm": 0.2351863533258438,
"kl": 0.00748443603515625,
"learning_rate": 6.466308972251785e-07,
"loss": 0.0003,
"reward": 1.1842002123594284,
"reward_std": 1.0244361460208893,
"rewards/cosine_scaled_reward": 0.15460011083632708,
"rewards/format_reward": 0.8750000149011612,
"step": 244
},
{
"completion_length": 1255.4167175292969,
"epoch": 0.28,
"grad_norm": 0.34309011697769165,
"kl": 0.01132965087890625,
"learning_rate": 6.435602608679916e-07,
"loss": 0.0005,
"reward": 0.8439341634511948,
"reward_std": 0.6376037150621414,
"rewards/cosine_scaled_reward": -0.025949583388864994,
"rewards/format_reward": 0.8958333432674408,
"step": 245
},
{
"completion_length": 1144.6250305175781,
"epoch": 0.28114285714285714,
"grad_norm": 0.27464959025382996,
"kl": 0.0088958740234375,
"learning_rate": 6.404850645156841e-07,
"loss": 0.0004,
"reward": 0.9792153835296631,
"reward_std": 0.5227902606129646,
"rewards/cosine_scaled_reward": 2.4352222681045532e-05,
"rewards/format_reward": 0.9791666716337204,
"step": 246
},
{
"completion_length": 985.2083587646484,
"epoch": 0.2822857142857143,
"grad_norm": 0.3299963176250458,
"kl": 0.01165771484375,
"learning_rate": 6.374054580489873e-07,
"loss": 0.0005,
"reward": 1.1308997794985771,
"reward_std": 0.699935294687748,
"rewards/cosine_scaled_reward": 0.09669988602399826,
"rewards/format_reward": 0.9375,
"step": 247
},
{
"completion_length": 1719.0208740234375,
"epoch": 0.2834285714285714,
"grad_norm": 0.27527329325675964,
"kl": 0.0107879638671875,
"learning_rate": 6.343215915635761e-07,
"loss": 0.0004,
"reward": 0.808366172015667,
"reward_std": 0.725908488035202,
"rewards/cosine_scaled_reward": -0.012483585625886917,
"rewards/format_reward": 0.8333333432674408,
"step": 248
},
{
"completion_length": 1970.2083740234375,
"epoch": 0.2845714285714286,
"grad_norm": 0.27165958285331726,
"kl": 0.011810302734375,
"learning_rate": 6.31233615362752e-07,
"loss": 0.0005,
"reward": 0.3851170837879181,
"reward_std": 0.7313886731863022,
"rewards/cosine_scaled_reward": -0.16160813719034195,
"rewards/format_reward": 0.7083333432674408,
"step": 249
},
{
"completion_length": 1342.3750305175781,
"epoch": 0.2857142857142857,
"grad_norm": 0.2968628704547882,
"kl": 0.00905609130859375,
"learning_rate": 6.281416799501187e-07,
"loss": 0.0004,
"reward": 0.8386699110269547,
"reward_std": 0.7474230378866196,
"rewards/cosine_scaled_reward": -0.02858173381537199,
"rewards/format_reward": 0.8958333432674408,
"step": 250
},
{
"completion_length": 1774.0208740234375,
"epoch": 0.28685714285714287,
"grad_norm": 0.39884844422340393,
"kl": 0.0139312744140625,
"learning_rate": 6.25045936022246e-07,
"loss": 0.0006,
"reward": 1.0918036848306656,
"reward_std": 1.1108526289463043,
"rewards/cosine_scaled_reward": 0.1917351707816124,
"rewards/format_reward": 0.708333358168602,
"step": 251
},
{
"completion_length": 1604.7500610351562,
"epoch": 0.288,
"grad_norm": 0.32672712206840515,
"kl": 0.0103912353515625,
"learning_rate": 6.219465344613258e-07,
"loss": 0.0004,
"reward": 0.7797054275870323,
"reward_std": 0.6628324761986732,
"rewards/cosine_scaled_reward": -0.06848062574863434,
"rewards/format_reward": 0.9166666716337204,
"step": 252
},
{
"completion_length": 1578.9167175292969,
"epoch": 0.28914285714285715,
"grad_norm": 0.3023277223110199,
"kl": 0.01739501953125,
"learning_rate": 6.188436263278172e-07,
"loss": 0.0007,
"reward": 0.7126835100352764,
"reward_std": 0.786981999874115,
"rewards/cosine_scaled_reward": -0.039491571485996246,
"rewards/format_reward": 0.7916666716337204,
"step": 253
},
{
"completion_length": 1142.6458435058594,
"epoch": 0.29028571428571426,
"grad_norm": 0.4636065363883972,
"kl": 0.0141448974609375,
"learning_rate": 6.157373628530852e-07,
"loss": 0.0006,
"reward": 1.1053122580051422,
"reward_std": 0.47931085526943207,
"rewards/cosine_scaled_reward": 0.16723946738056839,
"rewards/format_reward": 0.7708333432674408,
"step": 254
},
{
"completion_length": 1286.0833435058594,
"epoch": 0.2914285714285714,
"grad_norm": 0.28858116269111633,
"kl": 0.009735107421875,
"learning_rate": 6.126278954320294e-07,
"loss": 0.0004,
"reward": 1.1211883053183556,
"reward_std": 0.4671914726495743,
"rewards/cosine_scaled_reward": 0.09184413589537144,
"rewards/format_reward": 0.9375000149011612,
"step": 255
},
{
"completion_length": 1213.2083740234375,
"epoch": 0.2925714285714286,
"grad_norm": 0.32043886184692383,
"kl": 0.015838623046875,
"learning_rate": 6.095153756157051e-07,
"loss": 0.0006,
"reward": 0.823145791888237,
"reward_std": 0.5544667765498161,
"rewards/cosine_scaled_reward": -0.0675937756896019,
"rewards/format_reward": 0.9583333432674408,
"step": 256
},
{
"completion_length": 1316.8541870117188,
"epoch": 0.2937142857142857,
"grad_norm": 0.23760062456130981,
"kl": 0.0098114013671875,
"learning_rate": 6.06399955103937e-07,
"loss": 0.0004,
"reward": 0.49779732525348663,
"reward_std": 0.46498920768499374,
"rewards/cosine_scaled_reward": -0.23026802763342857,
"rewards/format_reward": 0.9583333432674408,
"step": 257
},
{
"completion_length": 1751.2708435058594,
"epoch": 0.2948571428571429,
"grad_norm": 0.254151314496994,
"kl": 0.0131378173828125,
"learning_rate": 6.032817857379256e-07,
"loss": 0.0005,
"reward": 0.6079542301595211,
"reward_std": 0.6472693011164665,
"rewards/cosine_scaled_reward": -0.10227290168404579,
"rewards/format_reward": 0.8125000149011612,
"step": 258
},
{
"completion_length": 978.6458435058594,
"epoch": 0.296,
"grad_norm": 0.29362747073173523,
"kl": 0.00855255126953125,
"learning_rate": 6.001610194928464e-07,
"loss": 0.0003,
"reward": 0.7723531350493431,
"reward_std": 0.7641957998275757,
"rewards/cosine_scaled_reward": -0.10340679436922073,
"rewards/format_reward": 0.9791666716337204,
"step": 259
},
{
"completion_length": 1655.1875305175781,
"epoch": 0.29714285714285715,
"grad_norm": 0.38929465413093567,
"kl": 0.0145721435546875,
"learning_rate": 5.97037808470444e-07,
"loss": 0.0006,
"reward": 0.37256848718971014,
"reward_std": 0.6071458011865616,
"rewards/cosine_scaled_reward": -0.18871578108519316,
"rewards/format_reward": 0.7500000298023224,
"step": 260
},
{
"completion_length": 1198.8333892822266,
"epoch": 0.29828571428571427,
"grad_norm": 0.2716739773750305,
"kl": 0.0102081298828125,
"learning_rate": 5.939123048916173e-07,
"loss": 0.0004,
"reward": 0.633615754544735,
"reward_std": 0.586112380027771,
"rewards/cosine_scaled_reward": -0.1519421450793743,
"rewards/format_reward": 0.9375000149011612,
"step": 261
},
{
"completion_length": 1484.1042175292969,
"epoch": 0.29942857142857143,
"grad_norm": 0.2628130316734314,
"kl": 0.0106353759765625,
"learning_rate": 5.907846610890011e-07,
"loss": 0.0004,
"reward": 0.4845110587775707,
"reward_std": 0.536693274974823,
"rewards/cosine_scaled_reward": -0.18482780829071999,
"rewards/format_reward": 0.8541666716337204,
"step": 262
},
{
"completion_length": 1064.0833587646484,
"epoch": 0.30057142857142854,
"grad_norm": 0.36321043968200684,
"kl": 0.0098876953125,
"learning_rate": 5.87655029499542e-07,
"loss": 0.0004,
"reward": 0.9257199168205261,
"reward_std": 0.8942077457904816,
"rewards/cosine_scaled_reward": -0.016306710429489613,
"rewards/format_reward": 0.9583333432674408,
"step": 263
},
{
"completion_length": 1316.1042175292969,
"epoch": 0.3017142857142857,
"grad_norm": 0.26243484020233154,
"kl": 0.00811004638671875,
"learning_rate": 5.845235626570683e-07,
"loss": 0.0003,
"reward": 0.6479791402816772,
"reward_std": 0.6652240380644798,
"rewards/cosine_scaled_reward": -0.15517710940912366,
"rewards/format_reward": 0.9583333432674408,
"step": 264
},
{
"completion_length": 1909.6250915527344,
"epoch": 0.3028571428571429,
"grad_norm": 0.2833210229873657,
"kl": 0.01206207275390625,
"learning_rate": 5.813904131848564e-07,
"loss": 0.0005,
"reward": 0.7709198147058487,
"reward_std": 0.6459413915872574,
"rewards/cosine_scaled_reward": 4.323199391365051e-05,
"rewards/format_reward": 0.7708333432674408,
"step": 265
},
{
"completion_length": 1645.9583740234375,
"epoch": 0.304,
"grad_norm": 0.3645997941493988,
"kl": 0.01513671875,
"learning_rate": 5.78255733788191e-07,
"loss": 0.0006,
"reward": 0.4407457821071148,
"reward_std": 0.6104073449969292,
"rewards/cosine_scaled_reward": -0.1858771131373942,
"rewards/format_reward": 0.8125000149011612,
"step": 266
},
{
"completion_length": 1644.0209045410156,
"epoch": 0.30514285714285716,
"grad_norm": 0.27599290013313293,
"kl": 0.014923095703125,
"learning_rate": 5.751196772469237e-07,
"loss": 0.0006,
"reward": 0.7639178857207298,
"reward_std": 0.5990442484617233,
"rewards/cosine_scaled_reward": -0.013874400407075882,
"rewards/format_reward": 0.7916666716337204,
"step": 267
},
{
"completion_length": 1681.4583587646484,
"epoch": 0.3062857142857143,
"grad_norm": 0.2269752472639084,
"kl": 0.01216888427734375,
"learning_rate": 5.71982396408026e-07,
"loss": 0.0005,
"reward": 0.5377051346004009,
"reward_std": 0.6556981913745403,
"rewards/cosine_scaled_reward": -0.14781412575393915,
"rewards/format_reward": 0.8333333358168602,
"step": 268
},
{
"completion_length": 1047.1250457763672,
"epoch": 0.30742857142857144,
"grad_norm": 0.3485262095928192,
"kl": 0.01123046875,
"learning_rate": 5.688440441781398e-07,
"loss": 0.0005,
"reward": 0.68592269718647,
"reward_std": 0.5478300377726555,
"rewards/cosine_scaled_reward": -0.13620532862842083,
"rewards/format_reward": 0.9583333432674408,
"step": 269
},
{
"completion_length": 1012.0208587646484,
"epoch": 0.30857142857142855,
"grad_norm": 0.3116324543952942,
"kl": 0.007354736328125,
"learning_rate": 5.657047735161255e-07,
"loss": 0.0003,
"reward": 0.9341000914573669,
"reward_std": 0.6274480298161507,
"rewards/cosine_scaled_reward": -0.03294998221099377,
"rewards/format_reward": 1.0,
"step": 270
},
{
"completion_length": 1732.6459350585938,
"epoch": 0.3097142857142857,
"grad_norm": 0.42208629846572876,
"kl": 0.016632080078125,
"learning_rate": 5.625647374256061e-07,
"loss": 0.0007,
"reward": 0.7596426885575056,
"reward_std": 0.77412149310112,
"rewards/cosine_scaled_reward": -0.0055953278206288815,
"rewards/format_reward": 0.7708333432674408,
"step": 271
},
{
"completion_length": 2177.270965576172,
"epoch": 0.31085714285714283,
"grad_norm": 0.5940233469009399,
"kl": 0.02764892578125,
"learning_rate": 5.594240889475106e-07,
"loss": 0.0011,
"reward": 0.4223189577460289,
"reward_std": 0.8890358135104179,
"rewards/cosine_scaled_reward": -0.10134052112698555,
"rewards/format_reward": 0.6250000223517418,
"step": 272
},
{
"completion_length": 1954.8333740234375,
"epoch": 0.312,
"grad_norm": 0.3138795495033264,
"kl": 0.02874755859375,
"learning_rate": 5.562829811526154e-07,
"loss": 0.0012,
"reward": 0.3654465600848198,
"reward_std": 0.5650510713458061,
"rewards/cosine_scaled_reward": -0.1506100632250309,
"rewards/format_reward": 0.6666666865348816,
"step": 273
},
{
"completion_length": 1267.4583587646484,
"epoch": 0.31314285714285717,
"grad_norm": 0.274538516998291,
"kl": 0.01172637939453125,
"learning_rate": 5.531415671340826e-07,
"loss": 0.0005,
"reward": 0.837937019765377,
"reward_std": 0.6332506015896797,
"rewards/cosine_scaled_reward": -0.03936483711004257,
"rewards/format_reward": 0.9166666716337204,
"step": 274
},
{
"completion_length": 2021.4583740234375,
"epoch": 0.3142857142857143,
"grad_norm": 0.30807891488075256,
"kl": 0.02382659912109375,
"learning_rate": 5.5e-07,
"loss": 0.001,
"reward": 0.9066380485892296,
"reward_std": 0.9834412485361099,
"rewards/cosine_scaled_reward": 0.1095690238289535,
"rewards/format_reward": 0.6875000223517418,
"step": 275
},
{
"completion_length": 1135.8750305175781,
"epoch": 0.31542857142857145,
