{
    "epoch": 1.0,
    "eval_logits/chosen": -2.1620335578918457,
    "eval_logits/rejected": -2.062356948852539,
    "eval_logps/chosen": -206.584716796875,
    "eval_logps/rejected": -214.93492126464844,
    "eval_loss": 0.521207869052887,
    "eval_rewards/accuracies": 0.7405660152435303,
    "eval_rewards/chosen": -2.5398268699645996,
    "eval_rewards/margins": 0.9912916421890259,
    "eval_rewards/rejected": -3.531118631362915,
    "eval_runtime": 424.2832,
    "eval_samples": 418,
    "eval_samples_per_second": 0.985,
    "eval_steps_per_second": 0.125
}