{
    "epoch": 1.996886351842242,
    "eval_logits/chosen": -0.16769856214523315,
    "eval_logits/rejected": 3.019533395767212,
    "eval_logps/chosen": -1311.4659423828125,
    "eval_logps/rejected": -2617.66845703125,
    "eval_loss": 0.15781185030937195,
    "eval_rewards/accuracies": 0.9215686321258545,
    "eval_rewards/chosen": -8.88057804107666,
    "eval_rewards/margins": 12.836706161499023,
    "eval_rewards/rejected": -21.71728515625,
    "eval_runtime": 297.083,
    "eval_samples": 6491,
    "eval_samples_per_second": 21.849,
    "eval_steps_per_second": 0.343,
    "total_flos": 0.0,
    "train_loss": 0.1585804910288655,
    "train_runtime": 28225.2793,
    "train_samples": 123309,
    "train_samples_per_second": 8.737,
    "train_steps_per_second": 0.034
}