"grad_norm": 0.2739965319633484,
"kl": 0.010345458984375,
"learning_rate": 5.468584328659172e-07,
"loss": 0.0004,
"reward": 1.0236308723688126,
"reward_std": 0.5140665993094444,
"rewards/cosine_scaled_reward": 0.04306542640551925,
"rewards/format_reward": 0.9375000149011612,
"step": 276
},
{
"completion_length": 1247.2500305175781,
"epoch": 0.31657142857142856,
"grad_norm": 0.37156689167022705,
"kl": 0.00988006591796875,
"learning_rate": 5.437170188473847e-07,
"loss": 0.0004,
"reward": 0.8927154019474983,
"reward_std": 0.8232090175151825,
"rewards/cosine_scaled_reward": -0.0015590004622936249,
"rewards/format_reward": 0.8958333432674408,
"step": 277
},
{
"completion_length": 1309.3125610351562,
"epoch": 0.3177142857142857,
"grad_norm": 0.2857230603694916,
"kl": 0.0121002197265625,
"learning_rate": 5.405759110524894e-07,
"loss": 0.0005,
"reward": 0.942589208483696,
"reward_std": 0.6875655725598335,
"rewards/cosine_scaled_reward": 0.012961250729858875,
"rewards/format_reward": 0.9166666716337204,
"step": 278
},
{
"completion_length": 1435.4791717529297,
"epoch": 0.31885714285714284,
"grad_norm": 0.43713676929473877,
"kl": 0.01840972900390625,
"learning_rate": 5.37435262574394e-07,
"loss": 0.0007,
"reward": 0.37974046915769577,
"reward_std": 0.5385972559452057,
"rewards/cosine_scaled_reward": -0.2163797914981842,
"rewards/format_reward": 0.8125,
"step": 279
},
{
"completion_length": 1156.062515258789,
"epoch": 0.32,
"grad_norm": 0.2634413242340088,
"kl": 0.0093536376953125,
"learning_rate": 5.342952264838747e-07,
"loss": 0.0004,
"reward": 0.980226680636406,
"reward_std": 0.5487889721989632,
"rewards/cosine_scaled_reward": 0.0005300038028508425,
"rewards/format_reward": 0.9791666716337204,
"step": 280
},
{
"completion_length": 1187.3125457763672,
"epoch": 0.3211428571428571,
"grad_norm": 0.32134756445884705,
"kl": 0.008182525634765625,
"learning_rate": 5.311559558218603e-07,
"loss": 0.0003,
"reward": 0.9637440145015717,
"reward_std": 0.8387380540370941,
"rewards/cosine_scaled_reward": 0.0027053444646298885,
"rewards/format_reward": 0.9583333432674408,
"step": 281
},
{
"completion_length": 1651.604232788086,
"epoch": 0.3222857142857143,
"grad_norm": 0.37761881947517395,
"kl": 0.0242919921875,
"learning_rate": 5.28017603591974e-07,
"loss": 0.001,
"reward": 0.5987202003598213,
"reward_std": 0.6141614019870758,
"rewards/cosine_scaled_reward": -0.10688992030918598,
"rewards/format_reward": 0.8125000149011612,
"step": 282
},
{
"completion_length": 1207.1250610351562,
"epoch": 0.32342857142857145,
"grad_norm": 0.3965184688568115,
"kl": 0.0155487060546875,
"learning_rate": 5.248803227530763e-07,
"loss": 0.0006,
"reward": 0.6064153388142586,
"reward_std": 0.6052896529436111,
"rewards/cosine_scaled_reward": -0.16554233682109043,
"rewards/format_reward": 0.9375000149011612,
"step": 283
},
{
"completion_length": 1520.5625610351562,
"epoch": 0.32457142857142857,
"grad_norm": 0.4802095890045166,
"kl": 0.01385498046875,
"learning_rate": 5.21744266211809e-07,
"loss": 0.0006,
"reward": 0.6298409104347229,
"reward_std": 0.7485504075884819,
"rewards/cosine_scaled_reward": -0.10174621269106865,
"rewards/format_reward": 0.833333358168602,
"step": 284
},
{
"completion_length": 1631.4166870117188,
"epoch": 0.32571428571428573,
"grad_norm": 0.5242775678634644,
"kl": 0.0250244140625,
"learning_rate": 5.186095868151436e-07,
"loss": 0.001,
"reward": 0.4431188479065895,
"reward_std": 0.6355130672454834,
"rewards/cosine_scaled_reward": -0.20552390813827515,
"rewards/format_reward": 0.8541666865348816,
"step": 285
},
{
"completion_length": 1278.2917175292969,
"epoch": 0.32685714285714285,
"grad_norm": 0.2374052256345749,
"kl": 0.009552001953125,
"learning_rate": 5.154764373429315e-07,
"loss": 0.0004,
"reward": 0.9486149102449417,
"reward_std": 0.6938119828701019,
"rewards/cosine_scaled_reward": -0.025692567229270935,
"rewards/format_reward": 1.0,
"step": 286
},
{
"completion_length": 1549.5208740234375,
"epoch": 0.328,
"grad_norm": 0.5496286749839783,
"kl": 0.0344696044921875,
"learning_rate": 5.123449705004581e-07,
"loss": 0.0014,
"reward": 0.6046592518687248,
"reward_std": 0.7527553886175156,
"rewards/cosine_scaled_reward": -0.12475371174514294,
"rewards/format_reward": 0.8541666865348816,
"step": 287
},
{
"completion_length": 906.6041717529297,
"epoch": 0.3291428571428571,
"grad_norm": 0.40935972332954407,
"kl": 0.014556884765625,
"learning_rate": 5.09215338910999e-07,
"loss": 0.0006,
"reward": 1.1515108793973923,
"reward_std": 0.670855775475502,
"rewards/cosine_scaled_reward": 0.07575542479753494,
"rewards/format_reward": 1.0,
"step": 288
},
{
"completion_length": 1302.3958587646484,
"epoch": 0.3302857142857143,
"grad_norm": 0.33956480026245117,
"kl": 0.01055145263671875,
"learning_rate": 5.060876951083828e-07,
"loss": 0.0004,
"reward": 0.8325834274291992,
"reward_std": 0.469268262386322,
"rewards/cosine_scaled_reward": -0.0628749430179596,
"rewards/format_reward": 0.9583333432674408,
"step": 289
},
{
"completion_length": 1219.1667022705078,
"epoch": 0.3314285714285714,
"grad_norm": 0.5517088174819946,
"kl": 0.0243072509765625,
"learning_rate": 5.02962191529556e-07,
"loss": 0.001,
"reward": 1.3289316296577454,
"reward_std": 0.8555012494325638,
"rewards/cosine_scaled_reward": 0.1957157626748085,
"rewards/format_reward": 0.9375000149011612,
"step": 290
},
{
"completion_length": 1010.8541870117188,
"epoch": 0.3325714285714286,
"grad_norm": 0.44879335165023804,
"kl": 0.020660400390625,
"learning_rate": 4.998389805071536e-07,
"loss": 0.0008,
"reward": 0.886590301990509,
"reward_std": 0.6713649779558182,
"rewards/cosine_scaled_reward": 0.00579514354467392,
"rewards/format_reward": 0.8750000149011612,
"step": 291
},
{
"completion_length": 1223.687515258789,
"epoch": 0.33371428571428574,
"grad_norm": 0.3632715344429016,
"kl": 0.0165252685546875,
"learning_rate": 4.967182142620745e-07,
"loss": 0.0007,
"reward": 0.8943772986531258,
"reward_std": 0.8097474128007889,
"rewards/cosine_scaled_reward": -0.0007280493155121803,
"rewards/format_reward": 0.8958333432674408,
"step": 292
},
{
"completion_length": 1374.0208740234375,
"epoch": 0.33485714285714285,
"grad_norm": 0.4008006155490875,
"kl": 0.0230865478515625,
"learning_rate": 4.93600044896063e-07,
"loss": 0.0009,
"reward": 0.6955159157514572,
"reward_std": 0.6655403971672058,
"rewards/cosine_scaled_reward": -0.11057540401816368,
"rewards/format_reward": 0.9166666865348816,
"step": 293
},
{
"completion_length": 1269.2917022705078,
"epoch": 0.336,
"grad_norm": 0.20706599950790405,
"kl": 0.00942230224609375,
"learning_rate": 4.904846243842949e-07,
"loss": 0.0004,
"reward": 1.2973814010620117,
"reward_std": 0.7491715997457504,
"rewards/cosine_scaled_reward": 0.16952402517199516,
"rewards/format_reward": 0.9583333432674408,
"step": 294
},
{
"completion_length": 2128.9375610351562,
"epoch": 0.33714285714285713,
"grad_norm": 0.3372494876384735,
"kl": 0.03973388671875,
"learning_rate": 4.873721045679706e-07,
"loss": 0.0016,
"reward": 0.4113108851015568,
"reward_std": 0.6322937309741974,
"rewards/cosine_scaled_reward": -0.11726122908294201,
"rewards/format_reward": 0.6458333432674408,
"step": 295
},
{
"completion_length": 938.1042022705078,
"epoch": 0.3382857142857143,
"grad_norm": 0.3853781521320343,
"kl": 0.01088714599609375,
"learning_rate": 4.842626371469149e-07,
"loss": 0.0004,
"reward": 1.2917412221431732,
"reward_std": 0.859960287809372,
"rewards/cosine_scaled_reward": 0.16670391708612442,
"rewards/format_reward": 0.9583333432674408,
"step": 296
},
{
"completion_length": 1234.5000305175781,
"epoch": 0.3394285714285714,
"grad_norm": 0.32493123412132263,
"kl": 0.02001953125,
"learning_rate": 4.811563736721829e-07,
"loss": 0.0008,
"reward": 0.5709987878799438,
"reward_std": 0.5739526003599167,
"rewards/cosine_scaled_reward": -0.1728339404799044,
"rewards/format_reward": 0.9166666865348816,
"step": 297
},
{
"completion_length": 1737.8750305175781,
"epoch": 0.3405714285714286,
"grad_norm": 0.6288060545921326,
"kl": 0.02978515625,
"learning_rate": 4.780534655386743e-07,
"loss": 0.0012,
"reward": 0.7676291763782501,
"reward_std": 0.6836749911308289,
"rewards/cosine_scaled_reward": -0.0328520848415792,
"rewards/format_reward": 0.833333358168602,
"step": 298
},
{
"completion_length": 1495.9583740234375,
"epoch": 0.3417142857142857,
"grad_norm": 0.5687032341957092,
"kl": 0.02923583984375,
"learning_rate": 4.749540639777539e-07,
"loss": 0.0012,
"reward": 0.7650880664587021,
"reward_std": 0.7526437044143677,
"rewards/cosine_scaled_reward": -0.0757893230766058,
"rewards/format_reward": 0.9166666716337204,
"step": 299
},
{
"completion_length": 1482.9167022705078,
"epoch": 0.34285714285714286,
"grad_norm": 0.307375431060791,
"kl": 0.0286102294921875,
"learning_rate": 4.7185832004988133e-07,
"loss": 0.0011,
"reward": 1.0939996913075447,
"reward_std": 0.6226199977099895,
"rewards/cosine_scaled_reward": 0.10949981957674026,
"rewards/format_reward": 0.875,
"step": 300
},
{
"completion_length": 1668.7708740234375,
"epoch": 0.344,
"grad_norm": 0.3655478358268738,
"kl": 0.0385589599609375,
"learning_rate": 4.68766384637248e-07,
"loss": 0.0015,
"reward": 0.9393804222345352,
"reward_std": 0.5472998023033142,
"rewards/cosine_scaled_reward": 0.04260684549808502,
"rewards/format_reward": 0.8541666865348816,
"step": 301
},
{
"completion_length": 1460.1667175292969,
"epoch": 0.34514285714285714,
"grad_norm": 0.5648999810218811,
"kl": 0.0326690673828125,
"learning_rate": 4.656784084364238e-07,
"loss": 0.0013,
"reward": 0.8058248609304428,
"reward_std": 0.5371805727481842,
"rewards/cosine_scaled_reward": -0.034587569534778595,
"rewards/format_reward": 0.8750000149011612,
"step": 302
},
{
"completion_length": 1176.3958740234375,
"epoch": 0.3462857142857143,
"grad_norm": 0.2893832325935364,
"kl": 0.01714324951171875,
"learning_rate": 4.6259454195101267e-07,
"loss": 0.0007,
"reward": 0.9501863233745098,
"reward_std": 0.7773154973983765,
"rewards/cosine_scaled_reward": 0.006343139801174402,
"rewards/format_reward": 0.9375000149011612,
"step": 303
},
{
"completion_length": 1049.2708587646484,
"epoch": 0.3474285714285714,
"grad_norm": 0.3396141231060028,
"kl": 0.017486572265625,
"learning_rate": 4.59514935484316e-07,
"loss": 0.0007,
"reward": 1.3814565241336823,
"reward_std": 0.7759077772498131,
"rewards/cosine_scaled_reward": 0.21156160347163677,
"rewards/format_reward": 0.9583333432674408,
"step": 304
},
{
"completion_length": 1675.1667175292969,
"epoch": 0.3485714285714286,
"grad_norm": 0.32787981629371643,
"kl": 0.0314483642578125,
"learning_rate": 4.5643973913200837e-07,
"loss": 0.0013,
"reward": 0.7963635921478271,
"reward_std": 0.7029048502445221,
"rewards/cosine_scaled_reward": -0.018484866246581078,
"rewards/format_reward": 0.8333333432674408,
"step": 305
},
{
"completion_length": 1478.7916870117188,
"epoch": 0.3497142857142857,
"grad_norm": 0.7724860310554504,
"kl": 0.04205322265625,
"learning_rate": 4.5336910277482155e-07,
"loss": 0.0017,
"reward": 0.934022843837738,
"reward_std": 0.6418131068348885,
"rewards/cosine_scaled_reward": 0.039928069338202477,
"rewards/format_reward": 0.8541666716337204,
"step": 306
},
{
"completion_length": 1466.062515258789,
"epoch": 0.35085714285714287,
"grad_norm": 0.7901880741119385,
"kl": 0.048187255859375,
"learning_rate": 4.503031760712397e-07,
"loss": 0.0019,
"reward": 1.208159700036049,
"reward_std": 0.9337977021932602,
"rewards/cosine_scaled_reward": 0.16657985746860504,
"rewards/format_reward": 0.8750000149011612,
"step": 307
},
{
"completion_length": 1011.0,
"epoch": 0.352,
"grad_norm": 0.43838247656822205,
"kl": 0.02608489990234375,
