{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 478,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 877.3754021335358,
      "learning_rate": 1.0416666666666666e-08,
      "logits/chosen": -1.689455509185791,
      "logits/rejected": -1.4794573783874512,
      "logps/chosen": -126.21005249023438,
      "logps/rejected": -98.13133239746094,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.02,
      "grad_norm": 774.0071219589337,
      "learning_rate": 1.0416666666666667e-07,
      "logits/chosen": -1.7073653936386108,
      "logits/rejected": -1.610068678855896,
      "logps/chosen": -139.72607421875,
      "logps/rejected": -91.37654113769531,
      "loss": 0.6859,
      "rewards/accuracies": 0.5138888955116272,
      "rewards/chosen": -0.0006362733547575772,
      "rewards/margins": 0.015451871789991856,
      "rewards/rejected": -0.016088144853711128,
      "step": 10
    },
    {
      "epoch": 0.04,
      "grad_norm": 288.7199289028814,
      "learning_rate": 2.0833333333333333e-07,
      "logits/chosen": -1.6388952732086182,
      "logits/rejected": -1.6494001150131226,
      "logps/chosen": -130.20901489257812,
      "logps/rejected": -94.0677490234375,
      "loss": 0.491,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": 0.4931251108646393,
      "rewards/margins": 0.6499455571174622,
      "rewards/rejected": -0.15682044625282288,
      "step": 20
    },
    {
      "epoch": 0.06,
      "grad_norm": 224.69456938109306,
      "learning_rate": 3.1249999999999997e-07,
      "logits/chosen": -1.6837990283966064,
      "logits/rejected": -1.6247406005859375,
      "logps/chosen": -130.29306030273438,
      "logps/rejected": -107.59431457519531,
      "loss": 0.2826,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.5809106826782227,
      "rewards/margins": 2.7498364448547363,
      "rewards/rejected": -1.1689256429672241,
      "step": 30
    },
    {
      "epoch": 0.08,
      "grad_norm": 183.79856437398223,
      "learning_rate": 4.1666666666666667e-07,
      "logits/chosen": -1.6083917617797852,
      "logits/rejected": -1.5386857986450195,
      "logps/chosen": -142.89288330078125,
      "logps/rejected": -118.29256439208984,
      "loss": 0.2069,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.7656170129776,
      "rewards/margins": 5.709896087646484,
      "rewards/rejected": -3.944279432296753,
      "step": 40
    },
    {
      "epoch": 0.1,
      "grad_norm": 299.30898027438695,
      "learning_rate": 4.999733114418725e-07,
      "logits/chosen": -1.5618069171905518,
      "logits/rejected": -1.589414119720459,
      "logps/chosen": -129.36355590820312,
      "logps/rejected": -135.2159881591797,
      "loss": 0.1965,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 1.1421730518341064,
      "rewards/margins": 8.362150192260742,
      "rewards/rejected": -7.219976902008057,
      "step": 50
    },
    {
      "epoch": 0.13,
      "grad_norm": 245.51415645278104,
      "learning_rate": 4.990398100856366e-07,
      "logits/chosen": -1.6810020208358765,
      "logits/rejected": -1.6362440586090088,
      "logps/chosen": -146.06399536132812,
      "logps/rejected": -148.35720825195312,
      "loss": 0.1815,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 0.9872430562973022,
      "rewards/margins": 10.306981086730957,
      "rewards/rejected": -9.319738388061523,
      "step": 60
    },
    {
      "epoch": 0.15,
      "grad_norm": 184.4357866731278,
      "learning_rate": 4.967775735898179e-07,
      "logits/chosen": -1.6720081567764282,
      "logits/rejected": -1.706272840499878,
      "logps/chosen": -138.72039794921875,
      "logps/rejected": -148.6911163330078,
      "loss": 0.1754,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.2756741046905518,
