{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:TOUS4IHOBE25O3Z5RMRXSDDR4D","short_pith_number":"pith:TOUS4IHO","canonical_record":{"source":{"id":"2203.11171","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:48:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8334a25495da89b12199124143689463367efebd64349e87f425d9b591cb6a20","abstract_canon_sha256":"91273e2e94f39d606d664234ff47cf5f017641df5b96e1b6289306ddcd6b33ab"},"schema_version":"1.0"},"canonical_sha256":"9ba92e20ee0935d76f3d8b23790c71e0e61c155b55cf6426e1e12035bb524f9b","source":{"kind":"arxiv","id":"2203.11171","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.11171","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"arxiv_version","alias_value":"2203.11171v4","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.11171","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"pith_short_12","alias_value":"TOUS4IHOBE25","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"pith_short_16","alias_value":"TOUS4IHOBE25O3Z5","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"pith_short_8","alias_value":"TOUS4IHO","created_at":"2026-07-05T05:48:32Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:TOUS4IHOBE25O3Z5RMRXSDDR4D","target":"record","payload":{"canonical_record":{"source":{"id":"2203.11171","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:48:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8334a25495da89b12199124143689463367efebd64349e87f425d9b591cb6a20","abstract_canon_sha256":"91273e2e94f39d606d664234ff47cf5f017641df5b96e1b6289306ddcd6b33ab"},"schema_version":"1.0"},"canonical_sha256":"9ba92e20ee0935d76f3d8b23790c71e0e61c155b55cf6426e1e12035bb524f9b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:48:32.918461Z","signature_b64":"KI/cmkeEOwRvGQtLesmIP6iMTq29eOVyxY3j25+BpHQGJGRLqFSfxTWZVMfgCusxiDPSWDaJrMm+ZmMIQr1QAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9ba92e20ee0935d76f3d8b23790c71e0e61c155b55cf6426e1e12035bb524f9b","last_reissued_at":"2026-07-05T05:48:32.918056Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:48:32.918056Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2203.11171","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:48:32Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gsXzglf1yEpj+la/JSKv6dKGCUVBHfUTMaIza5AE+wotyx+sVZkI4piFSBxdTh5BcjnuTucBcPTcigiCeUO/CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T11:59:39.101423Z"},"content_sha256":"418f26bd607051baebc40c4665fd12270ffa5093fb44274d64af174082bd12ba","schema_version":"1.0","event_id":"sha256:418f26bd607051baebc40c4665fd12270ffa5093fb44274d64af174082bd12ba"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:TOUS4IHOBE25O3Z5RMRXSDDR4D","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Self-Consistency Improves Chain of Thought Reasoning in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Self-consistency decoding replaces greedy search in chain-of-thought prompting and raises accuracy on reasoning benchmarks.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aakanksha Chowdhery, Dale Schuurmans, Denny Zhou, Ed Chi, Jason Wei, Quoc Le, Sharan Narang, Xuezhi Wang","submitted_at":"2022-03-21T17:48:52Z","abstract_excerpt":"Chain-of-thought prompting combined with pre-trained large language models has achieved encouraging results on complex reasoning tasks. In this paper, we propose a new decoding strategy, self-consistency, to replace the naive greedy decoding used in chain-of-thought prompting. It first samples a diverse set of reasoning paths instead of only taking the greedy one, and then selects the most consistent answer by marginalizing out the sampled reasoning paths. Self-consistency leverages the intuition that a complex reasoning problem typically admits multiple different ways of thinking leading to i"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"self-consistency boosts the performance of chain-of-thought prompting with a striking margin on a range of popular arithmetic and commonsense reasoning benchmarks, including GSM8K (+17.9%), SVAMP (+11.0%), AQuA (+12.2%), StrategyQA (+6.4%) and ARC-challenge (+3.9%).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the language model can generate a sufficiently diverse set of correct reasoning paths for problems where the answer is unique, so that majority voting over paths reliably surfaces the correct answer rather than an incorrect but consistent one; this enters in the description of the decoding procedure and the empirical evaluation setup.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Self-consistency decoding samples multiple reasoning paths and marginalizes to the most consistent answer, raising accuracy on GSM8K by 17.9 points and similar gains on SVAMP, AQuA, StrategyQA, and ARC-challenge.