eval_results.json

{
    "epoch": 0.9984168865435357,
    "eval_logits/chosen": -3.1016640663146973,
    "eval_logits/rejected": -2.988262891769409,
    "eval_logps/chosen": -530.1766967773438,
    "eval_logps/rejected": -572.2643432617188,
    "eval_loss": 0.621910035610199,
    "eval_rewards/accuracies": 0.6451612710952759,
    "eval_rewards/chosen": -2.672846555709839,
    "eval_rewards/margins": 0.31082332134246826,
    "eval_rewards/rejected": -2.9836699962615967,
    "eval_runtime": 324.3699,
    "eval_samples": 1976,
    "eval_samples_per_second": 6.092,
    "eval_steps_per_second": 0.382
}