{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NDOZ6LRGOOMNHVOCJ3YX5M6UAQ","short_pith_number":"pith:NDOZ6LRG","schema_version":"1.0","canonical_sha256":"68dd9f2e267398d3d5c24ef17eb3d4041d24968f3bbcdb8346a62ce2e4169086","source":{"kind":"arxiv","id":"2403.09849","version":1},"attestation_state":"computed","paper":{"title":"Self-Consistency Boosts Calibration for Math Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ante Wang, Baolin Peng, Dong Yu, Haitao Mi, Jinsong Su, Lifeng Jin, Linfeng Song, Ye Tian","submitted_at":"2024-03-14T20:17:10Z","abstract_excerpt":"Calibration, which establishes the correlation between accuracy and model confidence, is important for LLM development. We design three off-the-shelf calibration methods based on self-consistency (Wang et al., 2022) for math reasoning tasks. Evaluation on two popular benchmarks (GSM8K and MathQA) using strong open-source LLMs (Mistral and LLaMA2), our methods better bridge model confidence and accuracy than existing methods based on p(True) (Kadavath et al., 2022) or logit (Kadavath et al., 2022)."},"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":"2403.09849","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-14T20:17:10Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7551d867c3ef313215e549411eb99acef7f76ef83c4ecbe03bf6ae79a2e0e6b9","abstract_canon_sha256":"4ef94f0553df7b034d400164b3f874eb66285d0ab7a1a40f7405f7f878e4563e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:56:26.377617Z","signature_b64":"E13qoJmmFyEIyFIhzmPIzJv4BQV6Oc8WsvpAQKhZcI1omCCFTLwyv1LHJ5F7E9/C09MQOuFeJunoy/FVmlfYBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"68dd9f2e267398d3d5c24ef17eb3d4041d24968f3bbcdb8346a62ce2e4169086","last_reissued_at":"2026-07-05T07:56:26.377095Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:56:26.377095Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-Consistency Boosts Calibration for Math Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ante Wang, Baolin Peng, Dong Yu, Haitao Mi, Jinsong Su, Lifeng Jin, Linfeng Song, Ye Tian","submitted_at":"2024-03-14T20:17:10Z","abstract_excerpt":"Calibration, which establishes the correlation between accuracy and model confidence, is important for LLM development. We design three off-the-shelf calibration methods based on self-consistency (Wang et al., 2022) for math reasoning tasks. Evaluation on two popular benchmarks (GSM8K and MathQA) using strong open-source LLMs (Mistral and LLaMA2), our methods better bridge model confidence and accuracy than existing methods based on p(True) (Kadavath et al., 2022) or logit (Kadavath et al., 2022)."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.09849","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/2403.09849/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":"2403.09849","created_at":"2026-07-05T07:56:26.377153+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.09849v1","created_at":"2026-07-05T07:56:26.377153+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.09849","created_at":"2026-07-05T07:56:26.377153+00:00"},{"alias_kind":"pith_short_12","alias_value":"NDOZ6LRGOOMN","created_at":"2026-07-05T07:56:26.377153+00:00"},{"alias_kind":"pith_short_16","alias_value":"NDOZ6LRGOOMNHVOC","created_at":"2026-07-05T07:56:26.377153+00:00"},{"alias_kind":"pith_short_8","alias_value":"NDOZ6LRG","created_at":"2026-07-05T07:56:26.377153+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17312","citing_title":"Quantifying Consistency in LLM Logical Reasoning via Structural Uncertainty","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ","json":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ.json","graph_json":"https://pith.science/api/pith-number/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/graph.json","events_json":"https://pith.science/api/pith-number/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/events.json","paper":"https://pith.science/paper/NDOZ6LRG"},"agent_actions":{"view_html":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ","download_json":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ.json","view_paper":"https://pith.science/paper/NDOZ6LRG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.09849&json=true","fetch_graph":"https://pith.science/api/pith-number/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/graph.json","fetch_events":"https://pith.science/api/pith-number/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/action/storage_attestation","attest_author":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/action/author_attestation","sign_citation":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/action/citation_signature","submit_replication":"https://pith.science/pith/NDOZ6LRGOOMNHVOCJ3YX5M6UAQ/action/replication_record"}},"created_at":"2026-07-05T07:56:26.377153+00:00","updated_at":"2026-07-05T07:56:26.377153+00:00"}