{
    "epoch": 1.0,
    "eval_logits/chosen": -0.9196914434432983,
    "eval_logits/rejected": -0.40473511815071106,
    "eval_logps/chosen": -108.88490295410156,
    "eval_logps/rejected": -714.4129638671875,
    "eval_loss": 0.4451637268066406,
    "eval_rewards/accuracies": 0.9418103694915771,
    "eval_rewards/chosen": -0.031108301132917404,
    "eval_rewards/margins": 1.1164520978927612,
    "eval_rewards/rejected": -1.1475603580474854,
    "eval_runtime": 191.5694,
    "eval_samples": 1845,
    "eval_samples_per_second": 9.631,
    "eval_steps_per_second": 0.303,
    "train_loss": 0.41410978966289097,
    "train_runtime": 1778.1435,
    "train_samples": 6885,
    "train_samples_per_second": 3.872,
    "train_steps_per_second": 0.061
}