"learning_rate": 4.4724210845020494e-07,
"loss": 0.001,
"reward": 0.8291152790188789,
"reward_std": 0.4770050719380379,
"rewards/cosine_scaled_reward": -0.022942371666431427,
"rewards/format_reward": 0.875,
"step": 308
},
{
"completion_length": 1304.812515258789,
"epoch": 0.35314285714285715,
"grad_norm": 0.34449389576911926,
"kl": 0.027618408203125,
"learning_rate": 4.441860491038345e-07,
"loss": 0.0011,
"reward": 0.9580995887517929,
"reward_std": 0.9287254959344864,
"rewards/cosine_scaled_reward": 0.020716451108455658,
"rewards/format_reward": 0.9166666716337204,
"step": 309
},
{
"completion_length": 1587.1875610351562,
"epoch": 0.35428571428571426,
"grad_norm": 0.3894862234592438,
"kl": 0.0413360595703125,
"learning_rate": 4.4113514698014953e-07,
"loss": 0.0017,
"reward": 0.967779666185379,
"reward_std": 0.49595198780298233,
"rewards/cosine_scaled_reward": 0.04638980980962515,
"rewards/format_reward": 0.875,
"step": 310
},
{
"completion_length": 1005.9583587646484,
"epoch": 0.3554285714285714,
"grad_norm": 0.5111984610557556,
"kl": 0.0352020263671875,
"learning_rate": 4.3808955077581546e-07,
"loss": 0.0014,
"reward": 0.8524645194411278,
"reward_std": 0.6346431374549866,
"rewards/cosine_scaled_reward": -0.042517755180597305,
"rewards/format_reward": 0.9375,
"step": 311
},
{
"completion_length": 1219.1042175292969,
"epoch": 0.3565714285714286,
"grad_norm": 0.24494509398937225,
"kl": 0.0164337158203125,
"learning_rate": 4.350494089288943e-07,
"loss": 0.0007,
"reward": 1.1846633851528168,
"reward_std": 0.8452874422073364,
"rewards/cosine_scaled_reward": 0.09233169769868255,
"rewards/format_reward": 1.0,
"step": 312
},
{
"completion_length": 1772.6459045410156,
"epoch": 0.3577142857142857,
"grad_norm": 0.32232236862182617,
"kl": 0.05014801025390625,
"learning_rate": 4.3201486961161093e-07,
"loss": 0.002,
"reward": 0.7452734671533108,
"reward_std": 0.5130416378378868,
"rewards/cosine_scaled_reward": -0.09611329552717507,
"rewards/format_reward": 0.9375000149011612,
"step": 313
},
{
"completion_length": 1450.9167022705078,
"epoch": 0.3588571428571429,
"grad_norm": 0.7095328569412231,
"kl": 0.04193115234375,
"learning_rate": 4.2898608072313045e-07,
"loss": 0.0017,
"reward": 0.695801317691803,
"reward_std": 0.6888641864061356,
"rewards/cosine_scaled_reward": -0.07918267324566841,
"rewards/format_reward": 0.8541666865348816,
"step": 314
},
{
"completion_length": 890.2083435058594,
"epoch": 0.36,
"grad_norm": 0.8648350238800049,
"kl": 0.02002716064453125,
"learning_rate": 4.2596318988235037e-07,
"loss": 0.0008,
"reward": 0.8261366635560989,
"reward_std": 0.5876666381955147,
"rewards/cosine_scaled_reward": -0.07651501428335905,
"rewards/format_reward": 0.9791666716337204,
"step": 315
},
{
"completion_length": 1495.2917175292969,
"epoch": 0.36114285714285715,
"grad_norm": 0.7055426239967346,
"kl": 0.0330047607421875,
"learning_rate": 4.2294634442070553e-07,
"loss": 0.0013,
"reward": 0.5078188478946686,
"reward_std": 0.6738529801368713,
"rewards/cosine_scaled_reward": -0.18359058536589146,
"rewards/format_reward": 0.8750000149011612,
"step": 316
},
{
"completion_length": 2028.3959350585938,
"epoch": 0.36228571428571427,
"grad_norm": 0.6188425421714783,
"kl": 0.1023101806640625,
"learning_rate": 4.1993569137498776e-07,
"loss": 0.0041,
"reward": 0.6971778050065041,
"reward_std": 0.6844265758991241,
"rewards/cosine_scaled_reward": 0.015255570411682129,
"rewards/format_reward": 0.6666666865348816,
"step": 317
},
{
"completion_length": 1409.0833740234375,
"epoch": 0.36342857142857143,
"grad_norm": 0.4974954426288605,
"kl": 0.04302978515625,
"learning_rate": 4.1693137748017915e-07,
"loss": 0.0017,
"reward": 0.9140654609072953,
"reward_std": 0.590464636683464,
"rewards/cosine_scaled_reward": 0.040366058237850666,
"rewards/format_reward": 0.8333333432674408,
"step": 318
},
{
"completion_length": 1569.2500305175781,
"epoch": 0.36457142857142855,
"grad_norm": 0.3885625898838043,
"kl": 0.0399017333984375,
"learning_rate": 4.1393354916230005e-07,
"loss": 0.0016,
"reward": 0.7486050575971603,
"reward_std": 0.9239681512117386,
"rewards/cosine_scaled_reward": -0.08403081598225981,
"rewards/format_reward": 0.9166666865348816,
"step": 319
},
{
"completion_length": 1059.1875305175781,
"epoch": 0.3657142857142857,
"grad_norm": 0.5842506289482117,
"kl": 0.0222625732421875,
"learning_rate": 4.1094235253127374e-07,
"loss": 0.0009,
"reward": 0.872907280921936,
"reward_std": 0.7717489525675774,
"rewards/cosine_scaled_reward": -0.0635463809594512,
"rewards/format_reward": 1.0,
"step": 320
},
{
"completion_length": 1675.8334045410156,
"epoch": 0.3668571428571429,
"grad_norm": 1.7648111581802368,
"kl": 0.10723876953125,
"learning_rate": 4.079579333738039e-07,
"loss": 0.0043,
"reward": 0.6553980484604836,
"reward_std": 0.6528958007693291,
"rewards/cosine_scaled_reward": -0.06813432276248932,
"rewards/format_reward": 0.7916666865348816,
"step": 321
},
{
"completion_length": 1275.2708587646484,
"epoch": 0.368,
"grad_norm": 0.6659455895423889,
"kl": 0.0338134765625,
"learning_rate": 4.0498043714627006e-07,
"loss": 0.0013,
"reward": 1.113847702741623,
"reward_std": 0.7288860827684402,
"rewards/cosine_scaled_reward": 0.14025717787444592,
"rewards/format_reward": 0.8333333432674408,
"step": 322
},
{
"completion_length": 1612.9166717529297,
"epoch": 0.36914285714285716,
"grad_norm": 0.6660499572753906,
"kl": 0.0994873046875,
"learning_rate": 4.020100089676376e-07,
"loss": 0.004,
"reward": 1.0578741058707237,
"reward_std": 0.49694500118494034,
"rewards/cosine_scaled_reward": 0.09143703989684582,
"rewards/format_reward": 0.8750000149011612,
"step": 323
},
{
"completion_length": 1938.4584045410156,
"epoch": 0.3702857142857143,
"grad_norm": 1.3754856586456299,
"kl": 0.1189422607421875,
"learning_rate": 3.9904679361238526e-07,
"loss": 0.0047,
"reward": 0.27103549893945456,
"reward_std": 0.5563001856207848,
"rewards/cosine_scaled_reward": -0.23948227241635323,
"rewards/format_reward": 0.7500000149011612,
"step": 324
},
{
"completion_length": 953.395881652832,
"epoch": 0.37142857142857144,
"grad_norm": 0.36472955346107483,
"kl": 0.0133514404296875,
"learning_rate": 3.9609093550344907e-07,
"loss": 0.0005,
"reward": 1.0393343269824982,
"reward_std": 0.7779577225446701,
"rewards/cosine_scaled_reward": 0.01966716069728136,
"rewards/format_reward": 1.0,
"step": 325
},
{
"completion_length": 1773.9167175292969,
"epoch": 0.37257142857142855,
"grad_norm": 1.3614015579223633,
"kl": 0.09857177734375,
"learning_rate": 3.931425787051832e-07,
"loss": 0.0039,
"reward": 0.4826292358338833,
"reward_std": 0.624052882194519,
"rewards/cosine_scaled_reward": -0.1753520662896335,
"rewards/format_reward": 0.833333358168602,
"step": 326
},
{
"completion_length": 1750.354232788086,
"epoch": 0.3737142857142857,
"grad_norm": 1.2214010953903198,
"kl": 0.092681884765625,
"learning_rate": 3.902018669163384e-07,
"loss": 0.0037,
"reward": 0.6829137187451124,
"reward_std": 0.5776621401309967,
"rewards/cosine_scaled_reward": -0.05437648296356201,
"rewards/format_reward": 0.7916666865348816,
"step": 327
},
{
"completion_length": 1816.6250610351562,
"epoch": 0.37485714285714283,
"grad_norm": 0.6874251961708069,
"kl": 0.08489990234375,
"learning_rate": 3.872689434630585e-07,
"loss": 0.0034,
"reward": 0.796313688158989,
"reward_std": 1.0243088752031326,
"rewards/cosine_scaled_reward": -0.008093174546957016,
"rewards/format_reward": 0.8125000298023224,
"step": 328
},
{
"completion_length": 2015.5000305175781,
"epoch": 0.376,
"grad_norm": 1.2615182399749756,
"kl": 0.15081787109375,
"learning_rate": 3.843439512918949e-07,
"loss": 0.006,
"reward": 0.6527662584558129,
"reward_std": 0.7077807486057281,
"rewards/cosine_scaled_reward": -0.09028353914618492,
"rewards/format_reward": 0.833333358168602,
"step": 329
},
{
"completion_length": 2259.291717529297,
"epoch": 0.37714285714285717,
"grad_norm": 1.1345741748809814,
"kl": 0.1787109375,
"learning_rate": 3.8142703296283953e-07,
"loss": 0.0071,
"reward": 0.5901373848319054,
"reward_std": 0.9298846423625946,
"rewards/cosine_scaled_reward": -0.03826466016471386,
"rewards/format_reward": 0.6666666865348816,
"step": 330
},
{
"completion_length": 2008.9584045410156,
"epoch": 0.3782857142857143,
"grad_norm": 1.42428719997406,
"kl": 0.228515625,
"learning_rate": 3.785183306423767e-07,
"loss": 0.0092,
"reward": 0.315967773552984,
"reward_std": 0.7076915055513382,
"rewards/cosine_scaled_reward": -0.19618277810513973,
"rewards/format_reward": 0.7083333432674408,
"step": 331
},
{
"completion_length": 1344.7917175292969,
"epoch": 0.37942857142857145,
"grad_norm": 1.2130067348480225,
"kl": 0.17629241943359375,
"learning_rate": 3.7561798609655373e-07,
"loss": 0.007,
"reward": 1.3930619359016418,
"reward_std": 0.7922802865505219,
"rewards/cosine_scaled_reward": 0.26944761723279953,
"rewards/format_reward": 0.8541666716337204,
"step": 332
},
{
"completion_length": 1645.3750305175781,
"epoch": 0.38057142857142856,
"grad_norm": 1.1035624742507935,
"kl": 0.212158203125,
"learning_rate": 3.72726140684072e-07,
"loss": 0.0085,
"reward": 0.28750851564109325,
"reward_std": 0.3911990597844124,
"rewards/cosine_scaled_reward": -0.252079077064991,
"rewards/format_reward": 0.7916666865348816,
"step": 333
},
{
"completion_length": 1476.8333740234375,
"epoch": 0.38171428571428573,
"grad_norm": 1.2739105224609375,
"kl": 0.1046142578125,
"learning_rate": 3.6984293534939737e-07,
"loss": 0.0042,
"reward": 0.836659163236618,
"reward_std": 0.8671058118343353,
"rewards/cosine_scaled_reward": -0.04000374022871256,
"rewards/format_reward": 0.9166666865348816,
"step": 334
},
{
"completion_length": 1723.3333892822266,
"epoch": 0.38285714285714284,
"grad_norm": 1.672484040260315,
"kl": 0.21124267578125,
"learning_rate": 3.6696851061588994e-07,
"loss": 0.0084,
"reward": 0.5473275234689936,
"reward_std": 0.6023431569337845,
"rewards/cosine_scaled_reward": -0.11175291612744331,
"rewards/format_reward": 0.770833358168602,
"step": 335
},
{
"completion_length": 1291.8542175292969,
"epoch": 0.384,
"grad_norm": 1.1949892044067383,
"kl": 0.13958740234375,
"learning_rate": 3.641030065789562e-07,
"loss": 0.0056,
"reward": 0.43108636140823364,
"reward_std": 0.42728982865810394,
"rewards/cosine_scaled_reward": -0.22195683978497982,
"rewards/format_reward": 0.8750000149011612,
"step": 336
},
{
"completion_length": 1024.3333740234375,
"epoch": 0.3851428571428571,
"grad_norm": 0.902114748954773,
"kl": 0.11954498291015625,
"learning_rate": 3.612465628992203e-07,
"loss": 0.0048,
"reward": 1.0072421729564667,
"reward_std": 0.6713532879948616,
"rewards/cosine_scaled_reward": 0.045287732034921646,
"rewards/format_reward": 0.9166666865348816,
"step": 337
},
{
"completion_length": 1448.9375457763672,
"epoch": 0.3862857142857143,
"grad_norm": 1.2535922527313232,
"kl": 0.1561737060546875,
"learning_rate": 3.5839931879571725e-07,
"loss": 0.0062,
"reward": 0.5455589033663273,
"reward_std": 0.535648949444294,
"rewards/cosine_scaled_reward": -0.13347055204212666,
"rewards/format_reward": 0.8125000149011612,
"step": 338
},
{
"completion_length": 2035.6459045410156,
"epoch": 0.38742857142857146,
"grad_norm": 0.746334969997406,
"kl": 0.178955078125,
"learning_rate": 3.555614130391079e-07,
"loss": 0.0071,
"reward": 0.593096449971199,
"reward_std": 0.948510006070137,
"rewards/cosine_scaled_reward": -0.08886844478547573,
"rewards/format_reward": 0.770833358168602,
"step": 339
},
{
"completion_length": 1074.1041870117188,
"epoch": 0.38857142857142857,
"grad_norm": 0.5353119373321533,
"kl": 0.041778564453125,
"learning_rate": 3.5273298394491515e-07,
"loss": 0.0017,
