{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.957345971563981,
  "eval_steps": 100,
  "global_step": 156,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 265.9657020142556,
      "learning_rate": 3.125e-08,
      "logits/chosen": 123.11854553222656,
      "logits/rejected": 97.00198364257812,
      "logps/chosen": -425.18585205078125,
      "logps/rejected": -424.1869201660156,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.19,
      "grad_norm": 372.96475604860683,
      "learning_rate": 3.1249999999999997e-07,
      "logits/chosen": 117.39414978027344,
      "logits/rejected": 136.32305908203125,
      "logps/chosen": -442.28045654296875,
      "logps/rejected": -524.1576538085938,
      "loss": 0.7748,
      "rewards/accuracies": 0.4722222089767456,
      "rewards/chosen": 0.1107947826385498,
      "rewards/margins": 0.115071140229702,
      "rewards/rejected": -0.004276359919458628,
      "step": 10
    },
    {
      "epoch": 0.38,
      "grad_norm": 213.17360776034892,
      "learning_rate": 4.989935734988097e-07,
      "logits/chosen": 125.45426177978516,
      "logits/rejected": 133.0935821533203,
      "logps/chosen": -426.3124084472656,
      "logps/rejected": -493.6737365722656,
      "loss": 0.6502,
      "rewards/accuracies": 0.625,
      "rewards/chosen": -0.004780772142112255,
      "rewards/margins": 0.6115384101867676,
      "rewards/rejected": -0.6163192987442017,
      "step": 20
    },
    {
      "epoch": 0.57,
      "grad_norm": 204.32015271324505,
      "learning_rate": 4.877641290737883e-07,
      "logits/chosen": 123.5244369506836,
      "logits/rejected": 126.99784851074219,
      "logps/chosen": -466.4947814941406,
      "logps/rejected": -534.4097900390625,
      "loss": 0.6377,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -1.137466549873352,
      "rewards/margins": 1.5820324420928955,
      "rewards/rejected": -2.719499111175537,
      "step": 30
    },
    {
      "epoch": 0.76,
      "grad_norm": 200.5277184123756,
      "learning_rate": 4.646121984004665e-07,
      "logits/chosen": 126.0523910522461,
      "logits/rejected": 120.98759460449219,
      "logps/chosen": -489.78033447265625,
      "logps/rejected": -514.3892211914062,
      "loss": 0.6145,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -1.659891128540039,
      "rewards/margins": 1.3025611639022827,
      "rewards/rejected": -2.9624521732330322,
      "step": 40
    },
    {
      "epoch": 0.95,
      "grad_norm": 205.88684610219255,
      "learning_rate": 4.3069871595684787e-07,
      "logits/chosen": 134.4938201904297,
      "logits/rejected": 134.79849243164062,
      "logps/chosen": -502.4814453125,
      "logps/rejected": -523.4627685546875,
      "loss": 0.5901,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": -1.9658725261688232,
      "rewards/margins": 1.2555662393569946,
      "rewards/rejected": -3.2214386463165283,
      "step": 50
    },
    {
      "epoch": 1.14,
      "grad_norm": 87.33372866728267,
      "learning_rate": 3.877242453630256e-07,
      "logits/chosen": 133.0808563232422,
      "logits/rejected": 136.4200897216797,
      "logps/chosen": -467.0462951660156,
      "logps/rejected": -506.5567321777344,
      "loss": 0.2517,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -1.0280801057815552,
      "rewards/margins": 3.1976516246795654,
      "rewards/rejected": -4.22573184967041,
      "step": 60
    },
    {
      "epoch": 1.33,
      "grad_norm": 71.00765502932312,
      "learning_rate": 3.378437060203357e-07,
      "logits/chosen": 128.8972625732422,
      "logits/rejected": 129.42971801757812,
      "logps/chosen": -440.6957092285156,
      "logps/rejected": -552.7991943359375,
      "loss": 0.1231,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": -1.1781926155090332,
      "rewards/margins": 3.8879973888397217,
      "rewards/rejected": -5.066189765930176,
      "step": 70
    },
    {
      "epoch": 1.52,
      "grad_norm": 68.54852270837435,
      "learning_rate": 2.8355831645441387e-07,
      "logits/chosen": 133.30160522460938,
      "logits/rejected": 134.1454620361328,
      "logps/chosen": -499.032958984375,
      "logps/rejected": -585.234130859375,
      "loss": 0.1193,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.9932767748832703,
      "rewards/margins": 4.856534481048584,
      "rewards/rejected": -5.849810600280762,
      "step": 80
    },
    {
