{
  "id": "20261108-0001",
  "version": "v1",
  "title": "The Limits of Structural Consistency: A Token-Space Audit of LLM Counting",
  "authors": [
    {
      "name": "Wuming Pan",
      "affiliation": "Sichuan University"
    }
  ],
  "date": "2026-11-08",
  "type": "Research paper",
  "pages": 10,
  "subjects": [
    "Token Space",
    "Language model evaluation",
    "Consistency auditing",
    "Negative results"
  ],
  "summary": "An executable structural audit separates inconsistent answer vectors from coherent factual errors; staged experiments establish the limits of its diagnostic benefit.",
  "abstract": "Can exact mathematical consistency checks improve the diagnosis of language-model errors? We study finite counting questions through Token Space: source objects specify variables and conditions, maps specify answers, and kernel quotients compile companion questions into integer constraints. A violated constraint certifies that an answer vector has no common realizing structure without requiring reference facts. Passing the constraints does not certify factual correctness: every coherent alternative structure passes as well. We make this classical boundary executable and reassess completed experiments under their original endpoints. In a 32-target synthetic study, a twelve-question check rejected all five incorrect main answers but also six correct ones; twelve-answer repetition was correct on all targets. A later county-adjacency study admitted one model after 72/72 exact calibration replies. Names-only counts were exact on 24/45 requests, versus 45/45 with complete facts, but neither condition produced a stable incompatible vector. One domain yielded an incompatible first vector, already detectable from two coordinates, and two further vectors matching a coherent wrong graph. The audit separates repeated agreement, structural realizability and factual correctness, while the tested procedures establish no selective-answering advantage or stable incremental detection. The final factual evidence covers one admitted model and five selected domains, including one error-enriched domain; it does not establish general ineffectiveness.",
  "release_note": "Initial public preprint corresponding to the V9 manuscript submitted to TMLR on 8 November 2026. Author attribution, the repository imprint and the data-availability statement differ from the anonymous submission; scientific claims and experimental results are unchanged. This deposit does not imply acceptance or completed peer review.",
  "provenance": "OpenAI Codex assisted with writing, revision, experiment-code development and offline evidence checks. Claude Fable 5.1 provided evaluative feedback on every intermediate manuscript version. These contributions are not independent peer review or experimental evidence; the author remains responsible for the content.",
  "verification": "The public archive includes the eight-page technical supplement and typesetting sources. Raw model responses, experimental records and executable audit modules are retained in the separate supplementary package submitted to TMLR and are not included in this public archive. The submission package passed an offline audit of all 354 final-stage responses without new model calls. Results remain subject to the manuscript's stated sample and interpretation limits.",
  "namespace": "core",
  "number": 3,
  "identifier_profile": "1.1",
  "registration_date": "2026-11-08",
  "sequence": 1,
  "license": "CC BY 4.0",
  "rights": "Paper and supplement: Creative Commons Attribution 4.0 International. Copyright retained by Wuming Pan. Bundled third-party material retains its own notices; no blanket software licence is asserted.",
  "externalIdentifiers": [
    {
      "scheme": "OpenReview",
      "value": "HNUTW95Uto",
      "url": "https://openreview.net/forum?id=HNUTW95Uto"
    }
  ],
  "key": "core/20261108-0001",
  "schemaVersion": "1.0",
  "pid": "1id1:core:20261108-0001",
  "version_pid": "1id1:core:20261108-0001@v1",
  "url": "https://1id1.net/id/core/20261108-0001/",
  "version_url": "https://1id1.net/id/core/20261108-0001/v1/",
  "language": "en",
  "status": "preprint",
  "relations": [],
  "files": [
    {
      "name": "paper.pdf",
      "url": "https://1id1.net/records/core/20261108-0001/v1/paper.pdf",
      "media_type": "application/pdf",
      "bytes": 408733,
      "sha256": "9c0d57f94d3a7ede2d21daa726a2897ff43b073a270ea4e7b02d80a37ffe275b"
    },
    {
      "name": "source.zip",
      "url": "https://1id1.net/records/core/20261108-0001/v1/source.zip",
      "media_type": "application/zip",
      "bytes": 768067,
      "sha256": "c45bfddf951a0f56ec7d61483f79192e841596ae265390e69b491003e5fe5160"
    }
  ]
}
