{
  "apiVersion": "v1",
  "methodology": "https://sourcescore.org/methodology/",
  "canonical": "https://sourcescore.org/claims/824f830889daf33e/",
  "claim": {
    "vertical": "ai-ml",
    "subject": "TruthfulQA benchmark",
    "predicate": "introduced_in_paper",
    "object": "TruthfulQA: Measuring How Models Mimic Human Falsehoods (Lin et al., 2021)",
    "confidence": 0.92,
    "sources": [
      {
        "url": "https://arxiv.org/abs/2109.07958",
        "title": "TruthfulQA: Measuring How Models Mimic Human Falsehoods",
        "publisher": "arXiv (Lin, Hilton, Evans — University of Oxford + OpenAI)",
        "publishedDate": "2021-09-08",
        "accessedDate": "2026-05-31",
        "type": "preprint",
        "excerpt": "The benchmark comprises 817 questions that span 38 categories, including health, law, finance and politics."
      },
      {
        "url": "https://github.com/sylinrl/TruthfulQA",
        "title": "TruthfulQA benchmark repository",
        "publisher": "Stephanie Lin (sylinrl)",
        "publishedDate": "2021-09-08",
        "accessedDate": "2026-05-31",
        "type": "github-release"
      },
      {
        "url": "https://huggingface.co/papers/2109.07958",
        "title": "TruthfulQA (Hugging Face Papers)",
        "publisher": "Hugging Face",
        "accessedDate": "2026-05-31",
        "type": "docs"
      }
    ],
    "publishedAt": "2026-05-31T00:00:00Z",
    "lastVerified": "2026-05-31",
    "methodologyVersion": "veritas-v0.1",
    "tags": [
      "truthfulqa",
      "benchmark",
      "evaluation",
      "truthfulness",
      "hallucination",
      "lin",
      "2021"
    ],
    "id": "824f830889daf33e",
    "statement": "TruthfulQA benchmark introduced in paper: TruthfulQA: Measuring How Models Mimic Human Falsehoods (Lin et al., 2021)."
  },
  "signature": {
    "algorithm": "HMAC-SHA256",
    "signedBy": "did:web:sourcescore.org",
    "signedAt": "2026-05-31T00:00:00.000Z",
    "signature": "e2c3930772db1558dbd339b15dcd0ca813f95304244c7b3ffe199d610dfc4f27"
  },
  "citedAs": "TruthfulQA benchmark introduced in paper: TruthfulQA: Measuring How Models Mimic Human Falsehoods (Lin et al., 2021). — SourceScore Claim 824f830889daf33e (verified 2026-05-31, signed e2c39307…). https://sourcescore.org/claims/824f830889daf33e/"
}