      "rewards/margins": 11.420818328857422,
      "rewards/rejected": -10.14514446258545,
      "step": 70
    },
    {
      "epoch": 0.17,
      "grad_norm": 147.5148614226922,
      "learning_rate": 4.931986719649298e-07,
      "logits/chosen": -1.6803735494613647,
      "logits/rejected": -1.677125334739685,
      "logps/chosen": -131.1448211669922,
      "logps/rejected": -137.2668914794922,
      "loss": 0.1602,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.8630034923553467,
      "rewards/margins": 10.570993423461914,
      "rewards/rejected": -8.707988739013672,
      "step": 80
    },
    {
      "epoch": 0.19,
      "grad_norm": 231.45195159349998,
      "learning_rate": 4.883222001996351e-07,
      "logits/chosen": -1.7273483276367188,
      "logits/rejected": -1.7491174936294556,
      "logps/chosen": -137.68093872070312,
      "logps/rejected": -149.29808044433594,
      "loss": 0.1529,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 2.305860757827759,
      "rewards/margins": 11.786917686462402,
      "rewards/rejected": -9.481058120727539,
      "step": 90
    },
    {
      "epoch": 0.21,
      "grad_norm": 125.1418915635804,
      "learning_rate": 4.821741763807186e-07,
      "logits/chosen": -1.7639491558074951,
      "logits/rejected": -1.7565345764160156,
      "logps/chosen": -118.08821105957031,
      "logps/rejected": -132.42239379882812,
      "loss": 0.1523,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.9779144525527954,
      "rewards/margins": 10.331491470336914,
      "rewards/rejected": -8.35357666015625,
      "step": 100
    },
    {
      "epoch": 0.21,
      "eval_logits/chosen": -1.810097336769104,
      "eval_logits/rejected": -1.8013982772827148,
      "eval_logps/chosen": -125.05193328857422,
      "eval_logps/rejected": -137.99850463867188,
      "eval_loss": 0.13986752927303314,
      "eval_rewards/accuracies": 0.9375,
      "eval_rewards/chosen": 2.5440609455108643,
      "eval_rewards/margins": 11.49561595916748,
      "eval_rewards/rejected": -8.951555252075195,
      "eval_runtime": 97.654,
      "eval_samples_per_second": 20.48,
      "eval_steps_per_second": 0.328,
      "step": 100
    },
    {
      "epoch": 0.23,
      "grad_norm": 145.2147024977398,
      "learning_rate": 4.747874028753375e-07,
      "logits/chosen": -1.6646867990493774,
      "logits/rejected": -1.7233518362045288,
      "logps/chosen": -122.03011322021484,
      "logps/rejected": -140.46389770507812,
      "loss": 0.1545,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 2.0033175945281982,
      "rewards/margins": 9.661236763000488,
      "rewards/rejected": -7.657918453216553,
      "step": 110
    },
    {
      "epoch": 0.25,
      "grad_norm": 129.9317141592677,
      "learning_rate": 4.662012913161997e-07,
      "logits/chosen": -1.7326112985610962,
      "logits/rejected": -1.7273359298706055,
      "logps/chosen": -119.6495132446289,
      "logps/rejected": -140.51266479492188,
      "loss": 0.1492,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 2.637754440307617,
      "rewards/margins": 10.471738815307617,
      "rewards/rejected": -7.833985805511475,
      "step": 120
    },
    {
      "epoch": 0.27,
      "grad_norm": 81.32565764107251,
      "learning_rate": 4.5646165232345103e-07,
      "logits/chosen": -1.6919190883636475,
      "logits/rejected": -1.7214174270629883,
      "logps/chosen": -124.76143646240234,
      "logps/rejected": -136.91891479492188,
      "loss": 0.144,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 3.1906180381774902,
      "rewards/margins": 11.917051315307617,
      "rewards/rejected": -8.726433753967285,
      "step": 130
    },
    {
      "epoch": 0.29,
      "grad_norm": 233.71107343479284,
      "learning_rate": 4.456204510851956e-07,
