eval_results.json

{
    "epoch": 0.9988571428571429,
    "eval_logits/chosen": -2.3148677349090576,
    "eval_logits/rejected": -2.157057762145996,
    "eval_logps/chosen": -285.280517578125,
    "eval_logps/rejected": -258.34063720703125,
    "eval_loss": 0.03401629254221916,
    "eval_rewards/accuracies": 0.7413793206214905,
    "eval_rewards/chosen": -0.09676530957221985,
    "eval_rewards/margins": 0.2958797514438629,
    "eval_rewards/rejected": -0.39264509081840515,
    "eval_runtime": 90.641,
    "eval_samples": 1831,
    "eval_samples_per_second": 20.201,
    "eval_steps_per_second": 0.32
}