{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 135,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 6.3079497835343785,
      "learning_rate": 3.571428571428571e-08,
      "logits/chosen": -0.5727702975273132,
      "logits/rejected": -0.3682233691215515,
      "logps/chosen": -365.0997314453125,
      "logps/rejected": -573.4035034179688,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.07,
      "grad_norm": 6.996445402509399,
      "learning_rate": 3.5714285714285716e-07,
      "logits/chosen": -0.6761529445648193,
      "logits/rejected": -0.3929700255393982,
      "logps/chosen": -278.9779968261719,
      "logps/rejected": -682.2371215820312,
      "loss": 0.693,
      "rewards/accuracies": 0.4583333432674408,
      "rewards/chosen": -0.00010682118590921164,
      "rewards/margins": -0.0002022227126872167,
      "rewards/rejected": 9.540154132992029e-05,
      "step": 10
    },
    {
      "epoch": 0.15,
      "grad_norm": 6.50942716442102,
      "learning_rate": 4.969726470454313e-07,
      "logits/chosen": -0.5919066667556763,
      "logits/rejected": -0.32349973917007446,
      "logps/chosen": -302.1080627441406,
      "logps/rejected": -683.5020751953125,
      "loss": 0.6841,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.003587691579014063,
      "rewards/margins": 0.017893124371767044,
      "rewards/rejected": -0.02148081734776497,
      "step": 20
    },
    {
      "epoch": 0.22,
      "grad_norm": 7.122624559821119,
      "learning_rate": 4.787370064564882e-07,
      "logits/chosen": -0.5885516405105591,
      "logits/rejected": -0.3972572088241577,
      "logps/chosen": -299.45013427734375,
      "logps/rejected": -675.0470581054688,
      "loss": 0.6435,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.015654053539037704,
      "rewards/margins": 0.10273133218288422,
      "rewards/rejected": -0.11838539689779282,
      "step": 30
    },
    {
      "epoch": 0.3,
      "grad_norm": 5.610083056347874,
      "learning_rate": 4.45168462792932e-07,
      "logits/chosen": -0.5636590719223022,
      "logits/rejected": -0.4258306622505188,
      "logps/chosen": -372.2354736328125,
      "logps/rejected": -811.8500366210938,
      "loss": 0.5764,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -0.04066535085439682,
      "rewards/margins": 0.3092907965183258,
      "rewards/rejected": -0.3499561846256256,
      "step": 40
    },
    {
      "epoch": 0.37,
      "grad_norm": 5.430631727659141,
      "learning_rate": 3.9851721262034157e-07,
      "logits/chosen": -0.5452407002449036,
      "logits/rejected": -0.4972262978553772,
      "logps/chosen": -366.8607482910156,
      "logps/rejected": -841.2767333984375,
      "loss": 0.4857,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.07431399822235107,
      "rewards/margins": 0.8301405906677246,
      "rewards/rejected": -0.9044545888900757,
      "step": 50
    },
    {
      "epoch": 0.44,
      "grad_norm": 5.19789618190169,
      "learning_rate": 3.4191042415818e-07,
      "logits/chosen": -0.5970432758331299,
      "logits/rejected": -0.5481890439987183,
      "logps/chosen": -290.11712646484375,
      "logps/rejected": -911.9899291992188,
      "loss": 0.4219,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -0.11070326715707779,
      "rewards/margins": 1.2766830921173096,
      "rewards/rejected": -1.3873863220214844,
      "step": 60
    },
    {
      "epoch": 0.52,
      "grad_norm": 4.10574880535341,
      "learning_rate": 2.7914261414993976e-07,
      "logits/chosen": -0.5475656390190125,
      "logits/rejected": -0.48507851362228394,
      "logps/chosen": -352.1190490722656,
      "logps/rejected": -886.7733154296875,
      "loss": 0.3745,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.16023167967796326,
      "rewards/margins": 1.871180772781372,
      "rewards/rejected": -2.031412124633789,
      "step": 70
    },
