{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QXBET7U46BCPJVMDE445IAVB7Z","short_pith_number":"pith:QXBET7U4","schema_version":"1.0","canonical_sha256":"85c249fe9cf044f4d5832739d402a1fe6116e4e3a718328d3cb5bd32fd107bc6","source":{"kind":"arxiv","id":"2605.29288","version":1},"attestation_state":"computed","paper":{"title":"Diagnosing Harmful Continuation in Answer-Correct Long-CoT Training Traces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chen He, Fumin Shen, Lei Wang, Wenxuan Zhang, Yuhao Wu","submitted_at":"2026-05-28T03:12:15Z","abstract_excerpt":"Long chain-of-thought (CoT) traces are widely used as supervision for reasoning-oriented LLM SFT, yet answer-correct traces can still lead to markedly different fine-tuning outcomes. We study post-conclusion continuation in answer-correct long-CoT data: a continuation where the answer appears sufficiently supported, but the trace continues with additional reasoning that remains in the supervised target. To test its training effect, we use a delete-only editor to construct answer-preserving suffix removal and compare CoT-based SFT on the original and processed traces. We observe improved SFT ou"},"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":"2605.29288","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-28T03:12:15Z","cross_cats_sorted":[],"title_canon_sha256":"2d77f36a72dead3d6631e90667febf24beda157ee7889f4d82ae79a2ef10da9e","abstract_canon_sha256":"09a4a8398ee7a3bcca69a2c87036ab437b004ba25adf6c50e61d00f1f2dfef3b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:29.643628Z","signature_b64":"NEIt3ESkxY9vrlmtmTAhQ0vmrevJOXbX3Rqwo7KQPjPpqi8d7MB7Lb6169SvKC+ak1SZgYsag8tyv/YmUlnnBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"85c249fe9cf044f4d5832739d402a1fe6116e4e3a718328d3cb5bd32fd107bc6","last_reissued_at":"2026-05-29T01:05:29.642626Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:29.642626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Diagnosing Harmful Continuation in Answer-Correct Long-CoT Training Traces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chen He, Fumin Shen, Lei Wang, Wenxuan Zhang, Yuhao Wu","submitted_at":"2026-05-28T03:12:15Z","abstract_excerpt":"Long chain-of-thought (CoT) traces are widely used as supervision for reasoning-oriented LLM SFT, yet answer-correct traces can still lead to markedly different fine-tuning outcomes. We study post-conclusion continuation in answer-correct long-CoT data: a continuation where the answer appears sufficiently supported, but the trace continues with additional reasoning that remains in the supervised target. To test its training effect, we use a delete-only editor to construct answer-preserving suffix removal and compare CoT-based SFT on the original and processed traces. We observe improved SFT ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29288","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/2605.29288/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":"2605.29288","created_at":"2026-05-29T01:05:29.642960+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.29288v1","created_at":"2026-05-29T01:05:29.642960+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29288","created_at":"2026-05-29T01:05:29.642960+00:00"},{"alias_kind":"pith_short_12","alias_value":"QXBET7U46BCP","created_at":"2026-05-29T01:05:29.642960+00:00"},{"alias_kind":"pith_short_16","alias_value":"QXBET7U46BCPJVMD","created_at":"2026-05-29T01:05:29.642960+00:00"},{"alias_kind":"pith_short_8","alias_value":"QXBET7U4","created_at":"2026-05-29T01:05:29.642960+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/QXBET7U46BCPJVMDE445IAVB7Z","json":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z.json","graph_json":"https://pith.science/api/pith-number/QXBET7U46BCPJVMDE445IAVB7Z/graph.json","events_json":"https://pith.science/api/pith-number/QXBET7U46BCPJVMDE445IAVB7Z/events.json","paper":"https://pith.science/paper/QXBET7U4"},"agent_actions":{"view_html":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z","download_json":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z.json","view_paper":"https://pith.science/paper/QXBET7U4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.29288&json=true","fetch_graph":"https://pith.science/api/pith-number/QXBET7U46BCPJVMDE445IAVB7Z/graph.json","fetch_events":"https://pith.science/api/pith-number/QXBET7U46BCPJVMDE445IAVB7Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z/action/storage_attestation","attest_author":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z/action/author_attestation","sign_citation":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z/action/citation_signature","submit_replication":"https://pith.science/pith/QXBET7U46BCPJVMDE445IAVB7Z/action/replication_record"}},"created_at":"2026-05-29T01:05:29.642960+00:00","updated_at":"2026-05-29T01:05:29.642960+00:00"}