{
    "epoch": 1.0,
    "eval_logits/chosen": -5.382973670959473,
    "eval_logits/rejected": -5.635280132293701,
    "eval_logps/chosen": -592.3199462890625,
    "eval_logps/rejected": -744.3624877929688,
    "eval_loss": 0.7785406708717346,
    "eval_rewards/accuracies": 0.57421875,
    "eval_rewards/chosen": -2.0184545516967773,
    "eval_rewards/margins": 0.24129416048526764,
    "eval_rewards/rejected": -2.259748935699463,
    "eval_runtime": 97.5883,
    "eval_samples": 2000,
    "eval_samples_per_second": 20.494,
    "eval_steps_per_second": 0.328
}