"reward": 1.1473649591207504,
"reward_std": 0.7414836436510086,
"rewards/cosine_scaled_reward": 0.07368248514831066,
"rewards/format_reward": 1.0,
"step": 340
},
{
"completion_length": 1197.7083892822266,
"epoch": 0.38971428571428574,
"grad_norm": 1.1976772546768188,
"kl": 0.1300048828125,
"learning_rate": 3.4991416936678276e-07,
"loss": 0.0052,
"reward": 1.3422877192497253,
"reward_std": 0.6801351606845856,
"rewards/cosine_scaled_reward": 0.1919771609827876,
"rewards/format_reward": 0.9583333432674408,
"step": 341
},
{
"completion_length": 1645.479248046875,
"epoch": 0.39085714285714285,
"grad_norm": 1.725953459739685,
"kl": 0.33575439453125,
"learning_rate": 3.471051066897562e-07,
"loss": 0.0135,
"reward": 0.45842229574918747,
"reward_std": 0.5379992946982384,
"rewards/cosine_scaled_reward": -0.1457888763397932,
"rewards/format_reward": 0.7500000149011612,
"step": 342
},
{
"completion_length": 1599.0416870117188,
"epoch": 0.392,
"grad_norm": 1.9615150690078735,
"kl": 0.2664794921875,
"learning_rate": 3.4430593282358777e-07,
"loss": 0.0107,
"reward": 0.8948207944631577,
"reward_std": 0.6805202513933182,
"rewards/cosine_scaled_reward": 0.009910388849675655,
"rewards/format_reward": 0.8750000149011612,
"step": 343
},
{
"completion_length": 1742.8750610351562,
"epoch": 0.3931428571428571,
"grad_norm": 41.373931884765625,
"kl": 0.8536376953125,
"learning_rate": 3.4151678419606233e-07,
"loss": 0.0343,
"reward": 0.8157278522849083,
"reward_std": 0.7099937200546265,
"rewards/cosine_scaled_reward": -0.029636098071932793,
"rewards/format_reward": 0.8750000149011612,
"step": 344
},
{
"completion_length": 1436.8125610351562,
"epoch": 0.3942857142857143,
"grad_norm": 1.8816604614257812,
"kl": 0.26708984375,
"learning_rate": 3.387377967463493e-07,
"loss": 0.0107,
"reward": 0.9264494627714157,
"reward_std": 0.6063774973154068,
"rewards/cosine_scaled_reward": 0.015308059751987457,
"rewards/format_reward": 0.8958333432674408,
"step": 345
},
{
"completion_length": 1363.9792022705078,
"epoch": 0.3954285714285714,
"grad_norm": 2.3164970874786377,
"kl": 0.09549713134765625,
"learning_rate": 3.359691059183761e-07,
"loss": 0.0038,
"reward": 1.0577785670757294,
"reward_std": 0.8408964425325394,
"rewards/cosine_scaled_reward": 0.07055594399571419,
"rewards/format_reward": 0.9166666716337204,
"step": 346
},
{
"completion_length": 1337.7083892822266,
"epoch": 0.3965714285714286,
"grad_norm": 1.1808565855026245,
"kl": 0.2160491943359375,
"learning_rate": 3.3321084665422803e-07,
"loss": 0.0086,
"reward": 1.5314601063728333,
"reward_std": 0.8420404642820358,
"rewards/cosine_scaled_reward": 0.31781339878216386,
"rewards/format_reward": 0.895833358168602,
"step": 347
},
{
"completion_length": 1250.9791870117188,
"epoch": 0.3977142857142857,
"grad_norm": 1.317704439163208,
"kl": 0.178863525390625,
"learning_rate": 3.3046315338757026e-07,
"loss": 0.0072,
"reward": 0.746107667684555,
"reward_std": 0.5483127310872078,
"rewards/cosine_scaled_reward": -0.10611284070182592,
"rewards/format_reward": 0.9583333432674408,
"step": 348
},
{
"completion_length": 1193.000015258789,
"epoch": 0.39885714285714285,
"grad_norm": 1.4739476442337036,
"kl": 0.249725341796875,
"learning_rate": 3.2772616003709616e-07,
"loss": 0.01,
"reward": 1.2662739604711533,
"reward_std": 0.7532782405614853,
"rewards/cosine_scaled_reward": 0.18522025644779205,
"rewards/format_reward": 0.895833358168602,
"step": 349
},
{
"completion_length": 1177.3541870117188,
"epoch": 0.4,
"grad_norm": 2.2546327114105225,
"kl": 0.24993896484375,
"learning_rate": 3.250000000000001e-07,
"loss": 0.01,
"reward": 1.0857526510953903,
"reward_std": 0.6757391728460789,
"rewards/cosine_scaled_reward": 0.09495963307563215,
"rewards/format_reward": 0.8958333432674408,
"step": 350
},
{
"completion_length": 1196.0000457763672,
"epoch": 0.40114285714285713,
"grad_norm": 1.1820554733276367,
"kl": 0.15570068359375,
"learning_rate": 3.222848061454764e-07,
"loss": 0.0062,
"reward": 0.5796034894883633,
"reward_std": 0.5287084579467773,
"rewards/cosine_scaled_reward": -0.178948275744915,
"rewards/format_reward": 0.9375000149011612,
"step": 351
},
{
"completion_length": 1270.5416870117188,
"epoch": 0.4022857142857143,
"grad_norm": 1.5620218515396118,
"kl": 0.12603759765625,
"learning_rate": 3.195807108082429e-07,
"loss": 0.005,
"reward": 0.7529177367687225,
"reward_std": 0.6913501024246216,
"rewards/cosine_scaled_reward": -0.09229113161563873,
"rewards/format_reward": 0.9375000149011612,
"step": 352
},
{
"completion_length": 1443.1666870117188,
"epoch": 0.4034285714285714,
"grad_norm": 2.1429245471954346,
"kl": 0.18780517578125,
"learning_rate": 3.168878457820915e-07,
"loss": 0.0075,
"reward": 0.9008820652961731,
"reward_std": 0.6463882178068161,
"rewards/cosine_scaled_reward": -0.007892303168773651,
"rewards/format_reward": 0.9166666716337204,
"step": 353
},
{
"completion_length": 1230.6458587646484,
"epoch": 0.4045714285714286,
"grad_norm": 0.9503077268600464,
"kl": 0.2082061767578125,
"learning_rate": 3.142063423134644e-07,
"loss": 0.0083,
"reward": 0.4887809455394745,
"reward_std": 0.4611463025212288,
"rewards/cosine_scaled_reward": -0.23477619886398315,
"rewards/format_reward": 0.9583333432674408,
"step": 354
},
{
"completion_length": 930.0625305175781,
"epoch": 0.4057142857142857,
"grad_norm": 0.8462334275245667,
"kl": 0.1320648193359375,
"learning_rate": 3.115363310950578e-07,
"loss": 0.0053,
"reward": 0.8714643996208906,
"reward_std": 0.35595114156603813,
"rewards/cosine_scaled_reward": -0.022601131349802017,
"rewards/format_reward": 0.9166666716337204,
"step": 355
},
{
"completion_length": 1521.2500610351562,
"epoch": 0.40685714285714286,
"grad_norm": 3.6062703132629395,
"kl": 0.31671142578125,
"learning_rate": 3.0887794225945143e-07,
"loss": 0.0126,
"reward": 0.7512324824929237,
"reward_std": 0.7443573772907257,
"rewards/cosine_scaled_reward": -0.05146709643304348,
"rewards/format_reward": 0.8541666865348816,
"step": 356
},
{
"completion_length": 1307.0625305175781,
"epoch": 0.408,
"grad_norm": 2.073002576828003,
"kl": 0.2852783203125,
"learning_rate": 3.062313053727671e-07,
"loss": 0.0114,
"reward": 0.6376607120037079,
"reward_std": 0.5459994077682495,
"rewards/cosine_scaled_reward": -0.13950299471616745,
"rewards/format_reward": 0.9166666865348816,
"step": 357
},
{
"completion_length": 1298.2083587646484,
"epoch": 0.40914285714285714,
"grad_norm": 1.5467488765716553,
"kl": 0.34130859375,
"learning_rate": 3.0359654942835247e-07,
"loss": 0.0136,
"reward": 0.881379060447216,
"reward_std": 0.8507445156574249,
"rewards/cosine_scaled_reward": 0.003189507406204939,
"rewards/format_reward": 0.875,
"step": 358
},
{
"completion_length": 1172.7916870117188,
"epoch": 0.4102857142857143,
"grad_norm": 1.671886682510376,
"kl": 0.314453125,
"learning_rate": 3.0097380284049523e-07,
"loss": 0.0126,
"reward": 0.7981488406658173,
"reward_std": 0.5063908323645592,
"rewards/cosine_scaled_reward": -0.04884226247668266,
"rewards/format_reward": 0.895833358168602,
"step": 359
},
{
"completion_length": 990.9792175292969,
"epoch": 0.4114285714285714,
"grad_norm": 2.374612808227539,
"kl": 0.41845703125,
"learning_rate": 2.9836319343816397e-07,
"loss": 0.0168,
"reward": 1.1945213824510574,
"reward_std": 0.7698078602552414,
"rewards/cosine_scaled_reward": 0.11809402331709862,
"rewards/format_reward": 0.9583333432674408,
"step": 360
},
{
"completion_length": 1524.3541870117188,
"epoch": 0.4125714285714286,
"grad_norm": 3.4913723468780518,
"kl": 1.1523895263671875,
"learning_rate": 2.9576484845877793e-07,
"loss": 0.0461,
"reward": 0.6581477224826813,
"reward_std": 0.6756026446819305,
"rewards/cosine_scaled_reward": -0.05634281662059948,
"rewards/format_reward": 0.7708333432674408,
"step": 361
},
{
"completion_length": 1185.2708587646484,
"epoch": 0.4137142857142857,
"grad_norm": 34.620460510253906,
"kl": 0.832366943359375,
"learning_rate": 2.931788945420058e-07,
"loss": 0.0334,
"reward": 0.8447651118040085,
"reward_std": 0.7907865196466446,
"rewards/cosine_scaled_reward": -0.046367482747882605,
"rewards/format_reward": 0.9375000149011612,
"step": 362
},
{
"completion_length": 1141.0208892822266,
"epoch": 0.41485714285714287,
"grad_norm": 2.2585678100585938,
"kl": 0.352935791015625,
"learning_rate": 2.9060545772359305e-07,
"loss": 0.0141,
"reward": 0.5232757963240147,
"reward_std": 0.5396992526948452,
"rewards/cosine_scaled_reward": -0.1966954478994012,
"rewards/format_reward": 0.9166666865348816,
"step": 363
},
{
"completion_length": 1359.1667175292969,
"epoch": 0.416,
"grad_norm": 3.5956501960754395,
"kl": 0.715576171875,
"learning_rate": 2.8804466342921987e-07,
"loss": 0.0286,
"reward": 0.6145341023802757,
"reward_std": 0.5419690161943436,
"rewards/cosine_scaled_reward": -0.17189963907003403,
"rewards/format_reward": 0.9583333432674408,
"step": 364
},
{
"completion_length": 1670.229248046875,
"epoch": 0.41714285714285715,
"grad_norm": 3.5323920249938965,
"kl": 1.1943359375,
"learning_rate": 2.854966364683872e-07,
"loss": 0.0479,
"reward": 0.40414058696478605,
"reward_std": 0.691382423043251,
"rewards/cosine_scaled_reward": -0.17292970418930054,
"rewards/format_reward": 0.7500000074505806,
"step": 365
},
{
"completion_length": 1720.5000305175781,
"epoch": 0.41828571428571426,
"grad_norm": 3.34942364692688,
"kl": 1.042236328125,
"learning_rate": 2.829615010283344e-07,
"loss": 0.0417,
"reward": 0.6996188908815384,
"reward_std": 0.7247656881809235,
"rewards/cosine_scaled_reward": -0.0460239015519619,
"rewards/format_reward": 0.791666679084301,
"step": 366
},
{
"completion_length": 927.6458587646484,
"epoch": 0.41942857142857143,
"grad_norm": 1.461083173751831,
"kl": 0.38287353515625,
"learning_rate": 2.8043938066798645e-07,
"loss": 0.0153,
"reward": 1.2472570352256298,
"reward_std": 0.6670772060751915,
"rewards/cosine_scaled_reward": 0.14446185529232025,
"rewards/format_reward": 0.9583333432674408,
"step": 367
},
{
"completion_length": 1301.9792175292969,
"epoch": 0.4205714285714286,
"grad_norm": 2.861830711364746,
"kl": 0.337615966796875,
"learning_rate": 2.7793039831193133e-07,
"loss": 0.0135,
"reward": 1.2516463994979858,
"reward_std": 1.0124248266220093,
"rewards/cosine_scaled_reward": 0.17790652811527252,
"rewards/format_reward": 0.8958333432674408,
"step": 368
},
{
"completion_length": 1215.5625,
"epoch": 0.4217142857142857,
"grad_norm": 1.6978230476379395,
"kl": 0.335235595703125,
"learning_rate": 2.7543467624442956e-07,
"loss": 0.0134,
"reward": 1.1675912141799927,
"reward_std": 0.5593391507863998,
"rewards/cosine_scaled_reward": 0.13587890937924385,
"rewards/format_reward": 0.8958333432674408,
"step": 369
},
{
"completion_length": 1309.5208740234375,
"epoch": 0.4228571428571429,
"grad_norm": 2.2963132858276367,
"kl": 0.4232177734375,
"learning_rate": 2.729523361034538e-07,
"loss": 0.0169,
"reward": 0.5871782079339027,
"reward_std": 0.778529703617096,
"rewards/cosine_scaled_reward": -0.14391089417040348,
"rewards/format_reward": 0.8750000298023224,
"step": 370
},
{
"completion_length": 1090.1042022705078,
"epoch": 0.424,
"grad_norm": 0.9345715045928955,
"kl": 0.28076171875,
"learning_rate": 2.7048349887476037e-07,
"loss": 0.0112,
"reward": 0.6981654912233353,
"reward_std": 0.6373357623815536,
"rewards/cosine_scaled_reward": -0.09883392881602049,
"rewards/format_reward": 0.8958333432674408,
"step": 371
},
{
"completion_length": 1309.9375610351562,
"epoch": 0.42514285714285716,
"grad_norm": 0.7092093825340271,
"kl": 0.486602783203125,
"learning_rate": 2.6802828488599294e-07,
"loss": 0.0194,
"reward": 0.8811748586595058,
"reward_std": 0.5044156312942505,
