{
  "task": "ppo",
  "strategy": "full",
  "base_model": "gpt2-medium",
  "reward_model": "OmAhire369/safe-genai-reward-full",
  "n_train_prompts": 4000,
  "ppo_steps": 120,
  "batch_size": 32,
  "learning_rate": 1.41e-05,
  "param_stats": {
    "total_params": 354823168,
    "trainable_params": 354823168,
    "trainable_pct": 100.0,
    "total_M": 354.82,
    "trainable_M": 354.823
  },
  "build_notes": [],
  "reward_start": -4.85760498046875,
  "reward_final": -1.817138671875,
  "reward_best": 0.2688255310058594,
  "best_probe_tag": "step-80",
  "used_checkpoint": "best",
  "reward_delta": 3.04046630859375,
  "stopped_early": false,
  "length_collapsed": true,
  "length_start": 46.96875,
  "length_final": 48.0,
  "train_seconds": 1751.58,
  "peak_gpu_mb": 9386.4,
  "data_source": "hf:nrizwan/safe_ai_assignment_1"
}