{
  "status": "ok",
  "backend": "trl_transformers",
  "model_id": "hf-internal-testing/tiny-random-gpt2",
  "records": 4,
  "prompts_path": "/Users/daver/Desktop/Meta_Pytorch_OpenEnv_Scaler/polyguard-rl/data/processed/training_corpus_grpo_prompts.jsonl",
  "reward_summary": {
    "count": 2,
    "avg_reward": 0.764,
    "avg_reward_components": {
      "format_compliance_score": 0.999,
      "candidate_alignment_score": 0.999,
      "legality_score": 0.999,
      "safety_delta_score": 0.5,
      "burden_improvement_score": 0.5,
      "disease_stability_score": 0.9,
      "dosing_quality_score": 0.5,
      "abstention_quality_score": 0.56,
      "efficiency_score": 0.857,
      "process_fidelity_score": 0.92,
      "explanation_grounding_score": 0.8,
      "anti_cheat_score": 0.001,
      "uncertainty_calibration_score": 0.7
    },
    "avg_primary_reward_channels": {
      "safety_legality": 0.675,
      "clinical_improvement": 0.633,
      "dosing_quality": 0.53,
      "process_integrity": 0.894
    }
  },
  "reward_log": "/Users/daver/Desktop/Meta_Pytorch_OpenEnv_Scaler/polyguard-rl/checkpoints/grpo_reward_components.jsonl",
  "train_metrics": {
    "train_runtime": 1.0826,
    "train_samples_per_second": 1.847,
    "train_steps_per_second": 0.924,
    "total_flos": 0.0,
    "train_loss": 9.910791413858533e-05
  },
  "artifact_path": "/Users/daver/Desktop/Meta_Pytorch_OpenEnv_Scaler/polyguard-rl/checkpoints/grpo_adapter",
  "unsloth_available": false
}