"rewards/cosine_scaled_reward": -0.007329270243644714,
"rewards/format_reward": 0.8958333432674408,
"step": 372
},
{
"completion_length": 1460.7708587646484,
"epoch": 0.42628571428571427,
"grad_norm": 2.416278839111328,
"kl": 0.5400390625,
"learning_rate": 2.655868138008171e-07,
"loss": 0.0216,
"reward": 0.9694189727306366,
"reward_std": 0.6003681719303131,
"rewards/cosine_scaled_reward": 0.0472094789147377,
"rewards/format_reward": 0.8750000149011612,
"step": 373
},
{
"completion_length": 1132.5208587646484,
"epoch": 0.42742857142857144,
"grad_norm": 1.0274022817611694,
"kl": 0.29541015625,
"learning_rate": 2.631592046130896e-07,
"loss": 0.0118,
"reward": 1.1347306370735168,
"reward_std": 0.5300607345998287,
"rewards/cosine_scaled_reward": 0.07778198271989822,
"rewards/format_reward": 0.9791666716337204,
"step": 374
},
{
"completion_length": 1205.6042175292969,
"epoch": 0.42857142857142855,
"grad_norm": 1.4747833013534546,
"kl": 0.336669921875,
"learning_rate": 2.6074557564105724e-07,
"loss": 0.0135,
"reward": 0.914167582988739,
"reward_std": 0.7538186013698578,
"rewards/cosine_scaled_reward": -0.011666236445307732,
"rewards/format_reward": 0.9375000149011612,
"step": 375
},
{
"completion_length": 1323.9583740234375,
"epoch": 0.4297142857142857,
"grad_norm": 1.5346781015396118,
"kl": 0.40960693359375,
"learning_rate": 2.583460445215911e-07,
"loss": 0.0164,
"reward": 0.9741204902529716,
"reward_std": 0.8010425865650177,
"rewards/cosine_scaled_reward": 0.018310231156647205,
"rewards/format_reward": 0.9375000149011612,
"step": 376
},
{
"completion_length": 1173.8333892822266,
"epoch": 0.4308571428571429,
"grad_norm": 1.3860251903533936,
"kl": 0.32177734375,
"learning_rate": 2.5596072820445254e-07,
"loss": 0.0129,
"reward": 1.1005370616912842,
"reward_std": 0.8687849044799805,
"rewards/cosine_scaled_reward": 0.11276852712035179,
"rewards/format_reward": 0.8750000149011612,
"step": 377
},
{
"completion_length": 1077.0208587646484,
"epoch": 0.432,
"grad_norm": 1.6222304105758667,
"kl": 0.325439453125,
"learning_rate": 2.5358974294659373e-07,
"loss": 0.013,
"reward": 0.7557893544435501,
"reward_std": 0.46596937626600266,
"rewards/cosine_scaled_reward": -0.12210530787706375,
"rewards/format_reward": 1.0,
"step": 378
},
{
"completion_length": 1240.3333740234375,
"epoch": 0.43314285714285716,
"grad_norm": 1.0828309059143066,
"kl": 0.356201171875,
"learning_rate": 2.512332043064913e-07,
"loss": 0.0142,
"reward": 0.8545220792293549,
"reward_std": 0.6498586684465408,
"rewards/cosine_scaled_reward": -0.05190563574433327,
"rewards/format_reward": 0.9583333432674408,
"step": 379
},
{
"completion_length": 1270.6667175292969,
"epoch": 0.4342857142857143,
"grad_norm": 1.8103889226913452,
"kl": 0.383026123046875,
"learning_rate": 2.488912271385139e-07,
"loss": 0.0153,
"reward": 0.6777837425470352,
"reward_std": 0.7819753885269165,
"rewards/cosine_scaled_reward": -0.08819146640598774,
"rewards/format_reward": 0.8541666716337204,
"step": 380
},
{
"completion_length": 1318.7292175292969,
"epoch": 0.43542857142857144,
"grad_norm": 1.7676132917404175,
"kl": 0.482666015625,
"learning_rate": 2.465639255873246e-07,
"loss": 0.0193,
"reward": 0.7255931571125984,
"reward_std": 0.6526116281747818,
"rewards/cosine_scaled_reward": -0.10595342982560396,
"rewards/format_reward": 0.9375000149011612,
"step": 381
},
{
"completion_length": 1141.7917022705078,
"epoch": 0.43657142857142855,
"grad_norm": 1.832429051399231,
"kl": 0.3201904296875,
"learning_rate": 2.4425141308231765e-07,
"loss": 0.0128,
"reward": 0.7391386441886425,
"reward_std": 0.7291494160890579,
"rewards/cosine_scaled_reward": -0.07834736630320549,
"rewards/format_reward": 0.8958333432674408,
"step": 382
},
{
"completion_length": 1615.541732788086,
"epoch": 0.4377142857142857,
"grad_norm": 2.4247608184814453,
"kl": 1.1395721435546875,
"learning_rate": 2.4195380233209006e-07,
"loss": 0.0456,
"reward": 0.8856848031282425,
"reward_std": 0.5492001250386238,
"rewards/cosine_scaled_reward": 0.005342394113540649,
"rewards/format_reward": 0.8750000149011612,
"step": 383
},
{
"completion_length": 1421.0833435058594,
"epoch": 0.43885714285714283,
"grad_norm": 1.442662000656128,
"kl": 0.374847412109375,
"learning_rate": 2.3967120531894857e-07,
"loss": 0.015,
"reward": 0.32889158837497234,
"reward_std": 0.4332951605319977,
"rewards/cosine_scaled_reward": -0.26263754442334175,
"rewards/format_reward": 0.8541666865348816,
"step": 384
},
{
"completion_length": 1334.5208435058594,
"epoch": 0.44,
"grad_norm": 1.4730134010314941,
"kl": 0.422088623046875,
"learning_rate": 2.374037332934512e-07,
"loss": 0.0169,
"reward": 0.8442478328943253,
"reward_std": 0.6724153012037277,
"rewards/cosine_scaled_reward": -0.05704276263713837,
"rewards/format_reward": 0.9583333432674408,
"step": 385
},
{
"completion_length": 1111.4583740234375,
"epoch": 0.44114285714285717,
"grad_norm": 1.538129448890686,
"kl": 0.181640625,
"learning_rate": 2.3515149676898552e-07,
"loss": 0.0073,
"reward": 0.8161994330585003,
"reward_std": 0.4277452155947685,
"rewards/cosine_scaled_reward": -0.06065032631158829,
"rewards/format_reward": 0.9375000149011612,
"step": 386
},
{
"completion_length": 1218.0208435058594,
"epoch": 0.4422857142857143,
"grad_norm": 1.626770257949829,
"kl": 0.309906005859375,
"learning_rate": 2.3291460551638237e-07,
"loss": 0.0124,
"reward": 0.6058969795703888,
"reward_std": 0.7323561906814575,
"rewards/cosine_scaled_reward": -0.1658015362918377,
"rewards/format_reward": 0.9375000149011612,
"step": 387
},
{
"completion_length": 1258.187515258789,
"epoch": 0.44342857142857145,
"grad_norm": 1.7491554021835327,
"kl": 0.3914794921875,
"learning_rate": 2.306931685585657e-07,
"loss": 0.0156,
"reward": 0.7627854868769646,
"reward_std": 0.7414772808551788,
"rewards/cosine_scaled_reward": -0.05610728543251753,
"rewards/format_reward": 0.8750000149011612,
"step": 388
},
{
"completion_length": 1268.4375305175781,
"epoch": 0.44457142857142856,
"grad_norm": 1.6758959293365479,
"kl": 0.440826416015625,
"learning_rate": 2.2848729416523859e-07,
"loss": 0.0176,
"reward": 1.1525626480579376,
"reward_std": 0.7809968441724777,
"rewards/cosine_scaled_reward": 0.12836465798318386,
"rewards/format_reward": 0.895833358168602,
"step": 389
},
{
"completion_length": 940.9166870117188,
"epoch": 0.44571428571428573,
"grad_norm": 1.7198045253753662,
"kl": 0.3330078125,
"learning_rate": 2.2629708984760706e-07,
"loss": 0.0133,
"reward": 0.9264688044786453,
"reward_std": 0.6931325197219849,
"rewards/cosine_scaled_reward": -0.03676560753956437,
"rewards/format_reward": 1.0,
"step": 390
},
{
"completion_length": 978.7291717529297,
"epoch": 0.44685714285714284,
"grad_norm": 0.6671801805496216,
"kl": 0.029083251953125,
"learning_rate": 2.2412266235313973e-07,
"loss": 0.0012,
"reward": 0.6134733706712723,
"reward_std": 0.4728550612926483,
"rewards/cosine_scaled_reward": -0.1932633202522993,
"rewards/format_reward": 1.0,
"step": 391
},
{
"completion_length": 1309.5417175292969,
"epoch": 0.448,
"grad_norm": 2.3814332485198975,
"kl": 0.251953125,
"learning_rate": 2.2196411766036487e-07,
"loss": 0.0101,
"reward": 1.0050799548625946,
"reward_std": 0.8785246461629868,
"rewards/cosine_scaled_reward": 0.07545664254575968,
"rewards/format_reward": 0.8541666865348816,
"step": 392
},
{
"completion_length": 1097.3125305175781,
"epoch": 0.4491428571428571,
"grad_norm": 1.722952961921692,
"kl": 0.2423095703125,
"learning_rate": 2.1982156097370557e-07,
"loss": 0.0097,
"reward": 0.832743689417839,
"reward_std": 0.8567025661468506,
"rewards/cosine_scaled_reward": -0.0419615093851462,
"rewards/format_reward": 0.9166666865348816,
"step": 393
},
{
"completion_length": 1223.0833740234375,
"epoch": 0.4502857142857143,
"grad_norm": 2.990572452545166,
"kl": 0.4295654296875,
"learning_rate": 2.1769509671835223e-07,
"loss": 0.0172,
"reward": 0.5410304628312588,
"reward_std": 0.6248824968934059,
"rewards/cosine_scaled_reward": -0.13573478162288666,
"rewards/format_reward": 0.8125000149011612,
"step": 394
},
{
"completion_length": 1189.500015258789,
"epoch": 0.4514285714285714,
"grad_norm": 2.629859685897827,
"kl": 0.321044921875,
"learning_rate": 2.1558482853517253e-07,
"loss": 0.0128,
"reward": 0.5656278505921364,
"reward_std": 0.7980157136917114,
"rewards/cosine_scaled_reward": -0.1130194254219532,
"rewards/format_reward": 0.7916667014360428,
"step": 395
},
{
"completion_length": 1043.1458435058594,
"epoch": 0.45257142857142857,
"grad_norm": 1.0702786445617676,
"kl": 0.0563507080078125,
"learning_rate": 2.134908592756607e-07,
"loss": 0.0023,
"reward": 0.7221028283238411,
"reward_std": 0.5050450935959816,
"rewards/cosine_scaled_reward": -0.11811527609825134,
"rewards/format_reward": 0.9583333432674408,
"step": 396
},
{
"completion_length": 1230.7292175292969,
"epoch": 0.45371428571428574,
"grad_norm": 3.2597548961639404,
"kl": 0.392822265625,
"learning_rate": 2.1141329099692406e-07,
"loss": 0.0157,
"reward": 1.312897451221943,
"reward_std": 0.6480658948421478,
"rewards/cosine_scaled_reward": 0.2293653730303049,
"rewards/format_reward": 0.8541667014360428,
"step": 397
},
{
"completion_length": 1381.5625305175781,
"epoch": 0.45485714285714285,
"grad_norm": 1.6307339668273926,
"kl": 0.20714569091796875,
"learning_rate": 2.0935222495670968e-07,
"loss": 0.0083,
"reward": 0.7576578855514526,
"reward_std": 0.7408057749271393,
"rewards/cosine_scaled_reward": -0.10033774503972381,
"rewards/format_reward": 0.9583333432674408,
"step": 398
},
{
"completion_length": 1487.0208740234375,
"epoch": 0.456,
"grad_norm": 2.9457855224609375,
"kl": 0.426025390625,
"learning_rate": 2.0730776160846853e-07,
"loss": 0.0171,
"reward": 0.6220748424530029,
"reward_std": 0.5146789476275444,
"rewards/cosine_scaled_reward": -0.09521258249878883,
"rewards/format_reward": 0.8125000149011612,
"step": 399
},
{
"completion_length": 1041.7708587646484,
"epoch": 0.45714285714285713,
"grad_norm": 5.961241722106934,
"kl": 0.289337158203125,
"learning_rate": 2.0528000059645995e-07,
"loss": 0.0116,
"reward": 0.8267193324863911,
"reward_std": 0.6178570240736008,
"rewards/cosine_scaled_reward": -0.04497369006276131,
"rewards/format_reward": 0.9166666716337204,
"step": 400
},
{
"completion_length": 965.6250305175781,
"epoch": 0.4582857142857143,
"grad_norm": 1.4047675132751465,
"kl": 0.1659698486328125,
"learning_rate": 2.032690407508949e-07,
"loss": 0.0066,
"reward": 0.982437789440155,
"reward_std": 0.7459337636828423,
"rewards/cosine_scaled_reward": 0.0537188770249486,
"rewards/format_reward": 0.8750000149011612,
"step": 401
},
{
"completion_length": 1609.729248046875,
"epoch": 0.4594285714285714,
"grad_norm": 2.311453342437744,
"kl": 0.615234375,
"learning_rate": 2.0127498008311922e-07,
"loss": 0.0246,
"reward": 0.3766894303262234,
"reward_std": 0.658165842294693,
"rewards/cosine_scaled_reward": -0.2387386392802,
"rewards/format_reward": 0.8541666865348816,
"step": 402
},
{
"completion_length": 1466.9375610351562,
"epoch": 0.4605714285714286,
"grad_norm": 2.0651133060455322,
"kl": 0.7806396484375,
"learning_rate": 1.9929791578083655e-07,
"loss": 0.0312,
"reward": 0.5563420876860619,
"reward_std": 0.6794729232788086,
"rewards/cosine_scaled_reward": -0.1489123017527163,
"rewards/format_reward": 0.8541666865348816,
"step": 403
},
{
"completion_length": 1289.2708740234375,
"epoch": 0.4617142857142857,
"grad_norm": 2.5164127349853516,
"kl": 0.3046875,
"learning_rate": 1.9733794420337213e-07,
"loss": 0.0122,
"reward": 0.8672967702150345,
"reward_std": 0.7338433116674423,
"rewards/cosine_scaled_reward": -0.03510164050385356,
"rewards/format_reward": 0.9375000149011612,
