all_results.json

{
    "epoch": 3.97,
    "eval_logits/chosen": -0.00967690534889698,
    "eval_logits/rejected": 0.0881744921207428,
    "eval_logps/chosen": -306.2895202636719,
    "eval_logps/rejected": -278.5288391113281,
    "eval_loss": 0.0022096126340329647,
    "eval_rewards/accuracies": 0.4790000021457672,
    "eval_rewards/chosen": 0.0010312977246940136,
    "eval_rewards/margins": -0.0005611819215118885,
    "eval_rewards/rejected": 0.0015924794133752584,
    "eval_runtime": 426.4232,
    "eval_samples": 2000,
    "eval_samples_per_second": 4.69,
    "eval_steps_per_second": 1.173,
    "train_loss": 0.0016499216942447088,
    "train_runtime": 2658.6406,
    "train_samples": 61135,
    "train_samples_per_second": 1.505,
    "train_steps_per_second": 0.093
}