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Self-consistency decoding replaces greedy search in chain-of-thought prompting and raises accuracy on reasoning benchmarks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4d063756f7338b7f5ea1ae208fafa4b53b8cee5d6e12d44340288a426a6f8139"},"source":{"id":"2203.11171","kind":"arxiv","version":4},"verdict":{"id":"da004527-da46-44e9-b60b-74dd33637acd","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-24T11:35:43.766588Z","strongest_claim":"self-consistency boosts the performance of chain-of-thought prompting with a striking margin on a range of popular arithmetic and commonsense reasoning benchmarks, including GSM8K (+17.9%), SVAMP (+11.0%), AQuA (+12.2%), StrategyQA (+6.4%) and ARC-challenge (+3.9%).","one_line_summary":"Self-consistency decoding samples multiple reasoning paths and marginalizes to the most consistent answer, raising accuracy on GSM8K by 17.9 points and similar gains on SVAMP, AQuA, StrategyQA, and ARC-challenge.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the language model can generate a sufficiently diverse set of correct reasoning paths for problems where the answer is unique, so that majority voting over paths reliably surfaces the correct answer rather than an incorrect but consistent one; this enters in the description of the decoding procedure and the empirical evaluation setup.","pith_extraction_headline":"Self-consistency decoding replaces greedy search in chain-of-thought prompting and raises accuracy on reasoning benchmarks."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2203.11171/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"},"verdict_id":"da004527-da46-44e9-b60b-74dd33637acd"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:48:32Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uGGQM6lVoam8qxFxecd2zv82rBXcVsI2pZD8Hkc2ERJFHK2vpVUy2Cq+MIlvtdBV2PMa8SFTy6W9qN8q5jptCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T11:59:39.101869Z"},"content_sha256":"472ccb52f2a497254e56ded1b5bf2e00e6366f0bed7787acc1264aae2991f527","schema_version":"1.0","event_id":"sha256:472ccb52f2a497254e56ded1b5bf2e00e6366f0bed7787acc1264aae2991f527"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D/bundle.json","state_url":"https://pith.science/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-07T11:59:39Z","links":{"resolver":"https://pith.science/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D","bundle":"https://pith.science/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D/bundle.json","state":"https://pith.science/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TOUS4IHOBE25O3Z5RMRXSDDR4D/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:TOUS4IHOBE25O3Z5RMRXSDDR4D","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":"91273e2e94f39d606d664234ff47cf5f017641df5b96e1b6289306ddcd6b33ab","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:48:52Z","title_canon_sha256":"8334a25495da89b12199124143689463367efebd64349e87f425d9b591cb6a20"},"schema_version":"1.0","source":{"id":"2203.11171","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.11171","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"arxiv_version","alias_value":"2203.11171v4","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.11171","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"pith_short_12","alias_value":"TOUS4IHOBE25","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"pith_short_16","alias_value":"TOUS4IHOBE25O3Z5","created_at":"2026-07-05T05:48:32Z"},{"alias_kind":"pith_short_8","alias_value":"TOUS4IHO","created_at":"2026-07-05T05:48:32Z"}],"graph_snapshots":[{"event_id":"sha256:472ccb52f2a497254e56ded1b5bf2e00e6366f0bed7787acc1264aae2991f527","target":"graph","created_at":"2026-07-05T05:48:32Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"self-consistency boosts the performance of chain-of-thought prompting with a striking margin on a range of popular arithmetic and commonsense reasoning benchmarks, including GSM8K (+17.9%), SVAMP (+11.0%), AQuA (+12.2%), StrategyQA (+6.4%) and ARC-challenge (+3.9%)."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that the language model can generate a sufficiently diverse set of correct reasoning paths for problems where the answer is unique, so that majority voting over paths reliably surfaces the correct answer rather than an incorrect but consistent one; this enters in the description of the decoding procedure and the empirical evaluation setup."