"step": 404
},
{
"completion_length": 1246.5833587646484,
"epoch": 0.46285714285714286,
"grad_norm": 1.7397247552871704,
"kl": 0.14056396484375,
"learning_rate": 1.9539516087697517e-07,
"loss": 0.0056,
"reward": 0.7972104996442795,
"reward_std": 0.6757538244128227,
"rewards/cosine_scaled_reward": 0.0027718953788280487,
"rewards/format_reward": 0.7916667014360428,
"step": 405
},
{
"completion_length": 1569.5209045410156,
"epoch": 0.464,
"grad_norm": 3.1697490215301514,
"kl": 1.05712890625,
"learning_rate": 1.934696604901642e-07,
"loss": 0.0423,
"reward": 0.7342821173369884,
"reward_std": 0.6933150887489319,
"rewards/cosine_scaled_reward": -0.05994228646159172,
"rewards/format_reward": 0.8541666716337204,
"step": 406
},
{
"completion_length": 970.0625457763672,
"epoch": 0.46514285714285714,
"grad_norm": 2.7586686611175537,
"kl": 0.496337890625,
"learning_rate": 1.915615368891117e-07,
"loss": 0.0198,
"reward": 0.7943353094160557,
"reward_std": 0.5273813158273697,
"rewards/cosine_scaled_reward": -0.07158234342932701,
"rewards/format_reward": 0.9375000149011612,
"step": 407
},
{
"completion_length": 954.7083435058594,
"epoch": 0.4662857142857143,
"grad_norm": 1.7943812608718872,
"kl": 0.168701171875,
"learning_rate": 1.8967088307307e-07,
"loss": 0.0067,
"reward": 0.9154380261898041,
"reward_std": 0.63597172498703,
"rewards/cosine_scaled_reward": -0.02144765853881836,
"rewards/format_reward": 0.9583333432674408,
"step": 408
},
{
"completion_length": 1268.2291717529297,
"epoch": 0.4674285714285714,
"grad_norm": 0.9967200756072998,
"kl": 0.53662109375,
"learning_rate": 1.8779779118983867e-07,
"loss": 0.0215,
"reward": 0.9619596749544144,
"reward_std": 0.6098195463418961,
"rewards/cosine_scaled_reward": 0.022646483033895493,
"rewards/format_reward": 0.9166666865348816,
"step": 409
},
{
"completion_length": 1628.0000610351562,
"epoch": 0.4685714285714286,
"grad_norm": 5.845465660095215,
"kl": 0.8017578125,
"learning_rate": 1.8594235253127372e-07,
"loss": 0.0321,
"reward": 0.6380558162927628,
"reward_std": 0.5308569446206093,
"rewards/cosine_scaled_reward": -0.14972211251733825,
"rewards/format_reward": 0.9375000149011612,
"step": 410
},
{
"completion_length": 1665.7083740234375,
"epoch": 0.4697142857142857,
"grad_norm": 1.869497537612915,
"kl": 0.796875,
"learning_rate": 1.8410465752883758e-07,
"loss": 0.0319,
"reward": 0.31100673973560333,
"reward_std": 0.5528412610292435,
"rewards/cosine_scaled_reward": -0.27157998457551,
"rewards/format_reward": 0.8541666865348816,
"step": 411
},
{
"completion_length": 1114.604232788086,
"epoch": 0.47085714285714286,
"grad_norm": 1.987026333808899,
"kl": 0.284912109375,
"learning_rate": 1.822847957491922e-07,
"loss": 0.0114,
"reward": 0.7781837359070778,
"reward_std": 0.5904239565134048,
"rewards/cosine_scaled_reward": -0.09007478877902031,
"rewards/format_reward": 0.9583333432674408,
"step": 412
},
{
"completion_length": 1371.5417175292969,
"epoch": 0.472,
"grad_norm": 2.023742198944092,
"kl": 0.481689453125,
"learning_rate": 1.804828558898332e-07,
"loss": 0.0193,
"reward": 0.7588743381202221,
"reward_std": 0.7781134992837906,
"rewards/cosine_scaled_reward": -0.06847950583323836,
"rewards/format_reward": 0.8958333432674408,
"step": 413
},
{
"completion_length": 1710.7292175292969,
"epoch": 0.47314285714285714,
"grad_norm": 3.7236738204956055,
"kl": 1.0467529296875,
"learning_rate": 1.7869892577476722e-07,
"loss": 0.042,
"reward": 0.6331272795796394,
"reward_std": 0.7422880977392197,
"rewards/cosine_scaled_reward": -0.06885303813032806,
"rewards/format_reward": 0.770833358168602,
"step": 414
},
{
"completion_length": 1409.2708740234375,
"epoch": 0.4742857142857143,
"grad_norm": 2.454448938369751,
"kl": 0.506103515625,
"learning_rate": 1.7693309235023127e-07,
"loss": 0.0202,
"reward": 0.9847299754619598,
"reward_std": 0.6669813543558121,
"rewards/cosine_scaled_reward": 0.023614969104528427,
"rewards/format_reward": 0.9375000149011612,
"step": 415
},
{
"completion_length": 1729.4791870117188,
"epoch": 0.4754285714285714,
"grad_norm": 2.803572416305542,
"kl": 0.96875,
"learning_rate": 1.7518544168045524e-07,
"loss": 0.0387,
"reward": 0.4970005638897419,
"reward_std": 0.8149459362030029,
"rewards/cosine_scaled_reward": -0.10566640645265579,
"rewards/format_reward": 0.7083333432674408,
"step": 416
},
{
"completion_length": 1810.1875457763672,
"epoch": 0.4765714285714286,
"grad_norm": 5.9517717361450195,
"kl": 1.576904296875,
"learning_rate": 1.7345605894346726e-07,
"loss": 0.0632,
"reward": 0.5241810567677021,
"reward_std": 0.64292823523283,
"rewards/cosine_scaled_reward": -0.08165947627276182,
"rewards/format_reward": 0.6875000074505806,
"step": 417
},
{
"completion_length": 1492.8958740234375,
"epoch": 0.4777142857142857,
"grad_norm": 4.798679828643799,
"kl": 0.8544921875,
"learning_rate": 1.7174502842694212e-07,
"loss": 0.0342,
"reward": 0.4330623224377632,
"reward_std": 0.5981302931904793,
"rewards/cosine_scaled_reward": -0.22096885181963444,
"rewards/format_reward": 0.8750000149011612,
"step": 418
},
{
"completion_length": 1555.2708435058594,
"epoch": 0.47885714285714287,
"grad_norm": 4.279539585113525,
"kl": 0.86328125,
"learning_rate": 1.7005243352409333e-07,
"loss": 0.0346,
"reward": 0.5117529258131981,
"reward_std": 0.6584911718964577,
"rewards/cosine_scaled_reward": -0.1399568784981966,
"rewards/format_reward": 0.7916666865348816,
"step": 419
},
{
"completion_length": 1216.3542175292969,
"epoch": 0.48,
"grad_norm": 2.101778030395508,
"kl": 0.41534423828125,
"learning_rate": 1.6837835672960831e-07,
"loss": 0.0167,
"reward": 0.896861981600523,
"reward_std": 0.856042891740799,
"rewards/cosine_scaled_reward": 0.0005143135786056519,
"rewards/format_reward": 0.8958333432674408,
"step": 420
},
{
"completion_length": 1162.3750457763672,
"epoch": 0.48114285714285715,
"grad_norm": 2.8854591846466064,
"kl": 0.33807373046875,
"learning_rate": 1.6672287963562852e-07,
"loss": 0.0135,
"reward": 0.8688920065760612,
"reward_std": 0.7225290387868881,
"rewards/cosine_scaled_reward": -0.013470660895109177,
"rewards/format_reward": 0.895833358168602,
"step": 421
},
{
"completion_length": 1370.8541870117188,
"epoch": 0.48228571428571426,
"grad_norm": 1.7124054431915283,
"kl": 0.51220703125,
"learning_rate": 1.6508608292777203e-07,
"loss": 0.0205,
"reward": 0.7653668001294136,
"reward_std": 0.7626539617776871,
"rewards/cosine_scaled_reward": -0.07564992923289537,
"rewards/format_reward": 0.9166666865348816,
"step": 422
},
{
"completion_length": 1183.708396911621,
"epoch": 0.48342857142857143,
"grad_norm": 1.421221137046814,
"kl": 0.293243408203125,
"learning_rate": 1.6346804638120098e-07,
"loss": 0.0117,
"reward": 0.9164831042289734,
"reward_std": 0.533292543143034,
"rewards/cosine_scaled_reward": -0.01050846092402935,
"rewards/format_reward": 0.9375000149011612,
"step": 423
},
{
"completion_length": 1333.6250305175781,
"epoch": 0.4845714285714286,
"grad_norm": 2.020719289779663,
"kl": 0.547119140625,
"learning_rate": 1.6186884885673413e-07,
"loss": 0.0219,
"reward": 0.5903327092528343,
"reward_std": 0.5696832239627838,
"rewards/cosine_scaled_reward": -0.11108366213738918,
"rewards/format_reward": 0.8125000298023224,
"step": 424
},
{
"completion_length": 1426.3542175292969,
"epoch": 0.4857142857142857,
"grad_norm": 2.2707631587982178,
"kl": 0.79052734375,
"learning_rate": 1.6028856829700258e-07,
"loss": 0.0316,
"reward": 0.5219599902629852,
"reward_std": 0.5760443955659866,
"rewards/cosine_scaled_reward": -0.14527002349495888,
"rewards/format_reward": 0.8125000149011612,
"step": 425
},
{
"completion_length": 1259.3333740234375,
"epoch": 0.4868571428571429,
"grad_norm": 2.267915725708008,
"kl": 0.395263671875,
"learning_rate": 1.5872728172265146e-07,
"loss": 0.0158,
"reward": 0.8012158274650574,
"reward_std": 0.667873740196228,
"rewards/cosine_scaled_reward": -0.06814211048185825,
"rewards/format_reward": 0.9375000149011612,
"step": 426
},
{
"completion_length": 1615.0416870117188,
"epoch": 0.488,
"grad_norm": 2.34485125541687,
"kl": 0.85791015625,
"learning_rate": 1.5718506522858572e-07,
"loss": 0.0343,
"reward": 0.7913745269179344,
"reward_std": 0.7925111949443817,
"rewards/cosine_scaled_reward": -0.00014608167111873627,
"rewards/format_reward": 0.7916666865348816,
"step": 427
},
{
"completion_length": 1747.8333740234375,
"epoch": 0.48914285714285716,
"grad_norm": 2.9694297313690186,
"kl": 0.9736328125,
"learning_rate": 1.5566199398026147e-07,
"loss": 0.0389,
"reward": 0.55228191614151,
"reward_std": 0.6319544315338135,
"rewards/cosine_scaled_reward": -0.1509423702955246,
"rewards/format_reward": 0.8541666865348816,
"step": 428
},
{
"completion_length": 1631.5417175292969,
"epoch": 0.49028571428571427,
"grad_norm": 3.960716724395752,
"kl": 1.03509521484375,
"learning_rate": 1.5415814221002265e-07,
"loss": 0.0414,
"reward": 0.8934581205248833,
"reward_std": 0.5868038833141327,
"rewards/cosine_scaled_reward": 0.03006240352988243,
"rewards/format_reward": 0.833333358168602,
"step": 429
},
{
"completion_length": 1621.4375610351562,
"epoch": 0.49142857142857144,
"grad_norm": 8.593777656555176,
"kl": 1.48388671875,
"learning_rate": 1.5267358321348285e-07,
"loss": 0.0594,
"reward": 0.48539859987795353,
"reward_std": 0.7700510919094086,
"rewards/cosine_scaled_reward": -0.12188405683264136,
"rewards/format_reward": 0.7291666716337204,
"step": 430
},
{
"completion_length": 1201.0208587646484,
"epoch": 0.49257142857142855,
"grad_norm": 1.7287312746047974,
"kl": 0.67041015625,
"learning_rate": 1.5120838934595337e-07,
"loss": 0.0268,
"reward": 0.9445644542574883,
"reward_std": 0.5076880529522896,
"rewards/cosine_scaled_reward": 0.0035321786999702454,
"rewards/format_reward": 0.9375000149011612,
"step": 431
},
{
"completion_length": 1444.2084045410156,
"epoch": 0.4937142857142857,
"grad_norm": 3.417004346847534,
"kl": 0.6171875,
"learning_rate": 1.4976263201891613e-07,
"loss": 0.0247,
"reward": 0.6378493383526802,
"reward_std": 0.7811735272407532,
"rewards/cosine_scaled_reward": -0.09774199151434004,
"rewards/format_reward": 0.8333333432674408,
"step": 432
},
{
"completion_length": 1373.5417175292969,
"epoch": 0.4948571428571429,
"grad_norm": 1.7710994482040405,
"kl": 0.712890625,
"learning_rate": 1.483363816965435e-07,
"loss": 0.0286,
"reward": 0.6063527911901474,
"reward_std": 0.6604797914624214,
"rewards/cosine_scaled_reward": -0.144740279763937,
"rewards/format_reward": 0.8958333432674408,
"step": 433
},
{
"completion_length": 977.5208435058594,
"epoch": 0.496,
"grad_norm": 1.736846685409546,
"kl": 0.294708251953125,
"learning_rate": 1.469297078922642e-07,
"loss": 0.0118,
"reward": 1.2148061096668243,
"reward_std": 0.5557400360703468,
"rewards/cosine_scaled_reward": 0.11781970039010048,
"rewards/format_reward": 0.9791666716337204,
"step": 434
},
{
"completion_length": 1433.8542175292969,
"epoch": 0.49714285714285716,
"grad_norm": 2.2994189262390137,
"kl": 1.136810302734375,
"learning_rate": 1.4554267916537495e-07,
"loss": 0.0454,
"reward": 0.4324219524860382,
"reward_std": 0.8103198558092117,
"rewards/cosine_scaled_reward": -0.1900390349328518,
"rewards/format_reward": 0.8125000149011612,
"step": 435
},
{
"completion_length": 1499.3750305175781,
"epoch": 0.4982857142857143,
"grad_norm": 2.9299046993255615,
"kl": 0.6876220703125,
"learning_rate": 1.4417536311769885e-07,
"loss": 0.0275,
"reward": 0.6081612259149551,
"reward_std": 0.8525863587856293,
"rewards/cosine_scaled_reward": -0.1334194028750062,
"rewards/format_reward": 0.8750000149011612,
"step": 436
},
{
"completion_length": 1131.1458892822266,
"epoch": 0.49942857142857144,
