{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NHRPTQBLMKFE5H6HARC563LUVR","short_pith_number":"pith:NHRPTQBL","schema_version":"1.0","canonical_sha256":"69e2f9c02b628a4e9fc70445df6d74ac53c664cd3d7c2a7bf1cc63fa155bc94b","source":{"kind":"arxiv","id":"2503.20641","version":2},"attestation_state":"computed","paper":{"title":"Unlocking Efficient Long-to-Short LLM Reasoning with Model Merging","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Han Wu, Hui-Ling Zhen, Mingxuan Yuan, Shuqi Liu, Tao Zhong, Xiaojin Fu, Xing Li, Xiongwei Han, Yuxuan Yao, Zehua Liu","submitted_at":"2025-03-26T15:34:37Z","abstract_excerpt":"The transition from System 1 to System 2 reasoning in large language models (LLMs) has marked significant advancements in handling complex tasks through deliberate, iterative thinking. However, this progress often comes at the cost of efficiency, as models tend to overthink, generating redundant reasoning steps without proportional improvements in output quality. Long-to-Short (L2S) reasoning has emerged as a promising solution to this challenge, aiming to balance reasoning depth with practical efficiency. While existing approaches, such as supervised fine-tuning (SFT), reinforcement learning "},"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":"2503.20641","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-26T15:34:37Z","cross_cats_sorted":[],"title_canon_sha256":"699c36f9e732e7a607696681804a55d2c0efc05e87b6a48552bb16b6300e97d3","abstract_canon_sha256":"f30ce9c8951e627e7c4f73bc8713552e0afc38485f4c1e2e5617cedfc5e68154"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:08:18.348646Z","signature_b64":"0BTqOJ9X/R0fbBbeMuU/6Vv8l884YWE2lFs7BISeuSr30zbyn2eZc9GxXxbqphxyGYpn2k+OBI+GX8gtDt4VAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69e2f9c02b628a4e9fc70445df6d74ac53c664cd3d7c2a7bf1cc63fa155bc94b","last_reissued_at":"2026-07-05T11:08:18.348094Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:08:18.348094Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unlocking Efficient Long-to-Short LLM Reasoning with Model Merging","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Han Wu, Hui-Ling Zhen, Mingxuan Yuan, Shuqi Liu, Tao Zhong, Xiaojin Fu, Xing Li, Xiongwei Han, Yuxuan Yao, Zehua Liu","submitted_at":"2025-03-26T15:34:37Z","abstract_excerpt":"The transition from System 1 to System 2 reasoning in large language models (LLMs) has marked significant advancements in handling complex tasks through deliberate, iterative thinking. However, this progress often comes at the cost of efficiency, as models tend to overthink, generating redundant reasoning steps without proportional improvements in output quality. Long-to-Short (L2S) reasoning has emerged as a promising solution to this challenge, aiming to balance reasoning depth with practical efficiency. While existing approaches, such as supervised fine-tuning (SFT), reinforcement learning "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.20641","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/2503.20641/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":"2503.20641","created_at":"2026-07-05T11:08:18.348153+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.20641v2","created_at":"2026-07-05T11:08:18.348153+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.20641","created_at":"2026-07-05T11:08:18.348153+00:00"},{"alias_kind":"pith_short_12","alias_value":"NHRPTQBLMKFE","created_at":"2026-07-05T11:08:18.348153+00:00"},{"alias_kind":"pith_short_16","alias_value":"NHRPTQBLMKFE5H6H","created_at":"2026-07-05T11:08:18.348153+00:00"},{"alias_kind":"pith_short_8","alias_value":"NHRPTQBL","created_at":"2026-07-05T11:08:18.348153+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03344","citing_title":"RogueMerge: Robust and Unified Attacks against LLM Model Merging","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18083","citing_title":"A Data-Efficient Path to Multilingual LLMs: Language Expansion via Post-training PARAM$\\Delta$ Integration into Upcycled MoE","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2408.07666","citing_title":"Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities","ref_index":256,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04120","citing_title":"Shorter, but Still Trustworthy? An Empirical Study of Chain-of-Thought Compression","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12384","citing_title":"Scalable Token-Level Hallucination Detection in Large Language Models","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06111","citing_title":"Schedule-and-Calibrate: Utility-Guided Multi-Task Reinforcement Learning for Code LLMs","ref_index":34,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR","json":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR.json","graph_json":"https://pith.science/api/pith-number/NHRPTQBLMKFE5H6HARC563LUVR/graph.json","events_json":"https://pith.science/api/pith-number/NHRPTQBLMKFE5H6HARC563LUVR/events.json","paper":"https://pith.science/paper/NHRPTQBL"},"agent_actions":{"view_html":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR","download_json":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR.json","view_paper":"https://pith.science/paper/NHRPTQBL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.20641&json=true","fetch_graph":"https://pith.science/api/pith-number/NHRPTQBLMKFE5H6HARC563LUVR/graph.json","fetch_events":"https://pith.science/api/pith-number/NHRPTQBLMKFE5H6HARC563LUVR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR/action/storage_attestation","attest_author":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR/action/author_attestation","sign_citation":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR/action/citation_signature","submit_replication":"https://pith.science/pith/NHRPTQBLMKFE5H6HARC563LUVR/action/replication_record"}},"created_at":"2026-07-05T11:08:18.348153+00:00","updated_at":"2026-07-05T11:08:18.348153+00:00"}