      "logits/chosen": -1.6101233959197998,
      "logits/rejected": -1.5833767652511597,
      "logps/chosen": -121.59381103515625,
      "logps/rejected": -132.78977966308594,
      "loss": 0.1501,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 3.1950697898864746,
      "rewards/margins": 11.13970947265625,
      "rewards/rejected": -7.944640159606934,
      "step": 140
    },
    {
      "epoch": 0.31,
      "grad_norm": 177.51912736620577,
      "learning_rate": 4.337355301007335e-07,
      "logits/chosen": -1.7467788457870483,
      "logits/rejected": -1.7698566913604736,
      "logps/chosen": -119.5277328491211,
      "logps/rejected": -128.22744750976562,
      "loss": 0.1433,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 3.2382194995880127,
      "rewards/margins": 10.8007173538208,
      "rewards/rejected": -7.562497615814209,
      "step": 150
    },
    {
      "epoch": 0.33,
      "grad_norm": 248.7150598871683,
      "learning_rate": 4.2087030056579986e-07,
      "logits/chosen": -1.6178503036499023,
      "logits/rejected": -1.5939674377441406,
      "logps/chosen": -127.01255798339844,
      "logps/rejected": -134.47132873535156,
      "loss": 0.1508,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 2.990654468536377,
      "rewards/margins": 10.095832824707031,
      "rewards/rejected": -7.105177879333496,
      "step": 160
    },
    {
      "epoch": 0.36,
      "grad_norm": 201.08598013628404,
      "learning_rate": 4.070934040463998e-07,
      "logits/chosen": -1.7953569889068604,
      "logits/rejected": -1.7801055908203125,
      "logps/chosen": -122.10235595703125,
      "logps/rejected": -136.7244873046875,
      "loss": 0.1455,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 3.407832622528076,
      "rewards/margins": 11.618408203125,
      "rewards/rejected": -8.210575103759766,
      "step": 170
    },
    {
      "epoch": 0.38,
      "grad_norm": 169.66216294230117,
      "learning_rate": 3.9247834624635404e-07,
      "logits/chosen": -1.7412538528442383,
      "logits/rejected": -1.734012246131897,
      "logps/chosen": -126.0450439453125,
      "logps/rejected": -131.69639587402344,
      "loss": 0.1565,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 3.421628952026367,
      "rewards/margins": 11.586966514587402,
      "rewards/rejected": -8.165337562561035,
      "step": 180
    },
    {
      "epoch": 0.4,
      "grad_norm": 165.65601380930994,
      "learning_rate": 3.7710310482256523e-07,
      "logits/chosen": -1.7725350856781006,
      "logits/rejected": -1.7596399784088135,
      "logps/chosen": -115.76643371582031,
      "logps/rejected": -146.38294982910156,
      "loss": 0.1609,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 3.206624984741211,
      "rewards/margins": 12.064164161682129,
      "rewards/rejected": -8.857539176940918,
      "step": 190
    },
    {
      "epoch": 0.42,
      "grad_norm": 164.38352614855714,
      "learning_rate": 3.610497133404795e-07,
      "logits/chosen": -1.7693474292755127,
      "logits/rejected": -1.7811864614486694,
      "logps/chosen": -121.21810150146484,
      "logps/rejected": -129.79788208007812,
      "loss": 0.176,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 3.5836567878723145,
      "rewards/margins": 11.941611289978027,
      "rewards/rejected": -8.357954025268555,
      "step": 200
    },
    {
      "epoch": 0.42,
      "eval_logits/chosen": -1.8763718605041504,
      "eval_logits/rejected": -1.8761746883392334,
      "eval_logps/chosen": -120.78527069091797,
      "eval_logps/rejected": -137.00637817382812,
      "eval_loss": 0.13582415878772736,
      "eval_rewards/accuracies": 0.9375,
      "eval_rewards/chosen": 3.3973963260650635,
      "eval_rewards/margins": 12.150527954101562,