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Self-consistency decoding samples multiple reasoning paths and marginalizes to the most consistent answer, raising accuracy on GSM8K by 17.9 points and similar gains on SVAMP, AQuA, StrategyQA, and ARC-challenge."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Self-consistency decoding replaces greedy search in chain-of-thought prompting and raises accuracy on reasoning benchmarks."}],"snapshot_sha256":"4d063756f7338b7f5ea1ae208fafa4b53b8cee5d6e12d44340288a426a6f8139"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2203.11171/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Chain-of-thought prompting combined with pre-trained large language models has achieved encouraging results on complex reasoning tasks. In this paper, we propose a new decoding strategy, self-consistency, to replace the naive greedy decoding used in chain-of-thought prompting. It first samples a diverse set of reasoning paths instead of only taking the greedy one, and then selects the most consistent answer by marginalizing out the sampled reasoning paths. Self-consistency leverages the intuition that a complex reasoning problem typically admits multiple different ways of thinking leading to i","authors_text":"Aakanksha Chowdhery, Dale Schuurmans, Denny Zhou, Ed Chi, Jason Wei, Quoc Le, Sharan Narang, Xuezhi Wang","cross_cats":["cs.AI"],"headline":"Self-consistency decoding replaces greedy search in chain-of-thought prompting and raises accuracy on reasoning benchmarks.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:48:52Z","title":"Self-Consistency Improves Chain of Thought Reasoning in Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.11171","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-24T11:35:43.766588Z","id":"da004527-da46-44e9-b60b-74dd33637acd","model_set":{"reader":"grok-4.3"},"one_line_summary":"Self-consistency decoding samples multiple reasoning paths and marginalizes to the most consistent answer, raising accuracy on GSM8K by 17.9 points and similar gains on SVAMP, AQuA, StrategyQA, and ARC-challenge.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Self-consistency decoding replaces greedy search in chain-of-thought prompting and raises accuracy on reasoning benchmarks.","strongest_claim":"self-consistency boosts the performance of chain-of-thought prompting with a striking margin on a range of popular arithmetic and commonsense reasoning benchmarks, including GSM8K (+17.9%), SVAMP (+11.0%), AQuA (+12.2%), StrategyQA (+6.4%) and ARC-challenge (+3.9%).","weakest_assumption":"The assumption that the language model can generate a sufficiently diverse set of correct reasoning paths for problems where the answer is unique, so that majority voting over paths reliably surfaces the correct answer rather than an incorrect but consistent one; this enters in the description of the decoding procedure and the empirical evaluation setup."}},"verdict_id":"da004527-da46-44e9-b60b-74dd33637acd"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:418f26bd607051baebc40c4665fd12270ffa5093fb44274d64af174082bd12ba","target":"record","created_at":"2026-07-05T05:48:32Z","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":"91273e2e94f39d606d664234ff47cf5f017641df5b96e1b6289306ddcd6b33ab","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:48:52Z","title_canon_sha256":"8334a25495da89b12199124143689463367efebd64349e87f425d9b591cb6a20"},"schema_version":"1.0","source":{"id":"2203.11171","kind":"arxiv","version":4}},"canonical_sha256":"9ba92e20ee0935d76f3d8b23790c71e0e61c155b55cf6426e1e12035bb524f9b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9ba92e20ee0935d76f3d8b23790c71e0e61c155b55cf6426e1e12035bb524f9b","first_computed_at":"2026-07-05T05:48:32.918056Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:48:32.918056Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KI/cmkeEOwRvGQtLesmIP6iMTq29eOVyxY3j25+BpHQGJGRLqFSfxTWZVMfgCusxiDPSWDaJrMm+ZmMIQr1QAw==","signature_status":"signed_v1","signed_at":"2026-07-05T05:48:32.918461Z","signed_message":"canonical_sha256_bytes"},"source_id":"2203.11171","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:418f26bd607051baebc40c4665fd12270ffa5093fb44274d64af174082bd12ba","sha256:472ccb52f2a497254e56ded1b5bf2e00e6366f0bed7787acc1264aae2991f527"],"state_sha256":"c361382a1939cefe2989ab47c30d3ebee1aecbd990c956764ce13822dd321754"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4DhFGJ5jRcSwMh/ghb5Fh4qa2erEPHAKH2D8e0q39j5goUfV18OWQb8faFvkRGRYhC7TGoEBmblmmPLkL7VNBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T11:59:39.103928Z","bundle_sha256":"3aa39db2baaffa1997c7ea10487b5f46a4078d33129ed42c76436c98b47dceaf"}}