{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 328,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 5.561485449698178,
      "learning_rate": 1.5151515151515152e-07,
      "logits/chosen": -0.4639463722705841,
      "logits/rejected": -0.17450889945030212,
      "logps/chosen": -612.9666748046875,
      "logps/rejected": -1227.4912109375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 5.576753539954733,
      "learning_rate": 1.5151515151515152e-06,
      "logits/chosen": -0.41454941034317017,
      "logits/rejected": -0.577505350112915,
      "logps/chosen": -869.0648803710938,
      "logps/rejected": -1276.6817626953125,
      "loss": 0.6913,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.0039122202433645725,
      "rewards/margins": 0.003580428659915924,
      "rewards/rejected": -0.007492648903280497,
      "step": 10
    },
    {
      "epoch": 0.06,
      "grad_norm": 6.198667632843589,
      "learning_rate": 3.0303030303030305e-06,
      "logits/chosen": -0.3819029927253723,
      "logits/rejected": -0.46113020181655884,
      "logps/chosen": -997.2161865234375,
      "logps/rejected": -1214.39990234375,
      "loss": 0.6612,
      "rewards/accuracies": 0.643750011920929,
      "rewards/chosen": -0.15533407032489777,
      "rewards/margins": 0.047475699335336685,
      "rewards/rejected": -0.20280976593494415,
      "step": 20
    },
    {
      "epoch": 0.09,
      "grad_norm": 5.889725033116788,
      "learning_rate": 4.5454545454545455e-06,
      "logits/chosen": -0.24505889415740967,
      "logits/rejected": -0.24203690886497498,
      "logps/chosen": -911.0994873046875,
      "logps/rejected": -1363.7496337890625,
      "loss": 0.541,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.39497110247612,
      "rewards/margins": 0.5585557222366333,
      "rewards/rejected": -0.9535268545150757,
      "step": 30
    },
    {
      "epoch": 0.12,
      "grad_norm": 4.086637897535325,
      "learning_rate": 4.9930567839810125e-06,
      "logits/chosen": -0.17689356207847595,
      "logits/rejected": -0.16114920377731323,
      "logps/chosen": -1011.1546630859375,
      "logps/rejected": -1480.3541259765625,
      "loss": 0.4484,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.3060458302497864,
      "rewards/margins": 1.0700631141662598,
      "rewards/rejected": -1.3761091232299805,
      "step": 40
    },
    {
      "epoch": 0.15,
      "grad_norm": 4.522119109661391,
      "learning_rate": 4.959142005221991e-06,
      "logits/chosen": -0.11198903620243073,
      "logits/rejected": -0.15591135621070862,
      "logps/chosen": -899.4259033203125,
      "logps/rejected": -1485.4820556640625,
      "loss": 0.4169,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.2616763412952423,
      "rewards/margins": 1.9235137701034546,
      "rewards/rejected": -2.185190439224243,
      "step": 50
    },
    {
      "epoch": 0.18,
      "grad_norm": 5.555413111169151,
      "learning_rate": 4.897364164920515e-06,
      "logits/chosen": -0.1413438618183136,
      "logits/rejected": -0.248577281832695,
      "logps/chosen": -856.3390502929688,
      "logps/rejected": -1524.873291015625,
      "loss": 0.3649,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.2767874300479889,
      "rewards/margins": 1.9685882329940796,
      "rewards/rejected": -2.245375871658325,
      "step": 60
    },
    {
      "epoch": 0.21,
      "grad_norm": 5.814409469094451,
      "learning_rate": 4.808423230692374e-06,
      "logits/chosen": -0.23499484360218048,
      "logits/rejected": -0.3045286536216736,
      "logps/chosen": -930.0250854492188,
      "logps/rejected": -1563.339599609375,
      "loss": 0.3102,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -0.22926001250743866,
      "rewards/margins": 2.3029136657714844,
      "rewards/rejected": -2.5321736335754395,
      "step": 70
    },
    {
      "epoch": 0.24,
      "grad_norm": 7.080973924268642,
      "learning_rate": 4.693326938861367e-06,
      "logits/chosen": -0.19219891726970673,
      "logits/rejected": -0.29335862398147583,
      "logps/chosen": -1067.5966796875,
      "logps/rejected": -1646.7294921875,
      "loss": 0.3024,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.3816714882850647,
      "rewards/margins": 2.590569019317627,
      "rewards/rejected": -2.972240447998047,
      "step": 80
    },
    {
      "epoch": 0.27,
      "grad_norm": 4.198144592019097,
      "learning_rate": 4.553379376404085e-06,
      "logits/chosen": -0.16756731271743774,
      "logits/rejected": -0.2766783535480499,
      "logps/chosen": -901.4991455078125,