      "eval_rewards/rejected": -8.753131866455078,
      "eval_runtime": 97.4856,
      "eval_samples_per_second": 20.516,
      "eval_steps_per_second": 0.328,
      "step": 200
    },
    {
      "epoch": 0.44,
      "grad_norm": 177.62042791365712,
      "learning_rate": 3.4440382358952115e-07,
      "logits/chosen": -1.708108901977539,
      "logits/rejected": -1.654348611831665,
      "logps/chosen": -118.52095794677734,
      "logps/rejected": -123.7565689086914,
      "loss": 0.1417,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 2.8139612674713135,
      "rewards/margins": 9.644906997680664,
      "rewards/rejected": -6.830945014953613,
      "step": 210
    },
    {
      "epoch": 0.46,
      "grad_norm": 106.72676059840462,
      "learning_rate": 3.272542485937368e-07,
      "logits/chosen": -1.8347289562225342,
      "logits/rejected": -1.8747615814208984,
      "logps/chosen": -118.0849609375,
      "logps/rejected": -125.06263732910156,
      "loss": 0.1482,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 3.678068161010742,
      "rewards/margins": 11.184865951538086,
      "rewards/rejected": -7.506799221038818,
      "step": 220
    },
    {
      "epoch": 0.48,
      "grad_norm": 105.28694347115214,
      "learning_rate": 3.096924887558854e-07,
      "logits/chosen": -1.7717214822769165,
      "logits/rejected": -1.720580816268921,
      "logps/chosen": -131.0685577392578,
      "logps/rejected": -143.81317138671875,
      "loss": 0.1684,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 3.328054904937744,
      "rewards/margins": 12.128056526184082,
      "rewards/rejected": -8.800003051757812,
      "step": 230
    },
    {
      "epoch": 0.5,
      "grad_norm": 190.93999332658404,
      "learning_rate": 2.9181224366319943e-07,
      "logits/chosen": -1.7859230041503906,
      "logits/rejected": -1.802541971206665,
      "logps/chosen": -117.80191802978516,
      "logps/rejected": -135.9575653076172,
      "loss": 0.1559,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 2.8246419429779053,
      "rewards/margins": 10.490262031555176,
      "rewards/rejected": -7.665619850158691,
      "step": 240
    },
    {
      "epoch": 0.52,
      "grad_norm": 200.5301056579919,
      "learning_rate": 2.7370891215954565e-07,
      "logits/chosen": -1.7719627618789673,
      "logits/rejected": -1.8007062673568726,
      "logps/chosen": -120.75521087646484,
      "logps/rejected": -136.2689971923828,
      "loss": 0.1635,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 3.3319907188415527,
      "rewards/margins": 11.189115524291992,
      "rewards/rejected": -7.857124328613281,
      "step": 250
    },
    {
      "epoch": 0.54,
      "grad_norm": 109.42764564797797,
      "learning_rate": 2.55479083351317e-07,
      "logits/chosen": -1.8106673955917358,
      "logits/rejected": -1.8370100259780884,
      "logps/chosen": -125.1119613647461,
      "logps/rejected": -127.0175552368164,
      "loss": 0.1411,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 3.7430100440979004,
      "rewards/margins": 12.171258926391602,
      "rewards/rejected": -8.428248405456543,
      "step": 260
    },
    {
      "epoch": 0.56,
      "grad_norm": 107.44062116926116,
      "learning_rate": 2.3722002126275822e-07,
      "logits/chosen": -1.795784592628479,
      "logits/rejected": -1.726580262184143,
      "logps/chosen": -114.95989990234375,
      "logps/rejected": -129.06564331054688,
      "loss": 0.1393,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 3.4197163581848145,
      "rewards/margins": 10.986078262329102,
      "rewards/rejected": -7.566361427307129,
      "step": 270
    },
    {
      "epoch": 0.59,
      "grad_norm": 191.88979341493172,
      "learning_rate": 2.19029145890313e-07,
