baseline/results.json

{
  "_note": "Placeholder results. Run 'python inference.py' with API_BASE_URL, MODEL_NAME, and HF_TOKEN set to generate real LLM-based scores.",
  "model": "openai/gpt-4o-mini",
  "composite": null,
  "seed": 42,
  "easy": {
    "mean": null,
    "std": null,
    "scores": []
  },
  "medium": {
    "mean": null,
    "std": null,
    "scores": []
  },
  "hard": {
    "mean": null,
    "std": null,
    "scores": []
  },
  "elapsed_seconds": null
}