{
    "epoch": 0.9999197753710389,
    "eval_logits/chosen": -2.79840087890625,
    "eval_logits/rejected": -2.752188205718994,
    "eval_logps/chosen": -378.4407043457031,
    "eval_logps/rejected": -260.23724365234375,
    "eval_loss": 0.45464351773262024,
    "eval_rewards/accuracies": 0.7863636612892151,
    "eval_rewards/chosen": -1.0035351514816284,
    "eval_rewards/margins": 1.364441156387329,
    "eval_rewards/rejected": -2.367976188659668,
    "eval_runtime": 2783.2579,
    "eval_samples": 10555,
    "eval_samples_per_second": 3.792,
    "eval_steps_per_second": 0.474,
    "total_flos": 0.0,
    "train_loss": 0.44136011642577255,
    "train_runtime": 57646.808,
    "train_samples": 99717,
    "train_samples_per_second": 1.73,
    "train_steps_per_second": 0.108
}