      "logits/chosen": -1.6350713968276978,
      "logits/rejected": -1.7275726795196533,
      "logps/chosen": -121.347412109375,
      "logps/rejected": -143.2399444580078,
      "loss": 0.1733,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 3.141850233078003,
      "rewards/margins": 11.51993179321289,
      "rewards/rejected": -8.378082275390625,
      "step": 280
    },
    {
      "epoch": 0.61,
      "grad_norm": 187.2847938998864,
      "learning_rate": 2.0100351342479216e-07,
      "logits/chosen": -1.7116390466690063,
      "logits/rejected": -1.6921733617782593,
      "logps/chosen": -112.99186706542969,
      "logps/rejected": -130.45889282226562,
      "loss": 0.1534,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 3.034196376800537,
      "rewards/margins": 11.30268669128418,
      "rewards/rejected": -8.268491744995117,
      "step": 290
    },
    {
      "epoch": 0.63,
      "grad_norm": 427.8567531133513,
      "learning_rate": 1.8323929841460178e-07,
      "logits/chosen": -1.6803079843521118,
      "logits/rejected": -1.6214721202850342,
      "logps/chosen": -130.09739685058594,
      "logps/rejected": -148.15475463867188,
      "loss": 0.1509,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 3.2685546875,
      "rewards/margins": 11.6514892578125,
      "rewards/rejected": -8.382935523986816,
      "step": 300
    },
    {
      "epoch": 0.63,
      "eval_logits/chosen": -1.7874743938446045,
      "eval_logits/rejected": -1.7873148918151855,
      "eval_logps/chosen": -121.00535583496094,
      "eval_logps/rejected": -139.82205200195312,
      "eval_loss": 0.14029568433761597,
      "eval_rewards/accuracies": 0.93359375,
      "eval_rewards/chosen": 3.353379249572754,
      "eval_rewards/margins": 12.669642448425293,
      "eval_rewards/rejected": -9.316261291503906,
      "eval_runtime": 97.7473,
      "eval_samples_per_second": 20.461,
      "eval_steps_per_second": 0.327,
      "step": 300
    },
    {
      "epoch": 0.65,
      "grad_norm": 209.36571305183378,
      "learning_rate": 1.6583128063291573e-07,
      "logits/chosen": -1.6824207305908203,
      "logits/rejected": -1.7118114233016968,
      "logps/chosen": -122.2557144165039,
      "logps/rejected": -138.8484344482422,
      "loss": 0.1544,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 3.416393280029297,
      "rewards/margins": 11.478261947631836,
      "rewards/rejected": -8.061868667602539,
      "step": 310
    },
    {
      "epoch": 0.67,
      "grad_norm": 115.39699943542352,
      "learning_rate": 1.488723393865766e-07,
      "logits/chosen": -1.7209253311157227,
      "logits/rejected": -1.6945152282714844,
      "logps/chosen": -111.72103118896484,
      "logps/rejected": -138.40509033203125,
      "loss": 0.144,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 3.8151679039001465,
      "rewards/margins": 12.245112419128418,
      "rewards/rejected": -8.42994499206543,
      "step": 320
    },
    {
      "epoch": 0.69,
      "grad_norm": 101.70209071929943,
      "learning_rate": 1.3245295796480788e-07,
      "logits/chosen": -1.6543442010879517,
      "logits/rejected": -1.7321891784667969,
      "logps/chosen": -122.24208068847656,
      "logps/rejected": -133.8969268798828,
      "loss": 0.1435,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 3.18817400932312,
      "rewards/margins": 11.217942237854004,
      "rewards/rejected": -8.029767990112305,
      "step": 330
    },
    {
      "epoch": 0.71,
      "grad_norm": 123.95909153688254,
      "learning_rate": 1.1666074087171627e-07,
      "logits/chosen": -1.6039130687713623,
      "logits/rejected": -1.6741559505462646,
      "logps/chosen": -116.26409912109375,
      "logps/rejected": -122.7804946899414,
