{
  "_comment": "公開用 evidence sources（Loop 8, §13.1）。ソースは src/data/evidence/bibliography.json。論文は preprint と明記。",
  "schema": {
    "paperId": "string",
    "title": "string",
    "authors": "string",
    "year": "number",
    "url": "string",
    "type": "preprint|peer-reviewed",
    "note": "string"
  },
  "source": "https://cockpit.kawaii-girl.com/research/",
  "generatedAt": "2026-07-14",
  "updatePolicy": "論文追加時・年次見直し。DOI なしは推測せず arXiv ID のみ。",
  "license": "引用・リンクは可。",
  "papers": {
    "helm": { "title": "Holistic Evaluation of Language Models (HELM)", "authors": "Liang et al.", "year": 2022, "url": "https://arxiv.org/abs/2211.09110", "type": "preprint", "note": "多面的評価の提唱。現行2026モデルの性能証明には使用しない。" },
    "lost-in-the-middle": { "title": "Lost in the Middle: How Language Models Use Long Contexts", "authors": "Liu et al.", "year": 2023, "url": "https://arxiv.org/abs/2307.03172", "type": "preprint", "note": "context上限≠実効。現行モデルへの過剰一般化はしない。" },
    "longbench": { "title": "LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding", "authors": "Bai et al.", "year": 2023, "url": "https://arxiv.org/abs/2308.14508", "type": "preprint", "note": "中国語・英語タスク。日本語への直接一般化はしない。" },
    "swe-bench": { "title": "SWE-bench: Can Language Models Resolve Real-World GitHub Issues?", "authors": "Jimenez et al.", "year": 2023, "url": "https://arxiv.org/abs/2310.06770", "type": "preprint", "note": "2023論文内の古いモデルscoreを現行rankingに使用しない。" },
    "chatbot-arena": { "title": "Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference", "authors": "Chiang et al.", "year": 2024, "url": "https://arxiv.org/abs/2403.04132", "type": "preprint", "note": "preferenceとobjective correctnessは別。" },
    "llm-judge-position-bias": { "title": "Judging the Judges: A Systematic Study of Position Bias in LLM-as-a-Judge", "authors": "Shi et al.", "year": 2024, "url": "https://arxiv.org/abs/2406.07791", "type": "preprint", "note": "LLM judge の position bias。" },
    "llm-judge-prejudice": { "title": "Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge", "authors": "Ye et al.", "year": 2024, "url": "https://arxiv.org/abs/2410.02736", "type": "preprint", "note": "LLM judge の複数bias。" }
  }
}
