{
    "epoch": 1.0,
    "eval_logits/chosen": 98.9942626953125,
    "eval_logits/rejected": 98.98635864257812,
    "eval_logps/chosen": -32.3518180847168,
    "eval_logps/rejected": -35.857688903808594,
    "eval_loss": 2.3100497722625732,
    "eval_rewards/accuracies": 0.5215947031974792,
    "eval_rewards/chosen": 0.045678745955228806,
    "eval_rewards/margins": -0.00873997900635004,
    "eval_rewards/rejected": 0.05441872030496597,
    "eval_runtime": 103.7492,
    "eval_samples": 343,
    "eval_samples_per_second": 3.306,
    "eval_steps_per_second": 0.414
}