{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 939,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.003194888178913738,
      "grad_norm": 106.83801953543872,
      "learning_rate": 5.3191489361702125e-09,
      "logits/chosen": -3.578125,
      "logits/rejected": -3.421875,
      "logps/chosen": -282.0,
      "logps/rejected": -290.0,
      "loss": 0.6914,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03194888178913738,
      "grad_norm": 98.60082387787162,
      "learning_rate": 5.3191489361702123e-08,
      "logits/chosen": -3.484375,
      "logits/rejected": -3.390625,
      "logps/chosen": -274.0,
      "logps/rejected": -278.0,
      "loss": 0.6868,
      "rewards/accuracies": 0.1388888955116272,
      "rewards/chosen": -0.00970458984375,
      "rewards/margins": 6.794929504394531e-06,
      "rewards/rejected": -0.009765625,
      "step": 10
    },
    {
      "epoch": 0.06389776357827476,
      "grad_norm": 116.57506460735225,
      "learning_rate": 1.0638297872340425e-07,
      "logits/chosen": -3.421875,
      "logits/rejected": -3.421875,
      "logps/chosen": -280.0,
      "logps/rejected": -266.0,
      "loss": 0.6537,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": -0.03125,
      "rewards/margins": 0.12158203125,
      "rewards/rejected": -0.15234375,
      "step": 20
    },
    {
      "epoch": 0.09584664536741214,
      "grad_norm": 78.04564900745318,
      "learning_rate": 1.5957446808510638e-07,
      "logits/chosen": -3.34375,
      "logits/rejected": -3.484375,
      "logps/chosen": -276.0,
      "logps/rejected": -268.0,
      "loss": 0.6477,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.1552734375,
      "rewards/margins": 0.1826171875,
      "rewards/rejected": -0.337890625,
      "step": 30
    },
    {
      "epoch": 0.12779552715654952,
      "grad_norm": 104.09831779598169,
      "learning_rate": 2.127659574468085e-07,
      "logits/chosen": -3.46875,
      "logits/rejected": -3.421875,
      "logps/chosen": -272.0,
      "logps/rejected": -276.0,
      "loss": 0.5849,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.1923828125,
      "rewards/margins": 0.341796875,
      "rewards/rejected": -0.53515625,
      "step": 40
    },
    {
      "epoch": 0.1597444089456869,
      "grad_norm": 83.06087118508516,
      "learning_rate": 2.659574468085106e-07,
      "logits/chosen": -3.375,
      "logits/rejected": -3.40625,
      "logps/chosen": -278.0,
      "logps/rejected": -274.0,
      "loss": 0.6062,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.34375,
      "rewards/margins": 0.244140625,
      "rewards/rejected": -0.5859375,
      "step": 50
    },
    {
      "epoch": 0.19169329073482427,
      "grad_norm": 85.2326419243345,
      "learning_rate": 3.1914893617021275e-07,
      "logits/chosen": -3.28125,
      "logits/rejected": -3.3125,
      "logps/chosen": -280.0,
      "logps/rejected": -276.0,
      "loss": 0.5718,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.314453125,
      "rewards/margins": 0.5546875,
      "rewards/rejected": -0.87109375,
      "step": 60
    },
    {
      "epoch": 0.22364217252396165,
      "grad_norm": 85.35363076925235,
      "learning_rate": 3.7234042553191484e-07,
      "logits/chosen": -3.28125,
      "logits/rejected": -3.296875,
      "logps/chosen": -278.0,
      "logps/rejected": -278.0,
      "loss": 0.5627,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.244140625,
      "rewards/margins": 0.451171875,
      "rewards/rejected": -0.6953125,
      "step": 70
    },
    {
      "epoch": 0.25559105431309903,
      "grad_norm": 80.13630527816748,
      "learning_rate": 4.25531914893617e-07,
      "logits/chosen": -3.328125,
      "logits/rejected": -3.296875,
      "logps/chosen": -280.0,
      "logps/rejected": -274.0,
      "loss": 0.5984,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.3203125,
      "rewards/margins": 0.546875,
      "rewards/rejected": -0.8671875,
      "step": 80
    },
    {
      "epoch": 0.28753993610223644,
      "grad_norm": 64.74916812090929,
      "learning_rate": 4.787234042553192e-07,
      "logits/chosen": -3.375,
      "logits/rejected": -3.296875,
      "logps/chosen": -274.0,
      "logps/rejected": -278.0,
      "loss": 0.5594,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.224609375,
      "rewards/margins": 0.69921875,
      "rewards/rejected": -0.921875,
      "step": 90
    },
    {
      "epoch": 0.3194888178913738,
      "grad_norm": 62.874343395319244,
      "learning_rate": 4.964497041420119e-07,
      "logits/chosen": -3.34375,
      "logits/rejected": -3.34375,
      "logps/chosen": -280.0,
      "logps/rejected": -268.0,
      "loss": 0.5807,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.3203125,
      "rewards/margins": 0.55078125,
      "rewards/rejected": -0.87109375,
      "step": 100
    },
    {
      "epoch": 0.3514376996805112,
      "grad_norm": 65.70150692167816,
      "learning_rate": 4.905325443786982e-07,
      "logits/chosen": -3.28125,
      "logits/rejected": -3.328125,
      "logps/chosen": -276.0,
      "logps/rejected": -286.0,
      "loss": 0.5878,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.1162109375,
      "rewards/margins": 0.75,
      "rewards/rejected": -0.8671875,
      "step": 110
    },
    {
      "epoch": 0.38338658146964855,
