{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.971563981042654,
  "eval_steps": 128,
  "global_step": 104,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.018957345971563982,
      "grad_norm": 118.64707840566432,
      "learning_rate": 4.545454545454545e-08,
      "logits/chosen": 117.53560638427734,
      "logits/rejected": 126.8960952758789,
      "logps/chosen": -335.40118408203125,
      "logps/rejected": -439.16552734375,
      "loss": 0.6238,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 89.83608380616089,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": 135.038818359375,
      "logits/rejected": 138.375244140625,
      "logps/chosen": -396.12841796875,
      "logps/rejected": -439.4309387207031,
      "loss": 0.4619,
      "rewards/accuracies": 0.4236111044883728,
      "rewards/chosen": -0.00659214286133647,
      "rewards/margins": -0.0014215832343325019,
      "rewards/rejected": -0.005170559044927359,
      "step": 10
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 78.22469381016133,
      "learning_rate": 4.885348141000122e-07,
      "logits/chosen": 121.09458923339844,
      "logits/rejected": 124.804931640625,
      "logps/chosen": -370.9910583496094,
      "logps/rejected": -425.46539306640625,
      "loss": 0.3963,
      "rewards/accuracies": 0.6187499761581421,
      "rewards/chosen": 0.07104669511318207,
      "rewards/margins": 0.3447481393814087,
      "rewards/rejected": -0.27370142936706543,
      "step": 20
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 60.448251752592725,
      "learning_rate": 4.5025027361734613e-07,
      "logits/chosen": 140.5128631591797,
      "logits/rejected": 134.0343780517578,
      "logps/chosen": -430.7742614746094,
      "logps/rejected": -481.88446044921875,
      "loss": 0.3509,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -1.915453314781189,
      "rewards/margins": 1.1786596775054932,
      "rewards/rejected": -3.0941128730773926,
      "step": 30
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 60.84662620202295,
      "learning_rate": 3.893311157806091e-07,
      "logits/chosen": 125.0122299194336,
      "logits/rejected": 113.4749526977539,
      "logps/chosen": -417.31817626953125,
      "logps/rejected": -448.8020935058594,
      "loss": 0.3266,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": -3.1561999320983887,
      "rewards/margins": 1.3901399374008179,
      "rewards/rejected": -4.546339988708496,
      "step": 40
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 72.73427920034227,
      "learning_rate": 3.126631330646801e-07,
      "logits/chosen": 138.95956420898438,
      "logits/rejected": 143.02549743652344,
      "logps/chosen": -492.82830810546875,
      "logps/rejected": -581.9652709960938,
      "loss": 0.2986,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -3.6842238903045654,
      "rewards/margins": 1.597720980644226,
      "rewards/rejected": -5.28194522857666,
      "step": 50
    },
    {
      "epoch": 1.1374407582938388,
      "grad_norm": 44.0378737886917,
      "learning_rate": 2.2891223348923882e-07,
      "logits/chosen": 129.6141357421875,
      "logits/rejected": 133.13836669921875,
      "logps/chosen": -480.71624755859375,
      "logps/rejected": -576.3870849609375,
      "loss": 0.2058,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -4.1355299949646,
      "rewards/margins": 2.8655474185943604,
      "rewards/rejected": -7.001077175140381,
      "step": 60
    },
    {
      "epoch": 1.3270142180094786,
      "grad_norm": 40.42911482167594,
      "learning_rate": 1.4754491880085317e-07,
      "logits/chosen": 122.14341735839844,
      "logits/rejected": 124.2646484375,
      "logps/chosen": -461.14410400390625,
      "logps/rejected": -571.5001220703125,
      "loss": 0.1516,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": -4.70001220703125,
      "rewards/margins": 3.1992990970611572,
      "rewards/rejected": -7.8993120193481445,
      "step": 70
    },
    {
      "epoch": 1.5165876777251186,
      "grad_norm": 36.720234018841005,
      "learning_rate": 7.775827023107834e-08,
      "logits/chosen": 107.95296478271484,
      "logits/rejected": 124.5838623046875,
      "logps/chosen": -456.9344787597656,
      "logps/rejected": -586.04150390625,
      "loss": 0.1324,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": -5.164170742034912,
      "rewards/margins": 3.476015090942383,
      "rewards/rejected": -8.640186309814453,
      "step": 80
    },
    {
      "epoch": 1.7061611374407581,
      "grad_norm": 35.79600791893404,
      "learning_rate": 2.7440387297912122e-08,
      "logits/chosen": 107.18046569824219,
      "logits/rejected": 119.84127044677734,
      "logps/chosen": -484.3689880371094,
      "logps/rejected": -611.01416015625,
      "loss": 0.1218,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -4.896896839141846,
      "rewards/margins": 3.7697243690490723,
      "rewards/rejected": -8.666621208190918,
      "step": 90
    },
    {
      "epoch": 1.8957345971563981,
      "grad_norm": 38.68822400635857,
      "learning_rate": 2.27878296044029e-09,
      "logits/chosen": 113.38189697265625,
      "logits/rejected": 113.06996154785156,
      "logps/chosen": -471.37005615234375,
      "logps/rejected": -573.8251342773438,
      "loss": 0.1213,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -4.4697489738464355,
      "rewards/margins": 3.383305788040161,
      "rewards/rejected": -7.853055000305176,
      "step": 100
    },
    {
      "epoch": 1.971563981042654,
      "step": 104,
      "total_flos": 0.0,
      "train_loss": 0.2568448650149199,
      "train_runtime": 2175.3939,
      "train_samples_per_second": 6.206,
      "train_steps_per_second": 0.048
    }
  ],
  "logging_steps": 10,
  "max_steps": 104,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}