{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XFNGOAXY2MLCLSSJ5M7FKR7WLJ","short_pith_number":"pith:XFNGOAXY","schema_version":"1.0","canonical_sha256":"b95a6702f8d31625ca49eb3e5547f65a5060eece25e4a32d6bd20b7cbc40c7a3","source":{"kind":"arxiv","id":"2410.12934","version":1},"attestation_state":"computed","paper":{"title":"Enhancing Mathematical Reasoning in LLMs by Stepwise Correction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chao Shen, Meng Jiang, Qingkai Zeng, Zhaoxuan Tan, Zhenyu Wu, Zhihan Zhang","submitted_at":"2024-10-16T18:18:42Z","abstract_excerpt":"Best-of-N decoding methods instruct large language models (LLMs) to generate multiple solutions, score each using a scoring function, and select the highest scored as the final answer to mathematical reasoning problems. However, this repeated independent process often leads to the same mistakes, making the selected solution still incorrect. We propose a novel prompting method named Stepwise Correction (StepCo) that helps LLMs identify and revise incorrect steps in their generated reasoning paths. It iterates verification and revision phases that employ a process-supervised verifier. The verify"},"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":"2410.12934","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-16T18:18:42Z","cross_cats_sorted":[],"title_canon_sha256":"acf8cf8c1a7818929e1e3525d4fac89340ad0e2ee47904756f0a9e5bda2d1d8d","abstract_canon_sha256":"0cdef82978046286d94ef7063cbd5dc6ab5d6b3e254f1c64e2cc7f2aa2ae29c4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:21:52.967655Z","signature_b64":"nVnTwsa/CcEmzSgsxgw4HPCG0T7DQiT71R08/CThYJuyZtwdI3Q8FOhYeDlB88WSLtUPS0ejV8wc2CUm22L6BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b95a6702f8d31625ca49eb3e5547f65a5060eece25e4a32d6bd20b7cbc40c7a3","last_reissued_at":"2026-07-05T09:21:52.967187Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:21:52.967187Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Enhancing Mathematical Reasoning in LLMs by Stepwise Correction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chao Shen, Meng Jiang, Qingkai Zeng, Zhaoxuan Tan, Zhenyu Wu, Zhihan Zhang","submitted_at":"2024-10-16T18:18:42Z","abstract_excerpt":"Best-of-N decoding methods instruct large language models (LLMs) to generate multiple solutions, score each using a scoring function, and select the highest scored as the final answer to mathematical reasoning problems. However, this repeated independent process often leads to the same mistakes, making the selected solution still incorrect. We propose a novel prompting method named Stepwise Correction (StepCo) that helps LLMs identify and revise incorrect steps in their generated reasoning paths. It iterates verification and revision phases that employ a process-supervised verifier. The verify"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.12934","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/2410.12934/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":"2410.12934","created_at":"2026-07-05T09:21:52.967246+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.12934v1","created_at":"2026-07-05T09:21:52.967246+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.12934","created_at":"2026-07-05T09:21:52.967246+00:00"},{"alias_kind":"pith_short_12","alias_value":"XFNGOAXY2MLC","created_at":"2026-07-05T09:21:52.967246+00:00"},{"alias_kind":"pith_short_16","alias_value":"XFNGOAXY2MLCLSSJ","created_at":"2026-07-05T09:21:52.967246+00:00"},{"alias_kind":"pith_short_8","alias_value":"XFNGOAXY","created_at":"2026-07-05T09:21:52.967246+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2601.03559","citing_title":"DiffCoT: Diffusion-styled Chain-of-Thought Reasoning in LLMs","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25039","citing_title":"Dual-Track CoT: Budget-Aware Stepwise Guidance for Small LMs","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ","json":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ.json","graph_json":"https://pith.science/api/pith-number/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/graph.json","events_json":"https://pith.science/api/pith-number/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/events.json","paper":"https://pith.science/paper/XFNGOAXY"},"agent_actions":{"view_html":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ","download_json":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ.json","view_paper":"https://pith.science/paper/XFNGOAXY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.12934&json=true","fetch_graph":"https://pith.science/api/pith-number/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/graph.json","fetch_events":"https://pith.science/api/pith-number/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/action/storage_attestation","attest_author":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/action/author_attestation","sign_citation":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/action/citation_signature","submit_replication":"https://pith.science/pith/XFNGOAXY2MLCLSSJ5M7FKR7WLJ/action/replication_record"}},"created_at":"2026-07-05T09:21:52.967246+00:00","updated_at":"2026-07-05T09:21:52.967246+00:00"}