{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3PQRCAHWBS2F2Z7JAKSVEZSBR7","short_pith_number":"pith:3PQRCAHW","schema_version":"1.0","canonical_sha256":"dbe11100f60cb45d67e902a55266418fce19788fe0bfff86f21d6099eb33706f","source":{"kind":"arxiv","id":"2607.23386","version":1},"attestation_state":"computed","paper":{"title":"Confidently Wrong: Exception Chain Collapse in Frontier LLM Rule Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"John Kozak, Lisa Doake, Paul Simpson","submitted_at":"2026-07-25T22:40:02Z","abstract_excerpt":"We document a failure class in frontier large language models -- exception chain collapse -- observed in eligibility evaluation under nested conditional rules of the form \"A is required UNLESS B applies, UNLESS C overrides B\". The failure reproduces at first observation, but its empirical surface is unstable: between March and April 2026 several failure cells closed silently under the same model alias, with no version bump (GPT-5.4 on construction insurance moved from 96.6% to 100%, same prompt and harness). For regulated workflows, frontier-model accuracy is a moving compliance boundary that "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2607.23386","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-25T22:40:02Z","cross_cats_sorted":[],"title_canon_sha256":"449460d96a7751fa3c1b3c5b077418ee25ccb6a2cd5b52d8e51035dd7ac529aa","abstract_canon_sha256":"499165a699cba55a4945c624a5582396756d4085f4f722d05187996f99a31916"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-28T01:22:50.154828Z","signature_b64":"pwUnw/IiKpxE/bN8RBLBrYX9EVaD42iCjbSQOAwR74t0OUk5qCqTArWWvkJLHLNdjaP9mdx3vSQaQLFKRnV6Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbe11100f60cb45d67e902a55266418fce19788fe0bfff86f21d6099eb33706f","last_reissued_at":"2026-07-28T01:22:50.154039Z","signature_status":"signed_v1","first_computed_at":"2026-07-28T01:22:50.154039Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Confidently Wrong: Exception Chain Collapse in Frontier LLM Rule Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"John Kozak, Lisa Doake, Paul Simpson","submitted_at":"2026-07-25T22:40:02Z","abstract_excerpt":"We document a failure class in frontier large language models -- exception chain collapse -- observed in eligibility evaluation under nested conditional rules of the form \"A is required UNLESS B applies, UNLESS C overrides B\". The failure reproduces at first observation, but its empirical surface is unstable: between March and April 2026 several failure cells closed silently under the same model alias, with no version bump (GPT-5.4 on construction insurance moved from 96.6% to 100%, same prompt and harness). For regulated workflows, frontier-model accuracy is a moving compliance boundary that "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.23386","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.23386/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.23386","created_at":"2026-07-28T01:22:50.154454+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.23386v1","created_at":"2026-07-28T01:22:50.154454+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.23386","created_at":"2026-07-28T01:22:50.154454+00:00"},{"alias_kind":"pith_short_12","alias_value":"3PQRCAHWBS2F","created_at":"2026-07-28T01:22:50.154454+00:00"},{"alias_kind":"pith_short_16","alias_value":"3PQRCAHWBS2F2Z7J","created_at":"2026-07-28T01:22:50.154454+00:00"},{"alias_kind":"pith_short_8","alias_value":"3PQRCAHW","created_at":"2026-07-28T01:22:50.154454+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7","json":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7.json","graph_json":"https://pith.science/api/pith-number/3PQRCAHWBS2F2Z7JAKSVEZSBR7/graph.json","events_json":"https://pith.science/api/pith-number/3PQRCAHWBS2F2Z7JAKSVEZSBR7/events.json","paper":"https://pith.science/paper/3PQRCAHW"},"agent_actions":{"view_html":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7","download_json":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7.json","view_paper":"https://pith.science/paper/3PQRCAHW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.23386&json=true","fetch_graph":"https://pith.science/api/pith-number/3PQRCAHWBS2F2Z7JAKSVEZSBR7/graph.json","fetch_events":"https://pith.science/api/pith-number/3PQRCAHWBS2F2Z7JAKSVEZSBR7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7/action/storage_attestation","attest_author":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7/action/author_attestation","sign_citation":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7/action/citation_signature","submit_replication":"https://pith.science/pith/3PQRCAHWBS2F2Z7JAKSVEZSBR7/action/replication_record"}},"created_at":"2026-07-28T01:22:50.154454+00:00","updated_at":"2026-07-28T01:22:50.154454+00:00"}