{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IDZ7AHMZVCGRFDZXW5DOH3SEIO","short_pith_number":"pith:IDZ7AHMZ","schema_version":"1.0","canonical_sha256":"40f3f01d99a88d128f37b746e3ee4443977cba42177115387746075282ab71a1","source":{"kind":"arxiv","id":"2506.07104","version":2},"attestation_state":"computed","paper":{"title":"How Far Are We from Optimal Reasoning Efficiency?","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jiaxuan Gao, Kaifeng Lyu, Lu Yang, Qixin Tan, Shusheng Xu, Shu Yan, Wei Fu, Yi Wu, Zhiyu Mei","submitted_at":"2025-06-08T12:18:50Z","abstract_excerpt":"Large Reasoning Models (LRMs) demonstrate remarkable problem-solving capabilities through extended Chain-of-Thought (CoT) reasoning but often produce excessively verbose and redundant reasoning traces. This inefficiency incurs high inference costs and limits practical deployment. While existing fine-tuning methods aim to improve reasoning efficiency, assessing their efficiency gains remains challenging due to inconsistent evaluations. In this work, we introduce the reasoning efficiency frontiers, empirical upper bounds derived from fine-tuning base LRMs across diverse approaches and training c"},"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":"2506.07104","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-08T12:18:50Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7365253279ef5e1773ecad58b00bd6361b00f14694439bcb67bb198cca281935","abstract_canon_sha256":"dae0ee78da6ed2a9cc90833102dd8133cb9a54b920db1c83edc538dd0c52b01b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:08:03.921839Z","signature_b64":"GAMW0D15gKr4v0F/pt8SYtsBGQAeg+SVAlDOQJJKIQtDMwlqwzcuJk8qDLro2qjsNfeR0Teu0wzsCpQTQA5dBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"40f3f01d99a88d128f37b746e3ee4443977cba42177115387746075282ab71a1","last_reissued_at":"2026-07-05T12:08:03.921291Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:08:03.921291Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Far Are We from Optimal Reasoning Efficiency?","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jiaxuan Gao, Kaifeng Lyu, Lu Yang, Qixin Tan, Shusheng Xu, Shu Yan, Wei Fu, Yi Wu, Zhiyu Mei","submitted_at":"2025-06-08T12:18:50Z","abstract_excerpt":"Large Reasoning Models (LRMs) demonstrate remarkable problem-solving capabilities through extended Chain-of-Thought (CoT) reasoning but often produce excessively verbose and redundant reasoning traces. This inefficiency incurs high inference costs and limits practical deployment. While existing fine-tuning methods aim to improve reasoning efficiency, assessing their efficiency gains remains challenging due to inconsistent evaluations. In this work, we introduce the reasoning efficiency frontiers, empirical upper bounds derived from fine-tuning base LRMs across diverse approaches and training c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.07104","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/2506.07104/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":"2506.07104","created_at":"2026-07-05T12:08:03.921359+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.07104v2","created_at":"2026-07-05T12:08:03.921359+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.07104","created_at":"2026-07-05T12:08:03.921359+00:00"},{"alias_kind":"pith_short_12","alias_value":"IDZ7AHMZVCGR","created_at":"2026-07-05T12:08:03.921359+00:00"},{"alias_kind":"pith_short_16","alias_value":"IDZ7AHMZVCGRFDZX","created_at":"2026-07-05T12:08:03.921359+00:00"},{"alias_kind":"pith_short_8","alias_value":"IDZ7AHMZ","created_at":"2026-07-05T12:08:03.921359+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2601.11340","citing_title":"Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Models","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO","json":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO.json","graph_json":"https://pith.science/api/pith-number/IDZ7AHMZVCGRFDZXW5DOH3SEIO/graph.json","events_json":"https://pith.science/api/pith-number/IDZ7AHMZVCGRFDZXW5DOH3SEIO/events.json","paper":"https://pith.science/paper/IDZ7AHMZ"},"agent_actions":{"view_html":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO","download_json":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO.json","view_paper":"https://pith.science/paper/IDZ7AHMZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.07104&json=true","fetch_graph":"https://pith.science/api/pith-number/IDZ7AHMZVCGRFDZXW5DOH3SEIO/graph.json","fetch_events":"https://pith.science/api/pith-number/IDZ7AHMZVCGRFDZXW5DOH3SEIO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO/action/storage_attestation","attest_author":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO/action/author_attestation","sign_citation":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO/action/citation_signature","submit_replication":"https://pith.science/pith/IDZ7AHMZVCGRFDZXW5DOH3SEIO/action/replication_record"}},"created_at":"2026-07-05T12:08:03.921359+00:00","updated_at":"2026-07-05T12:08:03.921359+00:00"}