{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DUE2XCBYQXFXERPS4L7OH2YFHS","short_pith_number":"pith:DUE2XCBY","schema_version":"1.0","canonical_sha256":"1d09ab883885cb7245f2e2fee3eb053c84f8eb9e1b3eae27404e0614a5647648","source":{"kind":"arxiv","id":"2607.26102","version":1},"attestation_state":"computed","paper":{"title":"A Reference-Free Score for Detecting Silent Reasoning Failures in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Atul, Divya Mishra, Mehul Kumar Das, Varun Shukla, Vivek Shukla","submitted_at":"2026-07-28T07:28:38Z","abstract_excerpt":"Mathematical chain of thought (CoT) evaluation is commonly reduced to whether the final answer matches a reference. This conflates producing a correct conclusion with producing a valid derivation an invalid chain can accidentally reach the right answer, while a valid calculation can be followed by a transcription error. We call this mismatch the reasoning answer consistency gap. This framework paper introduces the Reasoning Answer Faithfulness Score (RAFS), a reference free, instance level diagnostic of whether an emitted mathematical trace is locally credible, supports its answer, and is stab"},"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":"2607.26102","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-07-28T07:28:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e7f745affe74d18293b97531e7a3f7c75ad1fbd6ac9ed6f8b8be2adea67cd50d","abstract_canon_sha256":"2a229448a39aa2a87e5886160715e3d966e78eb6b979da399f551732ce35604a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d09ab883885cb7245f2e2fee3eb053c84f8eb9e1b3eae27404e0614a5647648","last_reissued_at":"2026-07-30T00:08:21.313378Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-30T00:08:21.313378Z"},"graph_snapshot":{"paper":{"title":"A Reference-Free Score for Detecting Silent Reasoning Failures in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Atul, Divya Mishra, Mehul Kumar Das, Varun Shukla, Vivek Shukla","submitted_at":"2026-07-28T07:28:38Z","abstract_excerpt":"Mathematical chain of thought (CoT) evaluation is commonly reduced to whether the final answer matches a reference. This conflates producing a correct conclusion with producing a valid derivation an invalid chain can accidentally reach the right answer, while a valid calculation can be followed by a transcription error. We call this mismatch the reasoning answer consistency gap. This framework paper introduces the Reasoning Answer Faithfulness Score (RAFS), a reference free, instance level diagnostic of whether an emitted mathematical trace is locally credible, supports its answer, and is stab"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.26102","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/2607.26102/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":"2607.26102","created_at":"2026-07-30T00:08:21.317015+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.26102v1","created_at":"2026-07-30T00:08:21.317015+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.26102","created_at":"2026-07-30T00:08:21.317015+00:00"},{"alias_kind":"pith_short_12","alias_value":"DUE2XCBYQXFX","created_at":"2026-07-30T00:08:21.317015+00:00"},{"alias_kind":"pith_short_16","alias_value":"DUE2XCBYQXFXERPS","created_at":"2026-07-30T00:08:21.317015+00:00"},{"alias_kind":"pith_short_8","alias_value":"DUE2XCBY","created_at":"2026-07-30T00:08:21.317015+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/DUE2XCBYQXFXERPS4L7OH2YFHS","json":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS.json","graph_json":"https://pith.science/api/pith-number/DUE2XCBYQXFXERPS4L7OH2YFHS/graph.json","events_json":"https://pith.science/api/pith-number/DUE2XCBYQXFXERPS4L7OH2YFHS/events.json","paper":"https://pith.science/paper/DUE2XCBY"},"agent_actions":{"view_html":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS","download_json":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS.json","view_paper":"https://pith.science/paper/DUE2XCBY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.26102&json=true","fetch_graph":"https://pith.science/api/pith-number/DUE2XCBYQXFXERPS4L7OH2YFHS/graph.json","fetch_events":"https://pith.science/api/pith-number/DUE2XCBYQXFXERPS4L7OH2YFHS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS/action/storage_attestation","attest_author":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS/action/author_attestation","sign_citation":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS/action/citation_signature","submit_replication":"https://pith.science/pith/DUE2XCBYQXFXERPS4L7OH2YFHS/action/replication_record"}},"created_at":"2026-07-30T00:08:21.317015+00:00","updated_at":"2026-07-30T00:08:21.317015+00:00"}