      "loss": 0.1746,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 2.8452401161193848,
      "rewards/margins": 9.874032974243164,
      "rewards/rejected": -7.028792381286621,
      "step": 340
    },
    {
      "epoch": 0.73,
      "grad_norm": 294.0385926355559,
      "learning_rate": 1.0157994641835734e-07,
      "logits/chosen": -1.6423372030258179,
      "logits/rejected": -1.6570653915405273,
      "logps/chosen": -116.1936264038086,
      "logps/rejected": -144.1549072265625,
      "loss": 0.1291,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 2.7782697677612305,
      "rewards/margins": 11.297493934631348,
      "rewards/rejected": -8.519224166870117,
      "step": 350
    },
    {
      "epoch": 0.75,
      "grad_norm": 147.62661016854165,
      "learning_rate": 8.729103716819111e-08,
      "logits/chosen": -1.6380598545074463,
      "logits/rejected": -1.6626033782958984,
      "logps/chosen": -122.89555358886719,
      "logps/rejected": -138.44113159179688,
      "loss": 0.1469,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 3.050956964492798,
      "rewards/margins": 10.875678062438965,
      "rewards/rejected": -7.824721336364746,
      "step": 360
    },
    {
      "epoch": 0.77,
      "grad_norm": 250.3877641186192,
      "learning_rate": 7.387025063449081e-08,
      "logits/chosen": -1.749659776687622,
      "logits/rejected": -1.6942018270492554,
      "logps/chosen": -113.17866516113281,
      "logps/rejected": -127.69132232666016,
      "loss": 0.1488,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 2.642218828201294,
      "rewards/margins": 9.677666664123535,
      "rewards/rejected": -7.035447597503662,
      "step": 370
    },
    {
      "epoch": 0.79,
      "grad_norm": 179.11776110792954,
      "learning_rate": 6.138919252022435e-08,
      "logits/chosen": -1.7493488788604736,
      "logits/rejected": -1.777907133102417,
      "logps/chosen": -119.62767028808594,
      "logps/rejected": -144.23379516601562,
      "loss": 0.1547,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 3.405210494995117,
      "rewards/margins": 12.410168647766113,
      "rewards/rejected": -9.004958152770996,
      "step": 380
    },
    {
      "epoch": 0.82,
      "grad_norm": 105.79142677310188,
      "learning_rate": 4.991445467064689e-08,
      "logits/chosen": -1.6522718667984009,
      "logits/rejected": -1.657579779624939,
      "logps/chosen": -115.9373779296875,
      "logps/rejected": -131.8628387451172,
      "loss": 0.1191,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 3.3116908073425293,
      "rewards/margins": 11.62342643737793,
      "rewards/rejected": -8.311735153198242,
      "step": 390
    },
    {
      "epoch": 0.84,
      "grad_norm": 160.26625520373477,
      "learning_rate": 3.9507259776993954e-08,
      "logits/chosen": -1.6317713260650635,
      "logits/rejected": -1.717016577720642,
      "logps/chosen": -118.45027923583984,
      "logps/rejected": -131.49853515625,
      "loss": 0.2009,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 2.862239360809326,
      "rewards/margins": 10.845926284790039,
      "rewards/rejected": -7.983686923980713,
      "step": 400
    },
    {
      "epoch": 0.84,
      "eval_logits/chosen": -1.8097559213638306,
      "eval_logits/rejected": -1.8106379508972168,
      "eval_logps/chosen": -120.3245849609375,
      "eval_logps/rejected": -139.5015411376953,
      "eval_loss": 0.1389647275209427,
      "eval_rewards/accuracies": 0.9296875,
      "eval_rewards/chosen": 3.489531993865967,
      "eval_rewards/margins": 12.741693496704102,
      "eval_rewards/rejected": -9.252161026000977,
      "eval_runtime": 97.6986,
      "eval_samples_per_second": 20.471,
      "eval_steps_per_second": 0.328,
      "step": 400
    },
    {
      "epoch": 0.86,