"grad_norm": 1.970673680305481,
"kl": 0.24951171875,
"learning_rate": 1.4282782639029128e-07,
"loss": 0.01,
"reward": 0.8882552683353424,
"reward_std": 0.7682344913482666,
"rewards/cosine_scaled_reward": -0.02462236536666751,
"rewards/format_reward": 0.9375000149011612,
"step": 437
},
{
"completion_length": 1025.5208740234375,
"epoch": 0.5005714285714286,
"grad_norm": 1.1531387567520142,
"kl": 0.20703125,
"learning_rate": 1.4150013466019114e-07,
"loss": 0.0083,
"reward": 0.9214688986539841,
"reward_std": 0.6478733271360397,
"rewards/cosine_scaled_reward": -0.03926557023078203,
"rewards/format_reward": 1.0,
"step": 438
},
{
"completion_length": 1004.7708740234375,
"epoch": 0.5017142857142857,
"grad_norm": 1.5415436029434204,
"kl": 0.22320556640625,
"learning_rate": 1.4019235263722034e-07,
"loss": 0.0089,
"reward": 0.6858842521905899,
"reward_std": 0.581585705280304,
"rewards/cosine_scaled_reward": -0.1466412227600813,
"rewards/format_reward": 0.9791666716337204,
"step": 439
},
{
"completion_length": 1667.5000610351562,
"epoch": 0.5028571428571429,
"grad_norm": 1.8554582595825195,
"kl": 0.833984375,
"learning_rate": 1.3890454406082956e-07,
"loss": 0.0334,
"reward": 0.6555005759000778,
"reward_std": 0.8219043761491776,
"rewards/cosine_scaled_reward": -0.0784997058508452,
"rewards/format_reward": 0.8125000298023224,
"step": 440
},
{
"completion_length": 1507.8542175292969,
"epoch": 0.504,
"grad_norm": 3.88932466506958,
"kl": 0.70849609375,
"learning_rate": 1.3763677169699217e-07,
"loss": 0.0284,
"reward": 0.7302692234516144,
"reward_std": 0.6611650586128235,
"rewards/cosine_scaled_reward": -0.07236538827419281,
"rewards/format_reward": 0.8750000298023224,
"step": 441
},
{
"completion_length": 1471.1250305175781,
"epoch": 0.5051428571428571,
"grad_norm": 3.0517144203186035,
"kl": 0.61572265625,
"learning_rate": 1.3638909733514452e-07,
"loss": 0.0247,
"reward": 0.6592699624598026,
"reward_std": 0.7461230009794235,
"rewards/cosine_scaled_reward": -0.11828170018270612,
"rewards/format_reward": 0.8958333432674408,
"step": 442
},
{
"completion_length": 1453.1458740234375,
"epoch": 0.5062857142857143,
"grad_norm": 2.053025722503662,
"kl": 0.724609375,
"learning_rate": 1.351615817851748e-07,
"loss": 0.029,
"reward": 0.5722346976399422,
"reward_std": 0.643949382007122,
"rewards/cosine_scaled_reward": -0.1305493265390396,
"rewards/format_reward": 0.833333358168602,
"step": 443
},
{
"completion_length": 1375.6875305175781,
"epoch": 0.5074285714285715,
"grad_norm": 1.6717604398727417,
"kl": 0.594390869140625,
"learning_rate": 1.3395428487445914e-07,
"loss": 0.0237,
"reward": 1.0815926790237427,
"reward_std": 0.7383088618516922,
"rewards/cosine_scaled_reward": 0.08246299810707569,
"rewards/format_reward": 0.9166666865348816,
"step": 444
},
{
"completion_length": 1785.2708740234375,
"epoch": 0.5085714285714286,
"grad_norm": 3.6348721981048584,
"kl": 1.021484375,
"learning_rate": 1.3276726544494571e-07,
"loss": 0.0408,
"reward": 0.6278277039527893,
"reward_std": 0.6637818515300751,
"rewards/cosine_scaled_reward": -0.10275283083319664,
"rewards/format_reward": 0.8333333432674408,
"step": 445
},
{
"completion_length": 1080.1458587646484,
"epoch": 0.5097142857142857,
"grad_norm": 1.7892571687698364,
"kl": 0.1240234375,
"learning_rate": 1.316005813502869e-07,
"loss": 0.005,
"reward": 0.8095807060599327,
"reward_std": 0.6449108496308327,
"rewards/cosine_scaled_reward": -0.053542979061603546,
"rewards/format_reward": 0.9166666865348816,
"step": 446
},
{
"completion_length": 1139.9166870117188,
"epoch": 0.5108571428571429,
"grad_norm": 1.82309091091156,
"kl": 0.305145263671875,
"learning_rate": 1.3045428945301953e-07,
"loss": 0.0122,
"reward": 0.9205853343009949,
"reward_std": 0.7828942686319351,
"rewards/cosine_scaled_reward": -0.02929066913202405,
"rewards/format_reward": 0.9791666716337204,
"step": 447
},
{
"completion_length": 1278.1250305175781,
"epoch": 0.512,
"grad_norm": 2.259813070297241,
"kl": 0.251708984375,
"learning_rate": 1.2932844562179352e-07,
"loss": 0.0101,
"reward": 0.8140432685613632,
"reward_std": 0.658391922712326,
"rewards/cosine_scaled_reward": -0.05131170805543661,
"rewards/format_reward": 0.9166666865348816,
"step": 448
},
{
"completion_length": 1173.3125305175781,
"epoch": 0.5131428571428571,
"grad_norm": 1.361348032951355,
"kl": 0.3782958984375,
"learning_rate": 1.2822310472864885e-07,
"loss": 0.0151,
"reward": 0.9565356224775314,
"reward_std": 0.7290000915527344,
"rewards/cosine_scaled_reward": 0.019934438169002533,
"rewards/format_reward": 0.9166666716337204,
"step": 449
},
{
"completion_length": 1371.1667175292969,
"epoch": 0.5142857142857142,
"grad_norm": 3.834918737411499,
"kl": 0.513427734375,
"learning_rate": 1.2713832064634125e-07,
"loss": 0.0205,
"reward": 0.7457224242389202,
"reward_std": 0.5297245979309082,
"rewards/cosine_scaled_reward": -0.04380548745393753,
"rewards/format_reward": 0.8333333432674408,
"step": 450
},
{
"completion_length": 1255.6250305175781,
"epoch": 0.5154285714285715,
"grad_norm": 2.335192918777466,
"kl": 0.6416015625,
"learning_rate": 1.260741462457165e-07,
"loss": 0.0257,
"reward": 1.0948645919561386,
"reward_std": 0.7944803088903427,
"rewards/cosine_scaled_reward": 0.10993227222934365,
"rewards/format_reward": 0.8750000298023224,
"step": 451
},
{
"completion_length": 1190.6666870117188,
"epoch": 0.5165714285714286,
"grad_norm": 1.7608739137649536,
"kl": 0.500244140625,
"learning_rate": 1.2503063339313356e-07,
"loss": 0.02,
"reward": 0.5270581915974617,
"reward_std": 0.6462119966745377,
"rewards/cosine_scaled_reward": -0.1843875776976347,
"rewards/format_reward": 0.895833358168602,
"step": 452
},
{
"completion_length": 1610.3125610351562,
"epoch": 0.5177142857142857,
"grad_norm": 2.1905384063720703,
"kl": 0.96484375,
"learning_rate": 1.2400783294793668e-07,
"loss": 0.0386,
"reward": 0.5264739021658897,
"reward_std": 0.707670621573925,
"rewards/cosine_scaled_reward": -0.1325964080169797,
"rewards/format_reward": 0.7916666865348816,
"step": 453
},
{
"completion_length": 1178.0000305175781,
"epoch": 0.5188571428571429,
"grad_norm": 1.7162837982177734,
"kl": 0.55322265625,
"learning_rate": 1.2300579475997657e-07,
"loss": 0.0221,
"reward": 0.7971140295267105,
"reward_std": 0.5913056135177612,
"rewards/cosine_scaled_reward": -0.059776326175779104,
"rewards/format_reward": 0.9166666865348816,
"step": 454
},
{
"completion_length": 1480.6250305175781,
"epoch": 0.52,
"grad_norm": 4.159846305847168,
"kl": 1.05126953125,
"learning_rate": 1.220245676671809e-07,
"loss": 0.0421,
"reward": 0.6366595476865768,
"reward_std": 0.44621995836496353,
"rewards/cosine_scaled_reward": -0.11917022056877613,
"rewards/format_reward": 0.8750000149011612,
"step": 455
},
{
"completion_length": 1499.1458740234375,
"epoch": 0.5211428571428571,
"grad_norm": 3.277935743331909,
"kl": 0.708251953125,
"learning_rate": 1.2106419949317388e-07,
"loss": 0.0284,
"reward": 0.7240877486765385,
"reward_std": 0.5471076965332031,
"rewards/cosine_scaled_reward": -0.05462279508356005,
"rewards/format_reward": 0.8333333432674408,
"step": 456
},
{
"completion_length": 1379.7292175292969,
"epoch": 0.5222857142857142,
"grad_norm": 2.5969083309173584,
"kl": 0.6796875,
"learning_rate": 1.2012473704494537e-07,
"loss": 0.0272,
"reward": 0.869850842282176,
"reward_std": 0.700515478849411,
"rewards/cosine_scaled_reward": 0.02867540717124939,
"rewards/format_reward": 0.8125000298023224,
"step": 457
},
{
"completion_length": 1159.9791870117188,
"epoch": 0.5234285714285715,
"grad_norm": 2.355196952819824,
"kl": 0.31744384765625,
"learning_rate": 1.1920622611056974e-07,
"loss": 0.0127,
"reward": 0.923922210931778,
"reward_std": 0.5514720380306244,
"rewards/cosine_scaled_reward": -0.017205584794282913,
"rewards/format_reward": 0.9583333432674408,
"step": 458
},
{
"completion_length": 1273.5833892822266,
"epoch": 0.5245714285714286,
"grad_norm": 4.148768901824951,
"kl": 0.59765625,
"learning_rate": 1.1830871145697412e-07,
"loss": 0.0239,
"reward": 0.574098750948906,
"reward_std": 0.7631915658712387,
"rewards/cosine_scaled_reward": -0.16086730360984802,
"rewards/format_reward": 0.895833358168602,
"step": 459
},
{
"completion_length": 1453.9792175292969,
"epoch": 0.5257142857142857,
"grad_norm": 3.0608623027801514,
"kl": 0.8349609375,
"learning_rate": 1.1743223682775649e-07,
"loss": 0.0335,
"reward": 1.0540032014250755,
"reward_std": 0.7098206132650375,
"rewards/cosine_scaled_reward": 0.07908494677394629,
"rewards/format_reward": 0.8958333432674408,
"step": 460
},
{
"completion_length": 1280.1667175292969,
"epoch": 0.5268571428571428,
"grad_norm": 1.6031488180160522,
"kl": 0.51483154296875,
"learning_rate": 1.1657684494105386e-07,
"loss": 0.0206,
"reward": 1.0345441699028015,
"reward_std": 0.464998334646225,
"rewards/cosine_scaled_reward": 0.05893874540925026,
"rewards/format_reward": 0.9166666865348816,
"step": 461
},
{
"completion_length": 1348.2083740234375,
"epoch": 0.528,
"grad_norm": 2.087766647338867,
"kl": 0.288330078125,
"learning_rate": 1.1574257748745986e-07,
"loss": 0.0115,
"reward": 0.5173667185008526,
"reward_std": 0.5679958164691925,
"rewards/cosine_scaled_reward": -0.18923332425765693,
"rewards/format_reward": 0.8958333432674408,
"step": 462
},
{
"completion_length": 1184.8333892822266,
"epoch": 0.5291428571428571,
"grad_norm": 2.3618013858795166,
"kl": 0.67828369140625,
"learning_rate": 1.1492947512799328e-07,
"loss": 0.027,
"reward": 0.8587629348039627,
"reward_std": 0.5902151763439178,
"rewards/cosine_scaled_reward": 0.002298124134540558,
"rewards/format_reward": 0.8541666865348816,
"step": 463
},
{
"completion_length": 1381.3750610351562,
"epoch": 0.5302857142857142,
"grad_norm": 2.5598762035369873,
"kl": 0.4290771484375,
"learning_rate": 1.1413757749211602e-07,
"loss": 0.0172,
"reward": 0.8569861799478531,
"reward_std": 0.5644854605197906,
"rewards/cosine_scaled_reward": -0.04025692865252495,
"rewards/format_reward": 0.9375000149011612,
"step": 464
},
{
"completion_length": 1754.0000610351562,
"epoch": 0.5314285714285715,
"grad_norm": 3.649667739868164,
"kl": 1.2236328125,
"learning_rate": 1.1336692317580158e-07,
"loss": 0.049,
"reward": 0.6691107526421547,
"reward_std": 0.7533622533082962,
"rewards/cosine_scaled_reward": -0.050861308351159096,
"rewards/format_reward": 0.770833358168602,
"step": 465
},
{
"completion_length": 1626.979248046875,
"epoch": 0.5325714285714286,
"grad_norm": 4.014570713043213,
"kl": 1.1494140625,
"learning_rate": 1.1261754973965422e-07,
"loss": 0.046,
"reward": 0.39536508079618216,
"reward_std": 0.5428843200206757,
"rewards/cosine_scaled_reward": -0.18773413076996803,
"rewards/format_reward": 0.7708333432674408,
"step": 466
},
{
"completion_length": 1410.1875305175781,
"epoch": 0.5337142857142857,
"grad_norm": 6.032597064971924,
"kl": 0.89404296875,
"learning_rate": 1.1188949370707787e-07,
"loss": 0.0358,
"reward": 0.7157600894570351,
"reward_std": 0.6754554212093353,
"rewards/cosine_scaled_reward": -0.07961997389793396,
"rewards/format_reward": 0.8750000298023224,
"step": 467
},
{
"completion_length": 1505.0208740234375,
"epoch": 0.5348571428571428,
"grad_norm": 3.0937063694000244,
"kl": 1.083251953125,
"learning_rate": 1.1118279056249653e-07,
"loss": 0.0434,
"reward": 1.0479508265852928,
"reward_std": 0.7750177532434464,
"rewards/cosine_scaled_reward": 0.08647541608661413,
"rewards/format_reward": 0.8750000298023224,
"step": 468
},
{
"completion_length": 1116.0208435058594,
"epoch": 0.536,
"grad_norm": 2.7102174758911133,
"kl": 0.68310546875,
"learning_rate": 1.1049747474962444e-07,
"loss": 0.0273,
"reward": 0.8038362823426723,