      "logps/rejected": -1794.3072509765625,
      "loss": 0.249,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.32676273584365845,
      "rewards/margins": 4.70559549331665,
      "rewards/rejected": -5.032357215881348,
      "step": 90
    },
    {
      "epoch": 0.3,
      "grad_norm": 3.514239110297717,
      "learning_rate": 4.3901662051233755e-06,
      "logits/chosen": -0.13113151490688324,
      "logits/rejected": -0.23850134015083313,
      "logps/chosen": -943.93701171875,
      "logps/rejected": -1710.283203125,
      "loss": 0.2635,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": -0.5341259837150574,
      "rewards/margins": 4.174313545227051,
      "rewards/rejected": -4.708439826965332,
      "step": 100
    },
    {
      "epoch": 0.3,
      "eval_logits/chosen": -0.3091484606266022,
      "eval_logits/rejected": -0.10848761349916458,
      "eval_logps/chosen": -380.92041015625,
      "eval_logps/rejected": -1646.8138427734375,
      "eval_loss": 0.199021577835083,
      "eval_rewards/accuracies": 0.9166666865348816,
      "eval_rewards/chosen": -0.5114283561706543,
      "eval_rewards/margins": 9.306483268737793,
      "eval_rewards/rejected": -9.817912101745605,
      "eval_runtime": 29.898,
      "eval_samples_per_second": 6.422,
      "eval_steps_per_second": 0.201,
      "step": 100
    },
    {
      "epoch": 0.34,
      "grad_norm": 2.7779512658548504,
      "learning_rate": 4.205536695466524e-06,
      "logits/chosen": -0.13202831149101257,
      "logits/rejected": -0.31658726930618286,
      "logps/chosen": -881.8616333007812,
      "logps/rejected": -1843.311767578125,
      "loss": 0.2265,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.4379568099975586,
      "rewards/margins": 4.332350254058838,
      "rewards/rejected": -4.7703070640563965,
      "step": 110
    },
    {
      "epoch": 0.37,
      "grad_norm": 3.058731694797242,
      "learning_rate": 4.001582773552153e-06,
      "logits/chosen": -0.2267748862504959,
      "logits/rejected": -0.37614327669143677,
      "logps/chosen": -1021.9287109375,
      "logps/rejected": -1790.741455078125,
      "loss": 0.2448,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.7997490167617798,
      "rewards/margins": 3.615135908126831,
      "rewards/rejected": -4.4148850440979,
      "step": 120
    },
    {
      "epoch": 0.4,
      "grad_norm": 3.5646553088179957,
      "learning_rate": 3.7806153188114027e-06,
      "logits/chosen": -0.14952275156974792,
      "logits/rejected": -0.26807302236557007,
      "logps/chosen": -988.8695068359375,
      "logps/rejected": -1687.877197265625,
      "loss": 0.199,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.5245489478111267,
      "rewards/margins": 2.954119920730591,
      "rewards/rejected": -3.4786689281463623,
      "step": 130
    },
    {
      "epoch": 0.43,
      "grad_norm": 4.994600163477232,
      "learning_rate": 3.5451379808006014e-06,
      "logits/chosen": -0.1331457495689392,
      "logits/rejected": -0.23687326908111572,
      "logps/chosen": -885.8282470703125,
      "logps/rejected": -1685.4974365234375,
      "loss": 0.2011,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -0.609821617603302,
      "rewards/margins": 3.710085391998291,
      "rewards/rejected": -4.319907188415527,
      "step": 140
    },
    {
      "epoch": 0.46,
      "grad_norm": 4.000160476158621,
      "learning_rate": 3.2978188118513814e-06,
      "logits/chosen": -0.13606548309326172,
      "logits/rejected": -0.32709023356437683,
      "logps/chosen": -931.5257568359375,
      "logps/rejected": -1801.125,
      "loss": 0.1791,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -0.4420163631439209,
      "rewards/margins": 4.814242362976074,
      "rewards/rejected": -5.256258964538574,
      "step": 150
    },
    {
      "epoch": 0.49,
      "grad_norm": 5.561015522393931,
      "learning_rate": 3.041460036971664e-06,
      "logits/chosen": -0.20122027397155762,
      "logits/rejected": -0.37438589334487915,
      "logps/chosen": -960.2258911132812,
      "logps/rejected": -1840.1077880859375,
      "loss": 0.1646,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.548378050327301,
      "rewards/margins": 5.22580623626709,
      "rewards/rejected": -5.774184703826904,
      "step": 160
    },
    {
      "epoch": 0.52,
      "grad_norm": 3.0160283224172506,
      "learning_rate": 2.7789663035166035e-06,
      "logits/chosen": -0.2793598771095276,
      "logits/rejected": -0.4746064245700836,
      "logps/chosen": -1049.3095703125,
      "logps/rejected": -1993.3609619140625,