      "epoch": 1.71,
      "grad_norm": 96.90359229935069,
      "learning_rate": 2.2759017277414164e-07,
      "logits/chosen": 120.51957702636719,
      "logits/rejected": 123.01716613769531,
      "logps/chosen": -477.44366455078125,
      "logps/rejected": -548.8724365234375,
      "loss": 0.1182,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -1.5276060104370117,
      "rewards/margins": 4.922442436218262,
      "rewards/rejected": -6.450047969818115,
      "step": 90
    },
    {
      "epoch": 1.9,
      "grad_norm": 57.0241100183833,
      "learning_rate": 1.7274575140626315e-07,
      "logits/chosen": 134.63583374023438,
      "logits/rejected": 125.1494140625,
      "logps/chosen": -492.84930419921875,
      "logps/rejected": -590.646728515625,
      "loss": 0.1371,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": -1.4459283351898193,
      "rewards/margins": 5.340741157531738,
      "rewards/rejected": -6.7866692543029785,
      "step": 100
    },
    {
      "epoch": 1.9,
      "eval_logits/chosen": 102.37699127197266,
      "eval_logits/rejected": 96.37848663330078,
      "eval_logps/chosen": -483.2002868652344,
      "eval_logps/rejected": -502.7766418457031,
      "eval_loss": 0.5052046775817871,
      "eval_rewards/accuracies": 0.7708333134651184,
      "eval_rewards/chosen": -2.9795796871185303,
      "eval_rewards/margins": 2.143129587173462,
      "eval_rewards/rejected": -5.122709274291992,
      "eval_runtime": 48.3488,
      "eval_samples_per_second": 15.512,
      "eval_steps_per_second": 0.496,
      "step": 100
    },
    {
      "epoch": 2.09,
      "grad_norm": 49.58918461406247,
      "learning_rate": 1.2177518064852348e-07,
      "logits/chosen": 113.58479309082031,
      "logits/rejected": 127.19537353515625,
      "logps/chosen": -511.4662170410156,
      "logps/rejected": -620.75537109375,
      "loss": 0.0929,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -1.6378005743026733,
      "rewards/margins": 5.002486228942871,
      "rewards/rejected": -6.640286922454834,
      "step": 110
    },
    {
      "epoch": 2.27,
      "grad_norm": 30.33881666826054,
      "learning_rate": 7.723433775328384e-08,
      "logits/chosen": 124.79753112792969,
      "logits/rejected": 125.82017517089844,
      "logps/chosen": -493.6553649902344,
      "logps/rejected": -574.6585693359375,
      "loss": 0.053,
      "rewards/accuracies": 0.9937499761581421,
      "rewards/chosen": -1.6378300189971924,
      "rewards/margins": 5.504368305206299,
      "rewards/rejected": -7.142198085784912,
      "step": 120
    },
    {
      "epoch": 2.46,
      "grad_norm": 23.06668849787155,
      "learning_rate": 4.1356686569674335e-08,
      "logits/chosen": 129.1642608642578,
      "logits/rejected": 134.2584228515625,
      "logps/chosen": -508.4425354003906,
      "logps/rejected": -594.6097412109375,
      "loss": 0.0541,
      "rewards/accuracies": 0.9937499761581421,
      "rewards/chosen": -1.4683927297592163,
      "rewards/margins": 5.420845985412598,
      "rewards/rejected": -6.8892388343811035,
      "step": 130
    },
    {
      "epoch": 2.65,
      "grad_norm": 38.07392153670512,
      "learning_rate": 1.5941282340065697e-08,
      "logits/chosen": 114.87281799316406,
      "logits/rejected": 115.1664047241211,
      "logps/chosen": -465.3794860839844,
      "logps/rejected": -584.0101318359375,
      "loss": 0.0538,
      "rewards/accuracies": 1.0,
      "rewards/chosen": -1.9863094091415405,
      "rewards/margins": 5.705286979675293,
      "rewards/rejected": -7.691596984863281,
      "step": 140
    },
    {
      "epoch": 2.84,
      "grad_norm": 46.62238024751895,
      "learning_rate": 2.2625595580163247e-09,
      "logits/chosen": 121.98884582519531,
      "logits/rejected": 135.7744598388672,
      "logps/chosen": -492.438720703125,
      "logps/rejected": -580.8795166015625,
      "loss": 0.0589,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -1.7609163522720337,
      "rewards/margins": 5.512405872344971,
      "rewards/rejected": -7.273321628570557,
      "step": 150
    },
    {
      "epoch": 2.96,
      "step": 156,
      "total_flos": 0.0,
      "train_loss": 0.27917395799587935,
      "train_runtime": 1799.8719,
      "train_samples_per_second": 11.251,
      "train_steps_per_second": 0.087
    }
  ],
  "logging_steps": 10,
  "max_steps": 156,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}