{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:M2EDXAGA4SC6DNIBNSCKSASIDO","short_pith_number":"pith:M2EDXAGA","schema_version":"1.0","canonical_sha256":"66883b80c0e485e1b5016c84a902481b9e057178515d8d95c5b020cb9dbf3d5e","source":{"kind":"arxiv","id":"2407.03181","version":2},"attestation_state":"computed","paper":{"title":"Fine-Tuning on Diverse Reasoning Chains Drives Within-Inference CoT Refinement in LLMs","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Harish Tayyar Madabushi, Haritz Puerto, Iryna Gurevych, Tilek Chubakov, Xiaodan Zhu","submitted_at":"2024-07-03T15:01:18Z","abstract_excerpt":"Requiring a large language model (LLM) to generate intermediary reasoning steps, known as Chain of Thought (CoT), has been shown to be an effective way of boosting performance. Previous approaches have focused on generating multiple independent CoTs, combining them through ensembling or other post-hoc strategies to enhance reasoning. In this work, we introduce a novel approach where LLMs are fine-tuned to generate a sequence of Diverse Chains of Thought (DCoT) within a single inference step, which is fundamentally different from prior work that primarily operate on parallel CoT generations. DC"},"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":"2407.03181","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-03T15:01:18Z","cross_cats_sorted":[],"title_canon_sha256":"b17dcfd0096da82205b79da4b188df4e9201462e4906369ffef5a51fed1a03b5","abstract_canon_sha256":"732be2f6a7ccca7646d3cbeba0c74670392efecc4f92d2fa05d6278a19aaa6b9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:31.722811Z","signature_b64":"SOfvwXhCZ6klZDVGLFRRoo/pF5G/Qgj+FxZeRTx5iHRPi6/3yoyh0CElQvzOh/44Vip+xdgvSUATKvVLFHu6Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66883b80c0e485e1b5016c84a902481b9e057178515d8d95c5b020cb9dbf3d5e","last_reissued_at":"2026-07-05T11:10:31.722401Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:31.722401Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fine-Tuning on Diverse Reasoning Chains Drives Within-Inference CoT Refinement in LLMs","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Harish Tayyar Madabushi, Haritz Puerto, Iryna Gurevych, Tilek Chubakov, Xiaodan Zhu","submitted_at":"2024-07-03T15:01:18Z","abstract_excerpt":"Requiring a large language model (LLM) to generate intermediary reasoning steps, known as Chain of Thought (CoT), has been shown to be an effective way of boosting performance. Previous approaches have focused on generating multiple independent CoTs, combining them through ensembling or other post-hoc strategies to enhance reasoning. In this work, we introduce a novel approach where LLMs are fine-tuned to generate a sequence of Diverse Chains of Thought (DCoT) within a single inference step, which is fundamentally different from prior work that primarily operate on parallel CoT generations. DC"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.03181","kind":"arxiv","version":2},"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/2407.03181/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":"2407.03181","created_at":"2026-07-05T11:10:31.722461+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.03181v2","created_at":"2026-07-05T11:10:31.722461+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.03181","created_at":"2026-07-05T11:10:31.722461+00:00"},{"alias_kind":"pith_short_12","alias_value":"M2EDXAGA4SC6","created_at":"2026-07-05T11:10:31.722461+00:00"},{"alias_kind":"pith_short_16","alias_value":"M2EDXAGA4SC6DNIB","created_at":"2026-07-05T11:10:31.722461+00:00"},{"alias_kind":"pith_short_8","alias_value":"M2EDXAGA","created_at":"2026-07-05T11:10:31.722461+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.08221","citing_title":"NoisyCoconut: Counterfactual Consensus via Latent Space Reasoning","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO","json":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO.json","graph_json":"https://pith.science/api/pith-number/M2EDXAGA4SC6DNIBNSCKSASIDO/graph.json","events_json":"https://pith.science/api/pith-number/M2EDXAGA4SC6DNIBNSCKSASIDO/events.json","paper":"https://pith.science/paper/M2EDXAGA"},"agent_actions":{"view_html":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO","download_json":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO.json","view_paper":"https://pith.science/paper/M2EDXAGA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.03181&json=true","fetch_graph":"https://pith.science/api/pith-number/M2EDXAGA4SC6DNIBNSCKSASIDO/graph.json","fetch_events":"https://pith.science/api/pith-number/M2EDXAGA4SC6DNIBNSCKSASIDO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO/action/storage_attestation","attest_author":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO/action/author_attestation","sign_citation":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO/action/citation_signature","submit_replication":"https://pith.science/pith/M2EDXAGA4SC6DNIBNSCKSASIDO/action/replication_record"}},"created_at":"2026-07-05T11:10:31.722461+00:00","updated_at":"2026-07-05T11:10:31.722461+00:00"}