      "loss": 0.1263,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -0.5465644598007202,
      "rewards/margins": 6.312196731567383,
      "rewards/rejected": -6.858761787414551,
      "step": 170
    },
    {
      "epoch": 0.55,
      "grad_norm": 2.93594175440836,
      "learning_rate": 2.513311770373421e-06,
      "logits/chosen": -0.31422796845436096,
      "logits/rejected": -0.5171582102775574,
      "logps/chosen": -960.77978515625,
      "logps/rejected": -1946.6195068359375,
      "loss": 0.1157,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.6941060423851013,
      "rewards/margins": 5.883716583251953,
      "rewards/rejected": -6.577822685241699,
      "step": 180
    },
    {
      "epoch": 0.58,
      "grad_norm": 3.989156252845766,
      "learning_rate": 2.247506409552795e-06,
      "logits/chosen": -0.3567090332508087,
      "logits/rejected": -0.5727603435516357,
      "logps/chosen": -1111.8785400390625,
      "logps/rejected": -2083.59716796875,
      "loss": 0.1418,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.5756088495254517,
      "rewards/margins": 6.701145172119141,
      "rewards/rejected": -7.276753902435303,
      "step": 190
    },
    {
      "epoch": 0.61,
      "grad_norm": 3.220493331131803,
      "learning_rate": 1.9845619020032552e-06,
      "logits/chosen": -0.4283437132835388,
      "logits/rejected": -0.6103153228759766,
      "logps/chosen": -977.1068115234375,
      "logps/rejected": -1989.1517333984375,
      "loss": 0.1415,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.26823538541793823,
      "rewards/margins": 6.212519645690918,
      "rewards/rejected": -6.480755805969238,
      "step": 200
    },
    {
      "epoch": 0.61,
      "eval_logits/chosen": -0.48242926597595215,
      "eval_logits/rejected": -0.42804303765296936,
      "eval_logps/chosen": -420.5835876464844,
      "eval_logps/rejected": -1771.5301513671875,
      "eval_loss": 0.15017390251159668,
      "eval_rewards/accuracies": 0.9166666865348816,
      "eval_rewards/chosen": -0.9080603122711182,
      "eval_rewards/margins": 10.157015800476074,
      "eval_rewards/rejected": -11.065074920654297,
      "eval_runtime": 29.9491,
      "eval_samples_per_second": 6.411,
      "eval_steps_per_second": 0.2,
      "step": 200
    },
    {
      "epoch": 0.64,
      "grad_norm": 5.03992800632795,
      "learning_rate": 1.7274575140626318e-06,
      "logits/chosen": -0.40144267678260803,
      "logits/rejected": -0.6470441818237305,
      "logps/chosen": -992.0281372070312,
      "logps/rejected": -1956.958984375,
      "loss": 0.1581,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -0.683698832988739,
      "rewards/margins": 6.963919162750244,
      "rewards/rejected": -7.647618293762207,
      "step": 210
    },
    {
      "epoch": 0.67,
      "grad_norm": 2.5930618000995724,
      "learning_rate": 1.4791063411799938e-06,
      "logits/chosen": -0.4398053288459778,
      "logits/rejected": -0.6155021786689758,
      "logps/chosen": -901.8489990234375,
      "logps/rejected": -1997.6116943359375,
      "loss": 0.1258,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.4012775421142578,
      "rewards/margins": 6.776458740234375,
      "rewards/rejected": -7.177737236022949,
      "step": 220
    },
    {
      "epoch": 0.7,
      "grad_norm": 3.126671458993684,
      "learning_rate": 1.2423223013801946e-06,
      "logits/chosen": -0.4161323606967926,
      "logits/rejected": -0.6325975656509399,
      "logps/chosen": -993.7923583984375,
      "logps/rejected": -2073.91064453125,
      "loss": 0.1237,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.23834288120269775,
      "rewards/margins": 6.739261627197266,
      "rewards/rejected": -6.977604866027832,
      "step": 230
    },
    {
      "epoch": 0.73,
      "grad_norm": 3.053249955458482,
      "learning_rate": 1.019788252448267e-06,
      "logits/chosen": -0.41894397139549255,
      "logits/rejected": -0.5673054456710815,
      "logps/chosen": -946.7174682617188,
      "logps/rejected": -2171.60009765625,
      "loss": 0.0947,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": -0.34920141100883484,
      "rewards/margins": 8.334190368652344,
      "rewards/rejected": -8.683391571044922,
      "step": 240
    },
    {
      "epoch": 0.76,
      "grad_norm": 3.2492894915267074,
      "learning_rate": 8.140255940787059e-07,
      "logits/chosen": -0.4345846176147461,
      "logits/rejected": -0.5576716661453247,
      "logps/chosen": -990.5559692382812,
      "logps/rejected": -2134.9990234375,