    {
      "epoch": 0.59,
      "grad_norm": 4.473650758441227,
      "learning_rate": 2.1442129043167873e-07,
      "logits/chosen": -0.563640832901001,
      "logits/rejected": -0.4750913083553314,
      "logps/chosen": -358.35308837890625,
      "logps/rejected": -909.8238525390625,
      "loss": 0.3521,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -0.2543565332889557,
      "rewards/margins": 2.2235209941864014,
      "rewards/rejected": -2.477877378463745,
      "step": 80
    },
    {
      "epoch": 0.67,
      "grad_norm": 4.244427255212392,
      "learning_rate": 1.5208491051320744e-07,
      "logits/chosen": -0.5700355768203735,
      "logits/rejected": -0.6186022758483887,
      "logps/chosen": -375.76641845703125,
      "logps/rejected": -1033.01171875,
      "loss": 0.3265,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -0.30810004472732544,
      "rewards/margins": 3.5270843505859375,
      "rewards/rejected": -3.835183620452881,
      "step": 90
    },
    {
      "epoch": 0.74,
      "grad_norm": 5.017749470604324,
      "learning_rate": 9.631206222392479e-08,
      "logits/chosen": -0.5446823835372925,
      "logits/rejected": -0.6302965879440308,
      "logps/chosen": -325.74017333984375,
      "logps/rejected": -1016.5056762695312,
      "loss": 0.3165,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -0.2934742569923401,
      "rewards/margins": 3.200576066970825,
      "rewards/rejected": -3.4940502643585205,
      "step": 100
    },
    {
      "epoch": 0.74,
      "eval_logits/chosen": -0.8301539421081543,
      "eval_logits/rejected": -0.628753662109375,
      "eval_logps/chosen": -359.0135192871094,
      "eval_logps/rejected": -668.845947265625,
      "eval_loss": 0.5069088339805603,
      "eval_rewards/accuracies": 0.875,
      "eval_rewards/chosen": -0.15295247733592987,
      "eval_rewards/margins": 0.317004531621933,
      "eval_rewards/rejected": -0.46995699405670166,
      "eval_runtime": 13.8942,
      "eval_samples_per_second": 6.909,
      "eval_steps_per_second": 0.216,
      "step": 100
    },
    {
      "epoch": 0.81,
      "grad_norm": 4.353344275952825,
      "learning_rate": 5.0841360885690996e-08,
      "logits/chosen": -0.4664304256439209,
      "logits/rejected": -0.6527875661849976,
      "logps/chosen": -359.19305419921875,
      "logps/rejected": -1181.1474609375,
      "loss": 0.2813,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.31564438343048096,
      "rewards/margins": 4.183438301086426,
      "rewards/rejected": -4.499083042144775,
      "step": 110
    },
    {
      "epoch": 0.89,
      "grad_norm": 5.623756798134134,
      "learning_rate": 1.872083911532907e-08,
      "logits/chosen": -0.5250264406204224,
      "logits/rejected": -0.6364299058914185,
      "logps/chosen": -424.48333740234375,
      "logps/rejected": -1234.7606201171875,
      "loss": 0.2746,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -0.3566262423992157,
      "rewards/margins": 4.7986578941345215,
      "rewards/rejected": -5.1552839279174805,
      "step": 120
    },
    {
      "epoch": 0.96,
      "grad_norm": 6.011933917738528,
      "learning_rate": 2.1036283830834224e-09,
      "logits/chosen": -0.49347949028015137,
      "logits/rejected": -0.621853768825531,
      "logps/chosen": -399.3771667480469,
      "logps/rejected": -1258.150390625,
      "loss": 0.3003,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.3925880193710327,
      "rewards/margins": 5.569568157196045,
      "rewards/rejected": -5.962155818939209,
      "step": 130
    },
    {
      "epoch": 1.0,
      "step": 135,
      "total_flos": 0.0,
      "train_loss": 0.4338115497871681,
      "train_runtime": 2557.4167,
      "train_samples_per_second": 3.376,
      "train_steps_per_second": 0.053
    }
  ],
  "logging_steps": 10,
  "max_steps": 135,
  "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
}