{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-11-03-013",
    "source_story_id": "tmp-story-monofacts-scientific-discovery",
    "edition_id": "mp-2026-11-03-morning-0116",
    "edition_url": "https://themachinepress.com/edition/2026-11-03",
    "position": 13,
    "story_type": "dispatch",
    "section": "research",
    "editorial_classification": "editorial",
    "headline": "The Rare Fact Carries the Discovery Signal",
    "slug": "the-rare-fact-carries-the-discovery-signal",
    "dek": "Controlled training experiments link a higher rate of facts seen once—monofacts—to more model discoveries in materials science and regulatory genomics.",
    "summary": "Controlled training experiments link a higher rate of facts seen once—monofacts—to more model discoveries in materials science and regulatory genomics.",
    "body_text": "The study asks which property of training data enables a language model to propose discoveries rather than merely repeat common patterns. Its theory treats scientific-discovery capacity as missing probability mass and singles out monofacts, facts observed once during training. The authors directly vary monofact rates in materials-science and regulatory-genomics corpora. Across two open-weight model families, sampling strategies and reasoning budgets, more monofacts consistently produce more discoveries. Additional tests argue that ordinary dataset diversity does not explain the relationship and that the effect can transfer across domains when paired with in-domain data. The experiments establish a controlled association in the studied settings, not a guarantee that adding rare facts will make any model a reliable scientist.",
    "why_it_matters": "Controlled training experiments link a higher rate of facts seen once—monofacts—to more model discoveries in materials science and regulatory genomics.",
    "limitations": [
      "Additional tests argue that ordinary dataset diversity does not explain the relationship and that the effect can transfer across domains when paired with in-domain data.",
      "The experiments establish a controlled association in the studied settings, not a guarantee that adding rare facts will make any model a reliable scientist."
    ],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-03-013/the-rare-fact-carries-the-discovery-signal",
    "json_url": "https://themachinepress.com/story/mp-2026-11-03-013.json",
    "first_published_at": "2026-11-03T09:00:00.000-05:00",
    "modified_at": "2026-11-03T09:00:00.000-05:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-11-03-013-001",
        "text": "Controlled training experiments link a higher rate of facts seen once—monofacts—to more model discoveries in materials science and regulatory genomics.",
        "source_ids": [
          "source-2026-11-03-013"
        ],
        "qualification": "Additional tests argue that ordinary dataset diversity does not explain the relationship and that the effect can transfer across domains when paired with in-domain data."
      }
    ],
    "source_ids": [
      "source-2026-11-03-013"
    ],
    "tags": [
      "scientific discovery",
      "training data",
      "language models",
      "materials science"
    ],
    "image_url": "https://themachinepress.com/issues/2026-11-03/discovery-data.webp",
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-11-03-013",
      "title": "arXiv preprint 2611.03617, original v1",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2611.03617v1",
      "canonical_url": "https://arxiv.org/abs/2611.03617v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-11-02T13:50:54.000-05:00",
      "accessed_at": "2026-11-03T08:18:17.000-05:00",
      "supports_claim_ids": [
        "claim-mp-2026-11-03-013-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "The Rare Fact Carries the Discovery Signal",
    "publisher": "The Machine Press",
    "published_at": "2026-11-03T09:00:00.000-05:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-03-013/the-rare-fact-carries-the-discovery-signal"
  }
}