      "grad_norm": 66.45326771352913,
      "learning_rate": 4.846153846153846e-07,
      "logits/chosen": -3.296875,
      "logits/rejected": -3.375,
      "logps/chosen": -280.0,
      "logps/rejected": -276.0,
      "loss": 0.5633,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.1328125,
      "rewards/margins": 0.67578125,
      "rewards/rejected": -0.80859375,
      "step": 120
    },
    {
      "epoch": 0.41533546325878595,
      "grad_norm": 83.00130651053203,
      "learning_rate": 4.78698224852071e-07,
      "logits/chosen": -3.3125,
      "logits/rejected": -3.296875,
      "logps/chosen": -282.0,
      "logps/rejected": -276.0,
      "loss": 0.5831,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.05224609375,
      "rewards/margins": 1.0625,
      "rewards/rejected": -1.0078125,
      "step": 130
    },
    {
      "epoch": 0.4472843450479233,
      "grad_norm": 65.14922163628061,
      "learning_rate": 4.727810650887574e-07,
      "logits/chosen": -3.328125,
      "logits/rejected": -3.34375,
      "logps/chosen": -276.0,
      "logps/rejected": -284.0,
      "loss": 0.5408,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.02001953125,
      "rewards/margins": 0.97265625,
      "rewards/rejected": -0.953125,
      "step": 140
    },
    {
      "epoch": 0.4792332268370607,
      "grad_norm": 69.5805163425697,
      "learning_rate": 4.668639053254438e-07,
      "logits/chosen": -3.3125,
      "logits/rejected": -3.328125,
      "logps/chosen": -266.0,
      "logps/rejected": -278.0,
      "loss": 0.5394,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.047607421875,
      "rewards/margins": 0.890625,
      "rewards/rejected": -0.84375,
      "step": 150
    },
    {
      "epoch": 0.5111821086261981,
      "grad_norm": 67.64968360037635,
      "learning_rate": 4.6094674556213014e-07,
      "logits/chosen": -3.328125,
      "logits/rejected": -3.171875,
      "logps/chosen": -270.0,
      "logps/rejected": -288.0,
      "loss": 0.6143,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.26171875,
      "rewards/margins": 0.83984375,
      "rewards/rejected": -1.1015625,
      "step": 160
    },
    {
      "epoch": 0.5431309904153354,
      "grad_norm": 75.9892459922825,
      "learning_rate": 4.5502958579881655e-07,
      "logits/chosen": -3.265625,
      "logits/rejected": -3.171875,
      "logps/chosen": -278.0,
      "logps/rejected": -290.0,
      "loss": 0.5603,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.44140625,
      "rewards/margins": 0.796875,
      "rewards/rejected": -1.234375,
      "step": 170
    },
    {
      "epoch": 0.5750798722044729,
      "grad_norm": 90.65062968036268,
      "learning_rate": 4.491124260355029e-07,
      "logits/chosen": -3.203125,
      "logits/rejected": -3.1875,
      "logps/chosen": -278.0,
      "logps/rejected": -282.0,
      "loss": 0.6629,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.306640625,
      "rewards/margins": 0.75,
      "rewards/rejected": -1.0546875,
      "step": 180
    },
    {
      "epoch": 0.6070287539936102,
      "grad_norm": 108.28762791772537,
      "learning_rate": 4.4319526627218936e-07,
      "logits/chosen": -3.234375,
      "logits/rejected": -3.28125,
      "logps/chosen": -280.0,
      "logps/rejected": -280.0,
      "loss": 0.5251,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.208984375,
      "rewards/margins": 1.0390625,
      "rewards/rejected": -1.25,
      "step": 190
    },
    {
      "epoch": 0.6389776357827476,
      "grad_norm": 78.98605565885474,
      "learning_rate": 4.372781065088757e-07,
      "logits/chosen": -3.3125,
      "logits/rejected": -3.171875,
      "logps/chosen": -278.0,
      "logps/rejected": -288.0,
      "loss": 0.5118,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.296875,
      "rewards/margins": 1.015625,
      "rewards/rejected": -1.3125,
      "step": 200
    },
    {
      "epoch": 0.670926517571885,
      "grad_norm": 77.91455442318774,
      "learning_rate": 4.313609467455621e-07,
      "logits/chosen": -3.234375,
      "logits/rejected": -3.1875,
      "logps/chosen": -272.0,
      "logps/rejected": -288.0,
      "loss": 0.5157,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.14453125,
      "rewards/margins": 1.015625,
      "rewards/rejected": -1.15625,
      "step": 210
    },
    {
      "epoch": 0.7028753993610224,
      "grad_norm": 63.1172852867484,
      "learning_rate": 4.2544378698224847e-07,
      "logits/chosen": -3.1875,
      "logits/rejected": -3.28125,
      "logps/chosen": -268.0,
      "logps/rejected": -280.0,
      "loss": 0.5155,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -0.0301513671875,
      "rewards/margins": 0.9375,
      "rewards/rejected": -0.96875,
      "step": 220
    },
    {
      "epoch": 0.7348242811501597,
      "grad_norm": 75.1179190067346,
      "learning_rate": 4.195266272189349e-07,
      "logits/chosen": -3.21875,
      "logits/rejected": -3.3125,
      "logps/chosen": -280.0,
      "logps/rejected": -280.0,
      "loss": 0.5115,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.072265625,
      "rewards/margins": 1.1953125,
      "rewards/rejected": -1.265625,
      "step": 230
    },
    {
      "epoch": 0.7667731629392971,
      "grad_norm": 68.71250790991336,
      "learning_rate": 4.1360946745562133e-07,
