{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SXJKX4O647D34ED7QHN36DFEUW","short_pith_number":"pith:SXJKX4O6","schema_version":"1.0","canonical_sha256":"95d2abf1dee7c7be107f81dbbf0ca4a59bed2d92e5799e4ae3e6bec24243ca52","source":{"kind":"arxiv","id":"2505.15927","version":1},"attestation_state":"computed","paper":{"title":"CoT Information: Improved Sample Complexity under Chain-of-Thought Supervision","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Awni Altabaa, John Lafferty, Omar Montasser","submitted_at":"2025-05-21T18:28:54Z","abstract_excerpt":"Learning complex functions that involve multi-step reasoning poses a significant challenge for standard supervised learning from input-output examples. Chain-of-thought (CoT) supervision, which provides intermediate reasoning steps together with the final output, has emerged as a powerful empirical technique, underpinning much of the recent progress in the reasoning capabilities of large language models. This paper develops a statistical theory of learning under CoT supervision. A key characteristic of the CoT setting, in contrast to standard supervision, is the mismatch between the training o"},"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":"2505.15927","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2025-05-21T18:28:54Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"a2fcc5f526f34a41d8899d0cd5e6eef4f719faddf29e6047121408cf500fefc7","abstract_canon_sha256":"9cb6766f67b71338f79ab9395f765857b94ae896009a0aed56aefc496861b201"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:07:27.430672Z","signature_b64":"o4hZZX7scJ6Cvr9tceZid5b+vCtdt6ZtyVV/tV4ZWeW/LnzbkN+xzewJVlqf6RNHVm9RlqqzxkIEVQgHRlAjBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"95d2abf1dee7c7be107f81dbbf0ca4a59bed2d92e5799e4ae3e6bec24243ca52","last_reissued_at":"2026-07-05T11:07:27.430182Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:07:27.430182Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CoT Information: Improved Sample Complexity under Chain-of-Thought Supervision","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Awni Altabaa, John Lafferty, Omar Montasser","submitted_at":"2025-05-21T18:28:54Z","abstract_excerpt":"Learning complex functions that involve multi-step reasoning poses a significant challenge for standard supervised learning from input-output examples. Chain-of-thought (CoT) supervision, which provides intermediate reasoning steps together with the final output, has emerged as a powerful empirical technique, underpinning much of the recent progress in the reasoning capabilities of large language models. This paper develops a statistical theory of learning under CoT supervision. A key characteristic of the CoT setting, in contrast to standard supervision, is the mismatch between the training o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.15927","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/2505.15927/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":"2505.15927","created_at":"2026-07-05T11:07:27.430243+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.15927v1","created_at":"2026-07-05T11:07:27.430243+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.15927","created_at":"2026-07-05T11:07:27.430243+00:00"},{"alias_kind":"pith_short_12","alias_value":"SXJKX4O647D3","created_at":"2026-07-05T11:07:27.430243+00:00"},{"alias_kind":"pith_short_16","alias_value":"SXJKX4O647D34ED7","created_at":"2026-07-05T11:07:27.430243+00:00"},{"alias_kind":"pith_short_8","alias_value":"SXJKX4O6","created_at":"2026-07-05T11:07:27.430243+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.12013","citing_title":"Sample Complexity of Autoregressive Reasoning: Chain-of-Thought vs. End-to-End","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06819","citing_title":"A Theory of Online Learning with Autoregressive Chain-of-Thought Reasoning","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW","json":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW.json","graph_json":"https://pith.science/api/pith-number/SXJKX4O647D34ED7QHN36DFEUW/graph.json","events_json":"https://pith.science/api/pith-number/SXJKX4O647D34ED7QHN36DFEUW/events.json","paper":"https://pith.science/paper/SXJKX4O6"},"agent_actions":{"view_html":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW","download_json":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW.json","view_paper":"https://pith.science/paper/SXJKX4O6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.15927&json=true","fetch_graph":"https://pith.science/api/pith-number/SXJKX4O647D34ED7QHN36DFEUW/graph.json","fetch_events":"https://pith.science/api/pith-number/SXJKX4O647D34ED7QHN36DFEUW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW/action/storage_attestation","attest_author":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW/action/author_attestation","sign_citation":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW/action/citation_signature","submit_replication":"https://pith.science/pith/SXJKX4O647D34ED7QHN36DFEUW/action/replication_record"}},"created_at":"2026-07-05T11:07:27.430243+00:00","updated_at":"2026-07-05T11:07:27.430243+00:00"}