{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5PURD63EOZSO3PBICFK5OKRFBN","short_pith_number":"pith:5PURD63E","schema_version":"1.0","canonical_sha256":"ebe911fb647664edbc281155d72a250b5f963b35c9270b25a73604d178bdfca1","source":{"kind":"arxiv","id":"2505.18440","version":2},"attestation_state":"computed","paper":{"title":"Efficient Long CoT Reasoning in Small Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Huaxiu Yao, Hui Liu, Jinqi Jiang, Tian Qiu, Xianfeng Tang, Zhaoyang Wang","submitted_at":"2025-05-24T00:22:52Z","abstract_excerpt":"Recent large reasoning models such as DeepSeek-R1 exhibit strong complex problems solving abilities by generating long chain-of-thought (CoT) reasoning steps. It is challenging to directly train small language models (SLMs) to emerge long CoT. Thus, distillation becomes a practical method to enable SLMs for such reasoning ability. However, the long CoT often contains a lot of redundant contents (e.g., overthinking steps) which may make SLMs hard to learn considering their relatively poor capacity and generalization. To address this issue, we propose a simple-yet-effective method to prune unnec"},"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":"2505.18440","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-24T00:22:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"387f2e76ede752ab0cdb2a85e4df906c6096c978fb6378b3b452ed7b1d81477f","abstract_canon_sha256":"ad0d94543fc0400c5cb2ceaaec0584e15e1b4ab8b1eda2ff6a3a4dca69702146"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:28.514444Z","signature_b64":"/gzJvLLKTaav7qme2MkLtudxbp8l+YI/T7uTAA8nTKJ272/7TgiGPDK7AoIF+o0vc5cdlwN2P91UwtuKiaOPCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebe911fb647664edbc281155d72a250b5f963b35c9270b25a73604d178bdfca1","last_reissued_at":"2026-07-05T11:23:28.513911Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:28.513911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Long CoT Reasoning in Small Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Huaxiu Yao, Hui Liu, Jinqi Jiang, Tian Qiu, Xianfeng Tang, Zhaoyang Wang","submitted_at":"2025-05-24T00:22:52Z","abstract_excerpt":"Recent large reasoning models such as DeepSeek-R1 exhibit strong complex problems solving abilities by generating long chain-of-thought (CoT) reasoning steps. It is challenging to directly train small language models (SLMs) to emerge long CoT. Thus, distillation becomes a practical method to enable SLMs for such reasoning ability. However, the long CoT often contains a lot of redundant contents (e.g., overthinking steps) which may make SLMs hard to learn considering their relatively poor capacity and generalization. To address this issue, we propose a simple-yet-effective method to prune unnec"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18440","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/2505.18440/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":"2505.18440","created_at":"2026-07-05T11:23:28.513971+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.18440v2","created_at":"2026-07-05T11:23:28.513971+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18440","created_at":"2026-07-05T11:23:28.513971+00:00"},{"alias_kind":"pith_short_12","alias_value":"5PURD63EOZSO","created_at":"2026-07-05T11:23:28.513971+00:00"},{"alias_kind":"pith_short_16","alias_value":"5PURD63EOZSO3PBI","created_at":"2026-07-05T11:23:28.513971+00:00"},{"alias_kind":"pith_short_8","alias_value":"5PURD63E","created_at":"2026-07-05T11:23:28.513971+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN","json":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN.json","graph_json":"https://pith.science/api/pith-number/5PURD63EOZSO3PBICFK5OKRFBN/graph.json","events_json":"https://pith.science/api/pith-number/5PURD63EOZSO3PBICFK5OKRFBN/events.json","paper":"https://pith.science/paper/5PURD63E"},"agent_actions":{"view_html":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN","download_json":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN.json","view_paper":"https://pith.science/paper/5PURD63E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.18440&json=true","fetch_graph":"https://pith.science/api/pith-number/5PURD63EOZSO3PBICFK5OKRFBN/graph.json","fetch_events":"https://pith.science/api/pith-number/5PURD63EOZSO3PBICFK5OKRFBN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN/action/storage_attestation","attest_author":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN/action/author_attestation","sign_citation":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN/action/citation_signature","submit_replication":"https://pith.science/pith/5PURD63EOZSO3PBICFK5OKRFBN/action/replication_record"}},"created_at":"2026-07-05T11:23:28.513971+00:00","updated_at":"2026-07-05T11:23:28.513971+00:00"}