      "logits/chosen": -3.265625,
      "logits/rejected": -3.203125,
      "logps/chosen": -272.0,
      "logps/rejected": -290.0,
      "loss": 0.4746,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 0.029052734375,
      "rewards/margins": 1.6796875,
      "rewards/rejected": -1.6484375,
      "step": 240
    },
    {
      "epoch": 0.7987220447284346,
      "grad_norm": 55.46878046949881,
      "learning_rate": 4.076923076923077e-07,
      "logits/chosen": -3.265625,
      "logits/rejected": -3.328125,
      "logps/chosen": -280.0,
      "logps/rejected": -278.0,
      "loss": 0.4711,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.150390625,
      "rewards/margins": 1.140625,
      "rewards/rejected": -1.2890625,
      "step": 250
    },
    {
      "epoch": 0.8306709265175719,
      "grad_norm": 60.69436763385765,
      "learning_rate": 4.017751479289941e-07,
      "logits/chosen": -3.25,
      "logits/rejected": -3.28125,
      "logps/chosen": -274.0,
      "logps/rejected": -286.0,
      "loss": 0.5292,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.1650390625,
      "rewards/margins": 1.25,
      "rewards/rejected": -1.4140625,
      "step": 260
    },
    {
      "epoch": 0.8626198083067093,
      "grad_norm": 61.755235473426524,
      "learning_rate": 3.9585798816568044e-07,
      "logits/chosen": -3.28125,
      "logits/rejected": -3.3125,
      "logps/chosen": -274.0,
      "logps/rejected": -282.0,
      "loss": 0.4873,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.044921875,
      "rewards/margins": 1.4375,
      "rewards/rejected": -1.484375,
      "step": 270
    },
    {
      "epoch": 0.8945686900958466,
      "grad_norm": 64.8584152127513,
      "learning_rate": 3.8994082840236685e-07,
      "logits/chosen": -3.328125,
      "logits/rejected": -3.265625,
      "logps/chosen": -284.0,
      "logps/rejected": -280.0,
      "loss": 0.5342,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.1640625,
      "rewards/margins": 0.87109375,
      "rewards/rejected": -1.03125,
      "step": 280
    },
    {
      "epoch": 0.9265175718849841,
      "grad_norm": 71.13313796916569,
      "learning_rate": 3.840236686390532e-07,
      "logits/chosen": -3.296875,
      "logits/rejected": -3.28125,
      "logps/chosen": -268.0,
      "logps/rejected": -284.0,
      "loss": 0.455,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.0546875,
      "rewards/margins": 1.65625,
      "rewards/rejected": -1.7109375,
      "step": 290
    },
    {
      "epoch": 0.9584664536741214,
      "grad_norm": 147.24265223128356,
      "learning_rate": 3.7810650887573966e-07,
      "logits/chosen": -3.296875,
      "logits/rejected": -3.234375,
      "logps/chosen": -284.0,
      "logps/rejected": -284.0,
      "loss": 0.5612,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.27734375,
      "rewards/margins": 1.140625,
      "rewards/rejected": -1.421875,
      "step": 300
    },
    {
      "epoch": 0.9904153354632588,
      "grad_norm": 57.83608472539185,
      "learning_rate": 3.72189349112426e-07,
      "logits/chosen": -3.25,
      "logits/rejected": -3.25,
      "logps/chosen": -274.0,
      "logps/rejected": -282.0,
      "loss": 0.4624,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": 0.05712890625,
      "rewards/margins": 1.40625,
      "rewards/rejected": -1.3515625,
      "step": 310
    },
    {
      "epoch": 1.0,
      "eval_logits/chosen": -3.3125,
      "eval_logits/rejected": -3.34375,
      "eval_logps/chosen": -272.0,
      "eval_logps/rejected": -276.0,
      "eval_loss": 0.5753124952316284,
      "eval_rewards/accuracies": 0.75,
      "eval_rewards/chosen": -0.039306640625,
      "eval_rewards/margins": 0.703125,
      "eval_rewards/rejected": -0.7421875,
      "eval_runtime": 12.4952,
      "eval_samples_per_second": 16.006,
      "eval_steps_per_second": 0.56,
      "step": 313
    },
    {
      "epoch": 1.0223642172523961,
      "grad_norm": 23.67148600061787,
      "learning_rate": 3.662721893491124e-07,
      "logits/chosen": -3.296875,
      "logits/rejected": -3.296875,
      "logps/chosen": -276.0,
      "logps/rejected": -282.0,
      "loss": 0.232,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 0.482421875,
      "rewards/margins": 2.34375,
      "rewards/rejected": -1.859375,
      "step": 320
    },
    {
      "epoch": 1.0543130990415335,
      "grad_norm": 16.419046920806387,
      "learning_rate": 3.603550295857988e-07,
      "logits/chosen": -3.28125,
      "logits/rejected": -3.203125,
      "logps/chosen": -268.0,
      "logps/rejected": -302.0,
      "loss": 0.1279,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.51171875,
      "rewards/margins": 3.375,
      "rewards/rejected": -2.859375,
      "step": 330
    },
    {
      "epoch": 1.0862619808306708,
      "grad_norm": 11.936537156516739,
      "learning_rate": 3.544378698224852e-07,
      "logits/chosen": -3.1875,
      "logits/rejected": -3.203125,
      "logps/chosen": -276.0,
      "logps/rejected": -288.0,
      "loss": 0.1213,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 0.6015625,
      "rewards/margins": 3.171875,
      "rewards/rejected": -2.578125,
      "step": 340
    },
    {
      "epoch": 1.1182108626198084,
      "grad_norm": 20.043411609324348,
      "learning_rate": 3.485207100591716e-07,