      "grad_norm": 192.9545547193168,
      "learning_rate": 3.022313472693447e-08,
      "logits/chosen": -1.7275054454803467,
      "logits/rejected": -1.756888747215271,
      "logps/chosen": -129.18605041503906,
      "logps/rejected": -136.9790496826172,
      "loss": 0.1378,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 3.625795841217041,
      "rewards/margins": 12.600431442260742,
      "rewards/rejected": -8.974635124206543,
      "step": 410
    },
    {
      "epoch": 0.88,
      "grad_norm": 135.89148969120785,
      "learning_rate": 2.2111614344599684e-08,
      "logits/chosen": -1.697636365890503,
      "logits/rejected": -1.6976314783096313,
      "logps/chosen": -114.35809326171875,
      "logps/rejected": -131.75164794921875,
      "loss": 0.1286,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 2.790315866470337,
      "rewards/margins": 11.007737159729004,
      "rewards/rejected": -8.21742057800293,
      "step": 420
    },
    {
      "epoch": 0.9,
      "grad_norm": 163.87841082901525,
      "learning_rate": 1.521597710086439e-08,
      "logits/chosen": -1.6538774967193604,
      "logits/rejected": -1.667636513710022,
      "logps/chosen": -129.87210083007812,
      "logps/rejected": -131.872802734375,
      "loss": 0.1437,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 3.2474303245544434,
      "rewards/margins": 10.924890518188477,
      "rewards/rejected": -7.677459716796875,
      "step": 430
    },
    {
      "epoch": 0.92,
      "grad_norm": 117.88131682798563,
      "learning_rate": 9.57301420397924e-09,
      "logits/chosen": -1.7787322998046875,
      "logits/rejected": -1.7886247634887695,
      "logps/chosen": -121.3507080078125,
      "logps/rejected": -139.0587921142578,
      "loss": 0.142,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 3.5244851112365723,
      "rewards/margins": 12.655617713928223,
      "rewards/rejected": -9.131133079528809,
      "step": 440
    },
    {
      "epoch": 0.94,
      "grad_norm": 179.270368210056,
      "learning_rate": 5.212833302556258e-09,
      "logits/chosen": -1.792822241783142,
      "logits/rejected": -1.7586778402328491,
      "logps/chosen": -116.64122009277344,
      "logps/rejected": -135.26832580566406,
      "loss": 0.1538,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 3.7331383228302,
      "rewards/margins": 13.112436294555664,
      "rewards/rejected": -9.379298210144043,
      "step": 450
    },
    {
      "epoch": 0.96,
      "grad_norm": 118.73739536412977,
      "learning_rate": 2.158697848236607e-09,
      "logits/chosen": -1.6758760213851929,
      "logits/rejected": -1.6996933221817017,
      "logps/chosen": -120.97281646728516,
      "logps/rejected": -138.555419921875,
      "loss": 0.1252,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 3.676201581954956,
      "rewards/margins": 12.26156234741211,
      "rewards/rejected": -8.585359573364258,
      "step": 460
    },
    {
      "epoch": 0.98,
      "grad_norm": 152.28679005488013,
      "learning_rate": 4.269029751107489e-10,
      "logits/chosen": -1.6590917110443115,
      "logits/rejected": -1.6743139028549194,
      "logps/chosen": -115.36566162109375,
      "logps/rejected": -147.09939575195312,
      "loss": 0.1469,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 2.869194269180298,
      "rewards/margins": 11.534073829650879,
      "rewards/rejected": -8.66487979888916,
      "step": 470
    },
    {
      "epoch": 1.0,
      "step": 478,
      "total_flos": 0.0,
      "train_loss": 0.17478099153630405,
      "train_runtime": 7645.1097,
      "train_samples_per_second": 7.997,
      "train_steps_per_second": 0.063
    }
  ],
  "logging_steps": 10,
  "max_steps": 478,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}