all_results.json

{
    "epoch": 2.986666666666667,
    "eval_kl": 28.178438186645508,
    "eval_logps/chosen": -142.316337890625,
    "eval_loss": 0.5287721157073975,
    "eval_rewards/chosen": 2.6833560180664064,
    "eval_runtime": 25.1552,
    "eval_samples_per_second": 3.975,
    "eval_steps_per_second": 3.975,
    "total_flos": 3.5905450080116736e+16,
    "train_loss": 0.5264021051781518,
    "train_runtime": 1249.11,
    "train_samples_per_second": 2.162,
    "train_steps_per_second": 0.269
}