      "logits/chosen": -3.125,
      "logits/rejected": -3.15625,
      "logps/chosen": -276.0,
      "logps/rejected": -300.0,
      "loss": 0.1216,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.484375,
      "rewards/margins": 3.40625,
      "rewards/rejected": -2.921875,
      "step": 350
    },
    {
      "epoch": 1.1501597444089458,
      "grad_norm": 36.206031638132856,
      "learning_rate": 3.42603550295858e-07,
      "logits/chosen": -3.171875,
      "logits/rejected": -3.265625,
      "logps/chosen": -266.0,
      "logps/rejected": -298.0,
      "loss": 0.1088,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.4375,
      "rewards/margins": 3.390625,
      "rewards/rejected": -2.953125,
      "step": 360
    },
    {
      "epoch": 1.182108626198083,
      "grad_norm": 30.276531485727194,
      "learning_rate": 3.366863905325444e-07,
      "logits/chosen": -3.28125,
      "logits/rejected": -3.234375,
      "logps/chosen": -270.0,
      "logps/rejected": -300.0,
      "loss": 0.1304,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.51953125,
      "rewards/margins": 3.234375,
      "rewards/rejected": -2.71875,
      "step": 370
    },
    {
      "epoch": 1.2140575079872205,
      "grad_norm": 36.79686407358778,
      "learning_rate": 3.3076923076923075e-07,
      "logits/chosen": -3.171875,
      "logits/rejected": -3.234375,
      "logps/chosen": -274.0,
      "logps/rejected": -288.0,
      "loss": 0.1287,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 0.466796875,
      "rewards/margins": 3.15625,
      "rewards/rejected": -2.6875,
      "step": 380
    },
    {
      "epoch": 1.2460063897763578,
      "grad_norm": 45.670162060448156,
      "learning_rate": 3.2485207100591716e-07,
      "logits/chosen": -3.09375,
      "logits/rejected": -3.265625,
      "logps/chosen": -272.0,
      "logps/rejected": -304.0,
      "loss": 0.1189,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.87109375,
      "rewards/margins": 4.125,
      "rewards/rejected": -3.25,
      "step": 390
    },
    {
      "epoch": 1.2779552715654952,
      "grad_norm": 44.93419834702028,
      "learning_rate": 3.189349112426035e-07,
      "logits/chosen": -3.25,
      "logits/rejected": -3.28125,
      "logps/chosen": -274.0,
      "logps/rejected": -298.0,
      "loss": 0.139,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.63671875,
      "rewards/margins": 3.140625,
      "rewards/rejected": -2.5,
      "step": 400
    },
    {
      "epoch": 1.3099041533546325,
      "grad_norm": 34.60440373360165,
      "learning_rate": 3.130177514792899e-07,
      "logits/chosen": -3.1875,
      "logits/rejected": -3.21875,
      "logps/chosen": -266.0,
      "logps/rejected": -300.0,
      "loss": 0.1653,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.55078125,
      "rewards/margins": 3.375,
      "rewards/rejected": -2.828125,
      "step": 410
    },
    {
      "epoch": 1.34185303514377,
      "grad_norm": 23.20153131751831,
      "learning_rate": 3.071005917159763e-07,
      "logits/chosen": -3.203125,
      "logits/rejected": -3.171875,
      "logps/chosen": -268.0,
      "logps/rejected": -306.0,
      "loss": 0.1342,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.578125,
      "rewards/margins": 3.359375,
      "rewards/rejected": -2.796875,
      "step": 420
    },
    {
      "epoch": 1.3738019169329074,
      "grad_norm": 27.675463987741193,
      "learning_rate": 3.011834319526627e-07,
      "logits/chosen": -3.171875,
      "logits/rejected": -3.15625,
      "logps/chosen": -266.0,
      "logps/rejected": -304.0,
      "loss": 0.1054,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 0.6328125,
      "rewards/margins": 3.71875,
      "rewards/rejected": -3.09375,
      "step": 430
    },
    {
      "epoch": 1.4057507987220448,
      "grad_norm": 40.196786857901714,
      "learning_rate": 2.952662721893491e-07,
      "logits/chosen": -3.15625,
      "logits/rejected": -3.203125,
      "logps/chosen": -270.0,
      "logps/rejected": -292.0,
      "loss": 0.1471,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 0.71875,
      "rewards/margins": 3.21875,
      "rewards/rejected": -2.5,
      "step": 440
    },
    {
      "epoch": 1.4376996805111821,
      "grad_norm": 15.707432621538732,
      "learning_rate": 2.893491124260355e-07,
      "logits/chosen": -3.203125,
      "logits/rejected": -3.203125,
      "logps/chosen": -268.0,
      "logps/rejected": -296.0,
      "loss": 0.093,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.8828125,
      "rewards/margins": 3.59375,
      "rewards/rejected": -2.703125,
      "step": 450
    },
    {
      "epoch": 1.4696485623003195,
      "grad_norm": 28.76561180608655,
      "learning_rate": 2.834319526627219e-07,
      "logits/chosen": -3.1875,
      "logits/rejected": -3.140625,
      "logps/chosen": -264.0,
      "logps/rejected": -300.0,
      "loss": 0.1042,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.6328125,
      "rewards/margins": 3.671875,
      "rewards/rejected": -3.046875,
      "step": 460
    },
    {
      "epoch": 1.5015974440894568,
      "grad_norm": 37.202582805520755,
      "learning_rate": 2.7751479289940824e-07,
      "logits/chosen": -3.15625,
      "logits/rejected": -3.171875,
      "logps/chosen": -268.0,
      "logps/rejected": -302.0,
