{
    "epoch": 0.9982631930527722,
    "eval_logits/chosen": -1.386914849281311,
    "eval_logits/rejected": -1.3965359926223755,
    "eval_logps/chosen": -1.7612468004226685,
    "eval_logps/rejected": -2.3322505950927734,
    "eval_loss": 1.228195309638977,
    "eval_rewards/accuracies": 0.8414633870124817,
    "eval_rewards/chosen": -17.612468719482422,
    "eval_rewards/margins": 5.710038185119629,
    "eval_rewards/rejected": -23.322505950927734,
    "eval_runtime": 87.9064,
    "eval_samples": 1961,
    "eval_samples_per_second": 22.308,
    "eval_steps_per_second": 1.399
}