"reward_std": 0.7213334441184998,
"rewards/cosine_scaled_reward": -0.04599856585264206,
"rewards/format_reward": 0.8958333432674408,
"step": 469
},
{
"completion_length": 1321.1458740234375,
"epoch": 0.5371428571428571,
"grad_norm": 6.9517316818237305,
"kl": 1.054931640625,
"learning_rate": 1.0983357966978745e-07,
"loss": 0.0422,
"reward": 0.6858988218009472,
"reward_std": 0.6414925083518028,
"rewards/cosine_scaled_reward": -0.0424672719091177,
"rewards/format_reward": 0.7708333432674408,
"step": 470
},
{
"completion_length": 1012.8125457763672,
"epoch": 0.5382857142857143,
"grad_norm": 15.012682914733887,
"kl": 2.00054931640625,
"learning_rate": 1.0919113768029517e-07,
"loss": 0.0798,
"reward": 0.7205886021256447,
"reward_std": 0.697634182870388,
"rewards/cosine_scaled_reward": -0.1084557194262743,
"rewards/format_reward": 0.9375000149011612,
"step": 471
},
{
"completion_length": 1118.0000305175781,
"epoch": 0.5394285714285715,
"grad_norm": 1.921370506286621,
"kl": 0.70703125,
"learning_rate": 1.0857018009286381e-07,
"loss": 0.0282,
"reward": 0.8347459137439728,
"reward_std": 0.7502800822257996,
"rewards/cosine_scaled_reward": -0.03054371359758079,
"rewards/format_reward": 0.895833358168602,
"step": 472
},
{
"completion_length": 1343.3333740234375,
"epoch": 0.5405714285714286,
"grad_norm": 1.8898627758026123,
"kl": 0.716796875,
"learning_rate": 1.0797073717209013e-07,
"loss": 0.0287,
"reward": 0.3787691295146942,
"reward_std": 0.4741926044225693,
"rewards/cosine_scaled_reward": -0.2585321292281151,
"rewards/format_reward": 0.8958333432674408,
"step": 473
},
{
"completion_length": 1074.5417022705078,
"epoch": 0.5417142857142857,
"grad_norm": 1.3952125310897827,
"kl": 0.275390625,
"learning_rate": 1.0739283813397639e-07,
"loss": 0.011,
"reward": 1.0772456228733063,
"reward_std": 0.8115750551223755,
"rewards/cosine_scaled_reward": 0.09070614166557789,
"rewards/format_reward": 0.895833358168602,
"step": 474
},
{
"completion_length": 1536.7708740234375,
"epoch": 0.5428571428571428,
"grad_norm": 1.9422410726547241,
"kl": 0.89697265625,
"learning_rate": 1.068365111445064e-07,
"loss": 0.036,
"reward": 0.7740568369626999,
"reward_std": 0.7255712598562241,
"rewards/cosine_scaled_reward": -0.06088825827464461,
"rewards/format_reward": 0.8958333432674408,
"step": 475
},
{
"completion_length": 1238.6250305175781,
"epoch": 0.544,
"grad_norm": 1.791479468345642,
"kl": 0.49365234375,
"learning_rate": 1.063017833182728e-07,
"loss": 0.0198,
"reward": 1.1282098963856697,
"reward_std": 0.7448728978633881,
"rewards/cosine_scaled_reward": 0.1266049058176577,
"rewards/format_reward": 0.8750000298023224,
"step": 476
},
{
"completion_length": 926.4167175292969,
"epoch": 0.5451428571428572,
"grad_norm": 1.0713694095611572,
"kl": 0.20794677734375,
"learning_rate": 1.0578868071715544e-07,
"loss": 0.0083,
"reward": 1.242761254310608,
"reward_std": 0.48385217040777206,
"rewards/cosine_scaled_reward": 0.13179726898670197,
"rewards/format_reward": 0.9791666716337204,
"step": 477
},
{
"completion_length": 1389.8750457763672,
"epoch": 0.5462857142857143,
"grad_norm": 1.6024447679519653,
"kl": 0.315399169921875,
"learning_rate": 1.0529722834905125e-07,
"loss": 0.0126,
"reward": 0.5958605632185936,
"reward_std": 0.5410900861024857,
"rewards/cosine_scaled_reward": -0.17081973887979984,
"rewards/format_reward": 0.9375000149011612,
"step": 478
},
{
"completion_length": 1060.0000305175781,
"epoch": 0.5474285714285714,
"grad_norm": 1.9674124717712402,
"kl": 0.351318359375,
"learning_rate": 1.0482745016665526e-07,
"loss": 0.0141,
"reward": 0.9066205322742462,
"reward_std": 0.8546933829784393,
"rewards/cosine_scaled_reward": -0.036273106932640076,
"rewards/format_reward": 0.9791666716337204,
"step": 479
},
{
"completion_length": 1312.5208740234375,
"epoch": 0.5485714285714286,
"grad_norm": 1.8547768592834473,
"kl": 0.3245697021484375,
"learning_rate": 1.0437936906629334e-07,
"loss": 0.013,
"reward": 1.0583224594593048,
"reward_std": 0.5566529557108879,
"rewards/cosine_scaled_reward": 0.04999455437064171,
"rewards/format_reward": 0.9583333432674408,
"step": 480
},
{
"completion_length": 1382.7500610351562,
"epoch": 0.5497142857142857,
"grad_norm": 1.5025635957717896,
"kl": 0.718505859375,
"learning_rate": 1.0395300688680625e-07,
"loss": 0.0287,
"reward": 0.6311604380607605,
"reward_std": 0.6715650781989098,
"rewards/cosine_scaled_reward": -0.1323364470154047,
"rewards/format_reward": 0.8958333432674408,
"step": 481
},
{
"completion_length": 1061.8750305175781,
"epoch": 0.5508571428571428,
"grad_norm": 2.283987522125244,
"kl": 0.368408203125,
"learning_rate": 1.0354838440848501e-07,
"loss": 0.0148,
"reward": 0.9338921532034874,
"reward_std": 0.663320891559124,
"rewards/cosine_scaled_reward": -0.022637249901890755,
"rewards/format_reward": 0.9791666716337204,
"step": 482
},
{
"completion_length": 1664.0000305175781,
"epoch": 0.552,
"grad_norm": 3.247492551803589,
"kl": 1.43994140625,
"learning_rate": 1.0316552135205837e-07,
"loss": 0.0576,
"reward": 0.4382214695215225,
"reward_std": 0.6629593223333359,
"rewards/cosine_scaled_reward": -0.1454725954681635,
"rewards/format_reward": 0.7291666865348816,
"step": 483
},
{
"completion_length": 1166.9375457763672,
"epoch": 0.5531428571428572,
"grad_norm": 1.635184407234192,
"kl": 0.59814453125,
"learning_rate": 1.0280443637773163e-07,
"loss": 0.0239,
"reward": 1.209791585803032,
"reward_std": 0.48193909227848053,
"rewards/cosine_scaled_reward": 0.12572911009192467,
"rewards/format_reward": 0.9583333432674408,
"step": 484
},
{
"completion_length": 1562.9167022705078,
"epoch": 0.5542857142857143,
"grad_norm": 1.5177675485610962,
"kl": 1.32794189453125,
"learning_rate": 1.0246514708427701e-07,
"loss": 0.0532,
"reward": 0.6973680332303047,
"reward_std": 0.6229738295078278,
"rewards/cosine_scaled_reward": -0.07839931827038527,
"rewards/format_reward": 0.8541666716337204,
"step": 485
},
{
"completion_length": 1258.0416870117188,
"epoch": 0.5554285714285714,
"grad_norm": 1.6634312868118286,
"kl": 0.598846435546875,
"learning_rate": 1.0214767000817596e-07,
"loss": 0.0239,
"reward": 0.9936239048838615,
"reward_std": 0.7174008414149284,
"rewards/cosine_scaled_reward": 0.038478586822748184,
"rewards/format_reward": 0.9166666865348816,
"step": 486
},
{
"completion_length": 1263.3125610351562,
"epoch": 0.5565714285714286,
"grad_norm": 3.9874322414398193,
"kl": 0.56591796875,
"learning_rate": 1.0185202062281336e-07,
"loss": 0.0226,
"reward": 0.9569487422704697,
"reward_std": 0.776068776845932,
"rewards/cosine_scaled_reward": 0.009724359028041363,
"rewards/format_reward": 0.9375000149011612,
"step": 487
},
{
"completion_length": 1440.8333435058594,
"epoch": 0.5577142857142857,
"grad_norm": 3.2814691066741943,
"kl": 0.67236328125,
"learning_rate": 1.0157821333772304e-07,
"loss": 0.0269,
"reward": 0.4271909072995186,
"reward_std": 0.5705131366848946,
"rewards/cosine_scaled_reward": -0.2447379156947136,
"rewards/format_reward": 0.9166666865348816,
"step": 488
},
{
"completion_length": 983.0625457763672,
"epoch": 0.5588571428571428,
"grad_norm": 1.347419261932373,
"kl": 0.39642333984375,
"learning_rate": 1.013262614978859e-07,
"loss": 0.0158,
"reward": 1.1842274367809296,
"reward_std": 0.5909858122467995,
"rewards/cosine_scaled_reward": 0.11294705420732498,
"rewards/format_reward": 0.9583333432674408,
"step": 489
},
{
"completion_length": 1493.3958740234375,
"epoch": 0.56,
"grad_norm": 1.7785853147506714,
"kl": 0.92236328125,
"learning_rate": 1.0109617738307911e-07,
"loss": 0.0368,
"reward": 0.5224965363740921,
"reward_std": 0.5668257884681225,
"rewards/cosine_scaled_reward": -0.1450017336755991,
"rewards/format_reward": 0.8125000298023224,
"step": 490
},
{
"completion_length": 1626.2916870117188,
"epoch": 0.5611428571428572,
"grad_norm": 1.94631028175354,
"kl": 1.02734375,
"learning_rate": 1.0088797220727779e-07,
"loss": 0.0411,
"reward": 0.6662824004888535,
"reward_std": 0.6974209845066071,
"rewards/cosine_scaled_reward": -0.07310881093144417,
"rewards/format_reward": 0.8125000298023224,
"step": 491
},
{
"completion_length": 1049.5833740234375,
"epoch": 0.5622857142857143,
"grad_norm": 2.7264091968536377,
"kl": 0.43701171875,
"learning_rate": 1.0070165611810855e-07,
"loss": 0.0175,
"reward": 0.8967996649444103,
"reward_std": 0.40792329236865044,
"rewards/cosine_scaled_reward": -0.02035021036863327,
"rewards/format_reward": 0.9375,
"step": 492
},
{
"completion_length": 1251.0000610351562,
"epoch": 0.5634285714285714,
"grad_norm": 1.9115831851959229,
"kl": 0.535888671875,
"learning_rate": 1.005372381963547e-07,
"loss": 0.0215,
"reward": 0.7575453743338585,
"reward_std": 0.6094752550125122,
"rewards/cosine_scaled_reward": -0.11081065610051155,
"rewards/format_reward": 0.9791666716337204,
"step": 493
},
{
"completion_length": 1035.9375305175781,
"epoch": 0.5645714285714286,
"grad_norm": 3.7157654762268066,
"kl": 0.62744140625,
"learning_rate": 1.0039472645551372e-07,
"loss": 0.0251,
"reward": 0.49045146629214287,
"reward_std": 0.4288835674524307,
"rewards/cosine_scaled_reward": -0.21310760331107304,
"rewards/format_reward": 0.9166666865348816,
"step": 494
},
{
"completion_length": 1336.2291870117188,
"epoch": 0.5657142857142857,
"grad_norm": 1.2452759742736816,
"kl": 0.7421875,
"learning_rate": 1.002741278414069e-07,
"loss": 0.0296,
"reward": 1.002578854560852,
"reward_std": 0.5317458659410477,
"rewards/cosine_scaled_reward": 0.05337274447083473,
"rewards/format_reward": 0.895833358168602,
"step": 495
},
{
"completion_length": 1363.2500305175781,
"epoch": 0.5668571428571428,
"grad_norm": 1.7971941232681274,
"kl": 0.7275390625,
"learning_rate": 1.0017544823184055e-07,
"loss": 0.0291,
"reward": 1.0207700282335281,
"reward_std": 0.8502290099859238,
"rewards/cosine_scaled_reward": 0.09371834748890251,
"rewards/format_reward": 0.833333358168602,
"step": 496
},
{
"completion_length": 1141.7500305175781,
"epoch": 0.568,
"grad_norm": 2.350130796432495,
"kl": 0.56689453125,
"learning_rate": 1.0009869243631952e-07,
"loss": 0.0227,
"reward": 1.0808594226837158,
"reward_std": 0.7858606725931168,
"rewards/cosine_scaled_reward": 0.14459637086838484,
"rewards/format_reward": 0.7916666865348816,
"step": 497
},
{
"completion_length": 1354.0625610351562,
"epoch": 0.5691428571428572,
"grad_norm": 2.9282283782958984,
"kl": 0.8212890625,
"learning_rate": 1.000438641958131e-07,
"loss": 0.0329,
"reward": 0.7300854474306107,
"reward_std": 0.6824060678482056,
"rewards/cosine_scaled_reward": -0.08287395909428596,
"rewards/format_reward": 0.895833358168602,
"step": 498
},
{
"completion_length": 1752.0625610351562,
"epoch": 0.5702857142857143,
"grad_norm": 1.9427164793014526,
"kl": 0.99267578125,
"learning_rate": 1.0001096618257236e-07,
"loss": 0.0397,
"reward": 0.6041746586561203,
"reward_std": 0.5943188220262527,
"rewards/cosine_scaled_reward": -0.12499601114541292,
"rewards/format_reward": 0.8541666865348816,
"step": 499
},
{
"completion_length": 1389.7292175292969,
"epoch": 0.5714285714285714,
"grad_norm": 3.8614306449890137,
"kl": 0.90966796875,
"learning_rate": 1e-07,
"loss": 0.0364,
"reward": 0.8081704080104828,
"reward_std": 0.7910114228725433,
"rewards/cosine_scaled_reward": -0.022998109459877014,
"rewards/format_reward": 0.8541666716337204,
"step": 500
},
{
"epoch": 0.5714285714285714,
"step": 500,
"total_flos": 0.0,
"train_loss": 0.007592763372914685,
"train_runtime": 38120.128,
"train_samples_per_second": 0.63,
"train_steps_per_second": 0.013
}
],
"logging_steps": 1,
"max_steps": 500,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 50,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}