      "loss": 0.0995,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 0.6796875,
      "rewards/margins": 3.90625,
      "rewards/rejected": -3.21875,
      "step": 470
    },
    {
      "epoch": 1.5335463258785942,
      "grad_norm": 36.12380615792283,
      "learning_rate": 2.715976331360947e-07,
      "logits/chosen": -3.21875,
      "logits/rejected": -3.078125,
      "logps/chosen": -262.0,
      "logps/rejected": -312.0,
      "loss": 0.168,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 0.59765625,
      "rewards/margins": 3.4375,
      "rewards/rejected": -2.828125,
      "step": 480
    },
    {
      "epoch": 1.5654952076677318,
      "grad_norm": 31.441359412649966,
      "learning_rate": 2.6568047337278106e-07,
      "logits/chosen": -3.203125,
      "logits/rejected": -3.171875,
      "logps/chosen": -266.0,
      "logps/rejected": -296.0,
      "loss": 0.1083,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.494140625,
      "rewards/margins": 3.625,
      "rewards/rejected": -3.125,
      "step": 490
    },
    {
      "epoch": 1.5974440894568689,
      "grad_norm": 56.69663253139224,
      "learning_rate": 2.5976331360946746e-07,
      "logits/chosen": -3.125,
      "logits/rejected": -3.125,
      "logps/chosen": -264.0,
      "logps/rejected": -298.0,
      "loss": 0.1199,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.80078125,
      "rewards/margins": 3.859375,
      "rewards/rejected": -3.0625,
      "step": 500
    },
    {
      "epoch": 1.6293929712460065,
      "grad_norm": 42.26084157818348,
      "learning_rate": 2.538461538461538e-07,
      "logits/chosen": -3.140625,
      "logits/rejected": -3.21875,
      "logps/chosen": -268.0,
      "logps/rejected": -284.0,
      "loss": 0.1411,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 0.7421875,
      "rewards/margins": 3.265625,
      "rewards/rejected": -2.515625,
      "step": 510
    },
    {
      "epoch": 1.6613418530351438,
      "grad_norm": 24.290936395597605,
      "learning_rate": 2.479289940828402e-07,
      "logits/chosen": -3.15625,
      "logits/rejected": -3.125,
      "logps/chosen": -264.0,
      "logps/rejected": -294.0,
      "loss": 0.0927,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.74609375,
      "rewards/margins": 3.609375,
      "rewards/rejected": -2.859375,
      "step": 520
    },
    {
      "epoch": 1.6932907348242812,
      "grad_norm": 30.097818527606982,
      "learning_rate": 2.420118343195266e-07,
      "logits/chosen": -3.125,
      "logits/rejected": -3.125,
      "logps/chosen": -268.0,
      "logps/rejected": -294.0,
      "loss": 0.132,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.9140625,
      "rewards/margins": 3.96875,
      "rewards/rejected": -3.0625,
      "step": 530
    },
    {
      "epoch": 1.7252396166134185,
      "grad_norm": 25.610518794255373,
      "learning_rate": 2.36094674556213e-07,
      "logits/chosen": -3.09375,
      "logits/rejected": -3.140625,
      "logps/chosen": -270.0,
      "logps/rejected": -298.0,
      "loss": 0.135,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 0.5234375,
      "rewards/margins": 3.796875,
      "rewards/rejected": -3.28125,
      "step": 540
    },
    {
      "epoch": 1.7571884984025559,
      "grad_norm": 32.86888420257977,
      "learning_rate": 2.301775147928994e-07,
      "logits/chosen": -3.1875,
      "logits/rejected": -3.21875,
      "logps/chosen": -268.0,
      "logps/rejected": -294.0,
      "loss": 0.1477,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 0.921875,
      "rewards/margins": 3.546875,
      "rewards/rejected": -2.625,
      "step": 550
    },
    {
      "epoch": 1.7891373801916934,
      "grad_norm": 38.670593041590834,
      "learning_rate": 2.242603550295858e-07,
      "logits/chosen": -3.09375,
      "logits/rejected": -3.078125,
      "logps/chosen": -272.0,
      "logps/rejected": -300.0,
      "loss": 0.0912,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.66796875,
      "rewards/margins": 3.734375,
      "rewards/rejected": -3.0625,
      "step": 560
    },
    {
      "epoch": 1.8210862619808306,
      "grad_norm": 30.06581434037245,
      "learning_rate": 2.1834319526627217e-07,
      "logits/chosen": -3.125,
      "logits/rejected": -3.109375,
      "logps/chosen": -266.0,
      "logps/rejected": -312.0,
      "loss": 0.1155,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.81640625,
      "rewards/margins": 4.03125,
      "rewards/rejected": -3.21875,
      "step": 570
    },
    {
      "epoch": 1.8530351437699681,
      "grad_norm": 37.09573180706117,
      "learning_rate": 2.1242603550295858e-07,
      "logits/chosen": -3.0625,
      "logits/rejected": -3.0625,
      "logps/chosen": -272.0,
      "logps/rejected": -306.0,
      "loss": 0.1263,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.71484375,
      "rewards/margins": 3.84375,
      "rewards/rejected": -3.125,
      "step": 580
    },
    {
      "epoch": 1.8849840255591053,
      "grad_norm": 38.562399882898944,
      "learning_rate": 2.0650887573964496e-07,
      "logits/chosen": -3.125,
      "logits/rejected": -3.078125,
      "logps/chosen": -258.0,
      "logps/rejected": -302.0,
      "loss": 0.1334,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 0.9921875,
      "rewards/margins": 3.828125,
