eval_results.json

{
    "epoch": 0.9987943737441393,
    "eval_logits/chosen": -2.7708370685577393,
    "eval_logits/rejected": -2.7552716732025146,
    "eval_logps/chosen": -1.012373447418213,
    "eval_logps/rejected": -1.167529821395874,
    "eval_loss": 1.8871065378189087,
    "eval_rewards/accuracies": 0.6422872543334961,
    "eval_rewards/chosen": -1.012373447418213,
    "eval_rewards/margins": 0.1551561951637268,
    "eval_rewards/rejected": -1.167529821395874,
    "eval_runtime": 436.9003,
    "eval_samples": 2994,
    "eval_samples_per_second": 6.853,
    "eval_steps_per_second": 0.43
}