{
    "epoch": 0.9980582524271845,
    "eval_logits/chosen": -1.39054536819458,
    "eval_logits/rejected": -1.4322466850280762,
    "eval_logps/chosen": -245.04893493652344,
    "eval_logps/rejected": -228.5849151611328,
    "eval_loss": 0.6930710077285767,
    "eval_rewards/accuracies": 0.43386074900627136,
    "eval_rewards/chosen": 2.820852751028724e-05,
    "eval_rewards/margins": 0.00016564132238272578,
    "eval_rewards/rejected": -0.00013743281306233257,
    "eval_runtime": 970.4456,
    "eval_samples": 3156,
    "eval_samples_per_second": 3.252,
    "eval_steps_per_second": 0.407,
    "total_flos": 0.0,
    "train_loss": 0.6931419428220519,
    "train_runtime": 4635.9681,
    "train_samples": 4119,
    "train_samples_per_second": 0.888,
    "train_steps_per_second": 0.055
}