      "rewards/rejected": -2.84375,
      "step": 590
    },
    {
      "epoch": 1.9169329073482428,
      "grad_norm": 44.104454957376895,
      "learning_rate": 2.0059171597633133e-07,
      "logits/chosen": -3.140625,
      "logits/rejected": -3.1875,
      "logps/chosen": -270.0,
      "logps/rejected": -298.0,
      "loss": 0.1351,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.74609375,
      "rewards/margins": 3.765625,
      "rewards/rejected": -3.015625,
      "step": 600
    },
    {
      "epoch": 1.9488817891373802,
      "grad_norm": 36.65323926007583,
      "learning_rate": 1.9467455621301774e-07,
      "logits/chosen": -3.15625,
      "logits/rejected": -3.21875,
      "logps/chosen": -266.0,
      "logps/rejected": -296.0,
      "loss": 0.1479,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 0.8046875,
      "rewards/margins": 3.578125,
      "rewards/rejected": -2.78125,
      "step": 610
    },
    {
      "epoch": 1.9808306709265175,
      "grad_norm": 15.898706379798815,
      "learning_rate": 1.8875739644970412e-07,
      "logits/chosen": -3.09375,
      "logits/rejected": -3.140625,
      "logps/chosen": -272.0,
      "logps/rejected": -302.0,
      "loss": 0.11,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 0.3515625,
      "rewards/margins": 3.75,
      "rewards/rejected": -3.390625,
      "step": 620
    },
    {
      "epoch": 2.0,
      "eval_logits/chosen": -3.125,
      "eval_logits/rejected": -3.15625,
      "eval_logps/chosen": -278.0,
      "eval_logps/rejected": -282.0,
      "eval_loss": 0.6153125166893005,
      "eval_rewards/accuracies": 0.6428571343421936,
      "eval_rewards/chosen": -0.5234375,
      "eval_rewards/margins": 1.0078125,
      "eval_rewards/rejected": -1.53125,
      "eval_runtime": 12.4752,
      "eval_samples_per_second": 16.032,
      "eval_steps_per_second": 0.561,
      "step": 626
    },
    {
      "epoch": 2.012779552715655,
      "grad_norm": 9.922455136530488,
      "learning_rate": 1.8284023668639053e-07,
      "logits/chosen": -3.21875,
      "logits/rejected": -3.21875,
      "logps/chosen": -264.0,
      "logps/rejected": -300.0,
      "loss": 0.065,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 0.859375,
      "rewards/margins": 4.21875,
      "rewards/rejected": -3.359375,
      "step": 630
    },
    {
      "epoch": 2.0447284345047922,
      "grad_norm": 6.257694062913075,
      "learning_rate": 1.7692307692307693e-07,
      "logits/chosen": -3.15625,
      "logits/rejected": -3.109375,
      "logps/chosen": -268.0,
      "logps/rejected": -312.0,
      "loss": 0.0323,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.9375,
      "rewards/margins": 4.75,
      "rewards/rejected": -3.8125,
      "step": 640
    },
    {
      "epoch": 2.07667731629393,
      "grad_norm": 3.1244935227015445,
      "learning_rate": 1.710059171597633e-07,
      "logits/chosen": -3.09375,
      "logits/rejected": -3.0625,
      "logps/chosen": -258.0,
      "logps/rejected": -318.0,
      "loss": 0.0227,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 1.046875,
      "rewards/margins": 5.09375,
      "rewards/rejected": -4.0625,
      "step": 650
    },
    {
      "epoch": 2.108626198083067,
      "grad_norm": 3.263136591096294,
      "learning_rate": 1.650887573964497e-07,
      "logits/chosen": -3.0,
      "logits/rejected": -3.0625,
      "logps/chosen": -270.0,
      "logps/rejected": -308.0,
      "loss": 0.0312,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 1.03125,
      "rewards/margins": 5.3125,
      "rewards/rejected": -4.28125,
      "step": 660
    },
    {
      "epoch": 2.1405750798722045,
      "grad_norm": 7.503726583629619,
      "learning_rate": 1.591715976331361e-07,
      "logits/chosen": -3.046875,
      "logits/rejected": -3.140625,
      "logps/chosen": -264.0,
      "logps/rejected": -308.0,
      "loss": 0.0211,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.91796875,
      "rewards/margins": 5.0625,
      "rewards/rejected": -4.15625,
      "step": 670
    },
    {
      "epoch": 2.1725239616613417,
      "grad_norm": 4.390111479567184,
      "learning_rate": 1.5325443786982248e-07,
      "logits/chosen": -3.078125,
      "logits/rejected": -3.09375,
      "logps/chosen": -264.0,
      "logps/rejected": -316.0,
      "loss": 0.0207,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.51953125,
      "rewards/margins": 5.15625,
      "rewards/rejected": -4.65625,
      "step": 680
    },
    {
      "epoch": 2.2044728434504792,
      "grad_norm": 7.618343279103693,
      "learning_rate": 1.4733727810650885e-07,
      "logits/chosen": -3.078125,
      "logits/rejected": -3.0,
      "logps/chosen": -270.0,
      "logps/rejected": -316.0,
      "loss": 0.0339,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.65234375,
      "rewards/margins": 5.1875,
      "rewards/rejected": -4.53125,
      "step": 690
    },
    {
      "epoch": 2.236421725239617,
      "grad_norm": 7.443107454169407,
      "learning_rate": 1.4142011834319526e-07,
      "logits/chosen": -2.953125,
      "logits/rejected": -2.984375,
      "logps/chosen": -270.0,
      "logps/rejected": -310.0,
      "loss": 0.0225,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.68359375,
      "rewards/margins": 5.21875,
      "rewards/rejected": -4.5625,