      "loss": 0.1116,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.7247036695480347,
      "rewards/margins": 7.789909362792969,
      "rewards/rejected": -8.514613151550293,
      "step": 250
    },
    {
      "epoch": 0.79,
      "grad_norm": 2.064615334983041,
      "learning_rate": 6.273656994094232e-07,
      "logits/chosen": -0.4094814658164978,
      "logits/rejected": -0.6032238006591797,
      "logps/chosen": -973.67041015625,
      "logps/rejected": -2180.19677734375,
      "loss": 0.1092,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.7021085023880005,
      "rewards/margins": 8.40003776550293,
      "rewards/rejected": -9.102147102355957,
      "step": 260
    },
    {
      "epoch": 0.82,
      "grad_norm": 2.2441288839699647,
      "learning_rate": 4.619234996325314e-07,
      "logits/chosen": -0.4224594235420227,
      "logits/rejected": -0.5946496725082397,
      "logps/chosen": -1018.1749267578125,
      "logps/rejected": -2218.16552734375,
      "loss": 0.0905,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": -0.5338882803916931,
      "rewards/margins": 8.705224990844727,
      "rewards/rejected": -9.23911190032959,
      "step": 270
    },
    {
      "epoch": 0.85,
      "grad_norm": 4.0256681132412035,
      "learning_rate": 3.195735209788528e-07,
      "logits/chosen": -0.37011057138442993,
      "logits/rejected": -0.6081669330596924,
      "logps/chosen": -967.2252197265625,
      "logps/rejected": -2145.68798828125,
      "loss": 0.1005,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.7551916837692261,
      "rewards/margins": 7.838353157043457,
      "rewards/rejected": -8.593544960021973,
      "step": 280
    },
    {
      "epoch": 0.88,
      "grad_norm": 4.259685670607253,
      "learning_rate": 2.019286455866981e-07,
      "logits/chosen": -0.41176638007164,
      "logits/rejected": -0.5948176980018616,
      "logps/chosen": -1099.498779296875,
      "logps/rejected": -2072.706298828125,
      "loss": 0.1126,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -0.7403249740600586,
      "rewards/margins": 6.6156768798828125,
      "rewards/rejected": -7.356001853942871,
      "step": 290
    },
    {
      "epoch": 0.91,
      "grad_norm": 2.837717904545823,
      "learning_rate": 1.1032183690276754e-07,
      "logits/chosen": -0.38978689908981323,
      "logits/rejected": -0.554754376411438,
      "logps/chosen": -1006.9293212890625,
      "logps/rejected": -2237.94970703125,
      "loss": 0.0892,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.7637746334075928,
      "rewards/margins": 8.33720874786377,
      "rewards/rejected": -9.100981712341309,
      "step": 300
    },
    {
      "epoch": 0.91,
      "eval_logits/chosen": -0.48870301246643066,
      "eval_logits/rejected": -0.40482982993125916,
      "eval_logps/chosen": -455.8128662109375,
      "eval_logps/rejected": -1987.8828125,
      "eval_loss": 0.14211897552013397,
      "eval_rewards/accuracies": 0.9166666865348816,
      "eval_rewards/chosen": -1.2603524923324585,
      "eval_rewards/margins": 11.968250274658203,
      "eval_rewards/rejected": -13.228604316711426,
      "eval_runtime": 29.6183,
      "eval_samples_per_second": 6.482,
      "eval_steps_per_second": 0.203,
      "step": 300
    },
    {
      "epoch": 0.95,
      "grad_norm": 2.6436300030745867,
      "learning_rate": 4.579103667367385e-08,
      "logits/chosen": -0.4472607672214508,
      "logits/rejected": -0.6000027656555176,
      "logps/chosen": -1035.2442626953125,
      "logps/rejected": -2228.332275390625,
      "loss": 0.1212,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -0.6586896777153015,
      "rewards/margins": 9.017664909362793,
      "rewards/rejected": -9.676355361938477,
      "step": 310
    },
    {
      "epoch": 0.98,
      "grad_norm": 4.053048154617744,
      "learning_rate": 9.067404651211808e-09,
      "logits/chosen": -0.42106541991233826,
      "logits/rejected": -0.5793359279632568,
      "logps/chosen": -992.0040283203125,
      "logps/rejected": -2126.542724609375,
      "loss": 0.0966,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.767730176448822,
      "rewards/margins": 8.566682815551758,
      "rewards/rejected": -9.334413528442383,
      "step": 320
    },
    {
      "epoch": 1.0,
      "step": 328,
      "total_flos": 0.0,
      "train_loss": 0.22599055091055428,
      "train_runtime": 7017.6011,
      "train_samples_per_second": 2.987,
      "train_steps_per_second": 0.047
    }
  ],
  "logging_steps": 10,
  "max_steps": 328,
  "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
}