{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:SKFNQR3RNUIW7PFAT4RMM6466N","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"10e24de9756e678399295047bc3cc99f9ebc1e0eb5063812b696098225d70c22","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-04-07T18:42:01Z","title_canon_sha256":"68102a602a15e26d0ae82cac5aeb7474374edbe3a33527be8383e1367c1f4d25"},"schema_version":"1.0","source":{"id":"2504.05419","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.05419","created_at":"2026-07-05T10:45:42Z"},{"alias_kind":"arxiv_version","alias_value":"2504.05419v1","created_at":"2026-07-05T10:45:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.05419","created_at":"2026-07-05T10:45:42Z"},{"alias_kind":"pith_short_12","alias_value":"SKFNQR3RNUIW","created_at":"2026-07-05T10:45:42Z"},{"alias_kind":"pith_short_16","alias_value":"SKFNQR3RNUIW7PFA","created_at":"2026-07-05T10:45:42Z"},{"alias_kind":"pith_short_8","alias_value":"SKFNQR3R","created_at":"2026-07-05T10:45:42Z"}],"graph_snapshots":[{"event_id":"sha256:1f259f25a357a59a3acd1d9d10ecaf59c6c0bc9cf2ed38a0073830a573d7f371","target":"graph","created_at":"2026-07-05T10:45:42Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2504.05419/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reasoning models have achieved remarkable performance on tasks like math and logical reasoning thanks to their ability to search during reasoning. However, they still suffer from overthinking, often performing unnecessary reasoning steps even after reaching the correct answer. This raises the question: can models evaluate the correctness of their intermediate answers during reasoning? In this work, we study whether reasoning models encode information about answer correctness through probing the model's hidden states. The resulting probe can verify intermediate answers with high accuracy and pr","authors_text":"Anqi Zhang, Aurojit Panda, Chen Zhao, He He, Jane Pan, Jinyang Li, Yulin Chen","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-04-07T18:42:01Z","title":"Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.05419","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:957d4c06d4d05a037e2fdec603f5e4553fcf0d49ebc30135f3f2c444de8cbbea","target":"record","created_at":"2026-07-05T10:45:42Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"10e24de9756e678399295047bc3cc99f9ebc1e0eb5063812b696098225d70c22","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-04-07T18:42:01Z","title_canon_sha256":"68102a602a15e26d0ae82cac5aeb7474374edbe3a33527be8383e1367c1f4d25"},"schema_version":"1.0","source":{"id":"2504.05419","kind":"arxiv","version":1}},"canonical_sha256":"928ad847716d116fbca09f22c67b9ef3686169d010c1f27e14a10c0ece28b45d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"928ad847716d116fbca09f22c67b9ef3686169d010c1f27e14a10c0ece28b45d","first_computed_at":"2026-07-05T10:45:42.122593Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:45:42.122593Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8u8mMUg1Hf2oL2fEWP1zYu1f9C/NXkdJb3tZaWlc1ggQzLmoJO9pkjZSs/gd8o1cn8KyK+kbLTtNpW/LPXspCw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:45:42.123003Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.05419","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:957d4c06d4d05a037e2fdec603f5e4553fcf0d49ebc30135f3f2c444de8cbbea","sha256:1f259f25a357a59a3acd1d9d10ecaf59c6c0bc9cf2ed38a0073830a573d7f371"],"state_sha256":"399ae5d16b4d183631890cd5ee6c8d24c1517afc7c82b86dfc381f64a35f58f8"}