      "step": 700
    },
    {
      "epoch": 2.268370607028754,
      "grad_norm": 30.162725762855914,
      "learning_rate": 1.3550295857988164e-07,
      "logits/chosen": -3.03125,
      "logits/rejected": -2.96875,
      "logps/chosen": -266.0,
      "logps/rejected": -324.0,
      "loss": 0.0314,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.7890625,
      "rewards/margins": 5.4375,
      "rewards/rejected": -4.65625,
      "step": 710
    },
    {
      "epoch": 2.3003194888178915,
      "grad_norm": 11.871816606581527,
      "learning_rate": 1.2958579881656802e-07,
      "logits/chosen": -3.078125,
      "logits/rejected": -3.09375,
      "logps/chosen": -268.0,
      "logps/rejected": -308.0,
      "loss": 0.0281,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.828125,
      "rewards/margins": 5.09375,
      "rewards/rejected": -4.25,
      "step": 720
    },
    {
      "epoch": 2.3322683706070286,
      "grad_norm": 3.7256619480068776,
      "learning_rate": 1.2366863905325443e-07,
      "logits/chosen": -3.03125,
      "logits/rejected": -2.96875,
      "logps/chosen": -264.0,
      "logps/rejected": -324.0,
      "loss": 0.0195,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.61328125,
      "rewards/margins": 5.5,
      "rewards/rejected": -4.875,
      "step": 730
    },
    {
      "epoch": 2.364217252396166,
      "grad_norm": 3.3684622301561875,
      "learning_rate": 1.1775147928994082e-07,
      "logits/chosen": -3.0625,
      "logits/rejected": -2.984375,
      "logps/chosen": -258.0,
      "logps/rejected": -326.0,
      "loss": 0.0201,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.890625,
      "rewards/margins": 5.75,
      "rewards/rejected": -4.875,
      "step": 740
    },
    {
      "epoch": 2.3961661341853033,
      "grad_norm": 5.010929229836647,
      "learning_rate": 1.1183431952662721e-07,
      "logits/chosen": -3.015625,
      "logits/rejected": -2.953125,
      "logps/chosen": -270.0,
      "logps/rejected": -320.0,
      "loss": 0.0296,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.8828125,
      "rewards/margins": 6.0625,
      "rewards/rejected": -5.1875,
      "step": 750
    },
    {
      "epoch": 2.428115015974441,
      "grad_norm": 4.113820701091529,
      "learning_rate": 1.059171597633136e-07,
      "logits/chosen": -3.046875,
      "logits/rejected": -3.03125,
      "logps/chosen": -266.0,
      "logps/rejected": -320.0,
      "loss": 0.013,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.671875,
      "rewards/margins": 5.6875,
      "rewards/rejected": -5.0,
      "step": 760
    },
    {
      "epoch": 2.460063897763578,
      "grad_norm": 20.48578879210455,
      "learning_rate": 1e-07,
      "logits/chosen": -2.984375,
      "logits/rejected": -3.03125,
      "logps/chosen": -274.0,
      "logps/rejected": -320.0,
      "loss": 0.0176,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.578125,
      "rewards/margins": 5.78125,
      "rewards/rejected": -5.21875,
      "step": 770
    },
    {
      "epoch": 2.4920127795527156,
      "grad_norm": 6.654456223269256,
      "learning_rate": 9.408284023668639e-08,
      "logits/chosen": -3.015625,
      "logits/rejected": -2.875,
      "logps/chosen": -272.0,
      "logps/rejected": -328.0,
      "loss": 0.0206,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.0673828125,
      "rewards/margins": 5.5625,
      "rewards/rejected": -5.5,
      "step": 780
    },
    {
      "epoch": 2.523961661341853,
      "grad_norm": 24.63943461801463,
      "learning_rate": 8.816568047337278e-08,
      "logits/chosen": -2.953125,
      "logits/rejected": -3.015625,
      "logps/chosen": -278.0,
      "logps/rejected": -334.0,
      "loss": 0.027,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.0048828125,
      "rewards/margins": 5.9375,
      "rewards/rejected": -5.9375,
      "step": 790
    },
    {
      "epoch": 2.5559105431309903,
      "grad_norm": 12.499286028857556,
      "learning_rate": 8.224852071005916e-08,
      "logits/chosen": -3.0,
      "logits/rejected": -2.96875,
      "logps/chosen": -276.0,
      "logps/rejected": -332.0,
      "loss": 0.0179,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.37109375,
      "rewards/margins": 6.0625,
      "rewards/rejected": -5.71875,
      "step": 800
    },
    {
      "epoch": 2.587859424920128,
      "grad_norm": 14.115073034571477,
      "learning_rate": 7.633136094674555e-08,
      "logits/chosen": -3.078125,
      "logits/rejected": -3.03125,
      "logps/chosen": -262.0,
      "logps/rejected": -312.0,
      "loss": 0.0229,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.59375,
      "rewards/margins": 5.375,
      "rewards/rejected": -4.75,
      "step": 810
    },
    {
      "epoch": 2.619808306709265,
      "grad_norm": 11.454089939459458,
      "learning_rate": 7.041420118343195e-08,
      "logits/chosen": -3.015625,
      "logits/rejected": -3.109375,
      "logps/chosen": -268.0,
      "logps/rejected": -314.0,
      "loss": 0.0225,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.59375,
      "rewards/margins": 5.71875,
      "rewards/rejected": -5.125,
      "step": 820
    },
    {
      "epoch": 2.6517571884984026,
      "grad_norm": 4.235780738481745,
      "learning_rate": 6.449704142011835e-08,
      "logits/chosen": -2.859375,
