eval_results.json

{
    "epoch": 1.9979402677651905,
    "eval_logits/chosen": 0.47668975591659546,
    "eval_logits/rejected": 0.6139155626296997,
    "eval_logps/chosen": -348.3104248046875,
    "eval_logps/rejected": -444.2012634277344,
    "eval_loss": 0.5335665345191956,
    "eval_rewards/accuracies": 0.7265625,
    "eval_rewards/chosen": -0.9175618290901184,
    "eval_rewards/margins": 0.877814531326294,
    "eval_rewards/rejected": -1.7953763008117676,
    "eval_runtime": 24.4791,
    "eval_samples_per_second": 40.851,
    "eval_steps_per_second": 0.654
}