      "logits/rejected": -2.953125,
      "logps/chosen": -276.0,
      "logps/rejected": -336.0,
      "loss": 0.0266,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.44921875,
      "rewards/margins": 5.9375,
      "rewards/rejected": -5.46875,
      "step": 830
    },
    {
      "epoch": 2.68370607028754,
      "grad_norm": 19.998975812271155,
      "learning_rate": 5.857988165680473e-08,
      "logits/chosen": -2.96875,
      "logits/rejected": -2.96875,
      "logps/chosen": -272.0,
      "logps/rejected": -316.0,
      "loss": 0.0428,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.59765625,
      "rewards/margins": 5.6875,
      "rewards/rejected": -5.09375,
      "step": 840
    },
    {
      "epoch": 2.7156549520766773,
      "grad_norm": 8.111211588972713,
      "learning_rate": 5.266272189349112e-08,
      "logits/chosen": -3.015625,
      "logits/rejected": -2.984375,
      "logps/chosen": -262.0,
      "logps/rejected": -314.0,
      "loss": 0.0163,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.87109375,
      "rewards/margins": 5.71875,
      "rewards/rejected": -4.875,
      "step": 850
    },
    {
      "epoch": 2.747603833865815,
      "grad_norm": 4.3696948627392445,
      "learning_rate": 4.674556213017751e-08,
      "logits/chosen": -3.0,
      "logits/rejected": -2.90625,
      "logps/chosen": -268.0,
      "logps/rejected": -330.0,
      "loss": 0.0171,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.390625,
      "rewards/margins": 6.375,
      "rewards/rejected": -5.96875,
      "step": 860
    },
    {
      "epoch": 2.779552715654952,
      "grad_norm": 9.013340988525037,
      "learning_rate": 4.082840236686391e-08,
      "logits/chosen": -3.03125,
      "logits/rejected": -2.953125,
      "logps/chosen": -276.0,
      "logps/rejected": -324.0,
      "loss": 0.033,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.0693359375,
      "rewards/margins": 5.6875,
      "rewards/rejected": -5.625,
      "step": 870
    },
    {
      "epoch": 2.8115015974440896,
      "grad_norm": 10.6005436998683,
      "learning_rate": 3.4911242603550294e-08,
      "logits/chosen": -2.9375,
      "logits/rejected": -2.9375,
      "logps/chosen": -276.0,
      "logps/rejected": -342.0,
      "loss": 0.0289,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.1806640625,
      "rewards/margins": 6.375,
      "rewards/rejected": -6.1875,
      "step": 880
    },
    {
      "epoch": 2.8434504792332267,
      "grad_norm": 8.619929556764896,
      "learning_rate": 2.8994082840236687e-08,
      "logits/chosen": -2.953125,
      "logits/rejected": -3.0,
      "logps/chosen": -270.0,
      "logps/rejected": -324.0,
      "loss": 0.0273,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.359375,
      "rewards/margins": 6.3125,
      "rewards/rejected": -5.96875,
      "step": 890
    },
    {
      "epoch": 2.8753993610223643,
      "grad_norm": 8.796113060930784,
      "learning_rate": 2.3076923076923076e-08,
      "logits/chosen": -3.015625,
      "logits/rejected": -2.890625,
      "logps/chosen": -268.0,
      "logps/rejected": -330.0,
      "loss": 0.0314,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.087890625,
      "rewards/margins": 5.71875,
      "rewards/rejected": -5.625,
      "step": 900
    },
    {
      "epoch": 2.9073482428115014,
      "grad_norm": 10.748451177210375,
      "learning_rate": 1.7159763313609465e-08,
      "logits/chosen": -2.9375,
      "logits/rejected": -2.859375,
      "logps/chosen": -276.0,
      "logps/rejected": -332.0,
      "loss": 0.0263,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 0.4609375,
      "rewards/margins": 6.25,
      "rewards/rejected": -5.8125,
      "step": 910
    },
    {
      "epoch": 2.939297124600639,
      "grad_norm": 25.184969388385202,
      "learning_rate": 1.1242603550295858e-08,
      "logits/chosen": -3.078125,
      "logits/rejected": -2.90625,
      "logps/chosen": -270.0,
      "logps/rejected": -336.0,
      "loss": 0.0173,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 0.05517578125,
      "rewards/margins": 5.8125,
      "rewards/rejected": -5.75,
      "step": 920
    },
    {
      "epoch": 2.9712460063897765,
      "grad_norm": 18.238796703298632,
      "learning_rate": 5.325443786982248e-09,
      "logits/chosen": -2.96875,
      "logits/rejected": -2.9375,
      "logps/chosen": -270.0,
      "logps/rejected": -326.0,
      "loss": 0.0287,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 0.3515625,
      "rewards/margins": 5.875,
      "rewards/rejected": -5.5,
      "step": 930
    },
    {
      "epoch": 3.0,
      "eval_logits/chosen": -2.984375,
      "eval_logits/rejected": -3.015625,
      "eval_logps/chosen": -286.0,
      "eval_logps/rejected": -296.0,
      "eval_loss": 0.7446093559265137,
      "eval_rewards/accuracies": 0.6964285969734192,
      "eval_rewards/chosen": -1.34375,
      "eval_rewards/margins": 1.40625,
      "eval_rewards/rejected": -2.75,
      "eval_runtime": 15.7054,
      "eval_samples_per_second": 12.734,
      "eval_steps_per_second": 0.446,
      "step": 939
    }
  ],
  "logging_steps": 10,
  "max_steps": 939,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "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": 8,
  "trial_name": null,
  "trial_params": null
}