{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7WF6YZUQDW43ACUNRBUT3CMPTJ","short_pith_number":"pith:7WF6YZUQ","schema_version":"1.0","canonical_sha256":"fd8bec66901db9b00a8d88693d898f9a7b6b076ea3d3ea4da10b0ade68cfa2fd","source":{"kind":"arxiv","id":"2505.17155","version":2},"attestation_state":"computed","paper":{"title":"TrimR: Verifier-based Training-Free Thinking Compression for Efficient Test-Time Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hui-Ling Zhen, Mingxuan Yuan, Weizhe Lin, Wulong Liu, Xianzhi Yu, Xiaojin Fu, Xiaosong Li, Xing Li, Yaoyuan Wang, Zhiyuan Yang","submitted_at":"2025-05-22T12:23:30Z","abstract_excerpt":"Large Reasoning Models (LRMs) demonstrate exceptional capability in tackling complex mathematical, logical, and coding tasks by leveraging extended Chain-of-Thought (CoT) reasoning. Test-time scaling methods, such as prolonging CoT with explicit token-level exploration, can push LRMs' accuracy boundaries, but they incur significant decoding overhead. A key inefficiency source is LRMs often generate redundant thinking CoTs, which demonstrate clear structured overthinking and underthinking patterns. Inspired by human cognitive reasoning processes and numerical optimization theories, we propose T"},"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":"2505.17155","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-22T12:23:30Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"1ffa30e56bc3b36ea6ef5925300a76821f13808e8f347aab82702b018806889d","abstract_canon_sha256":"424486a515ccc43f0cb18c27c56eaf42a66226f3bce4af718ea0d2c3c2aac8e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:26.303950Z","signature_b64":"6a20yupChemS654QS15Mll4hsPHfGVKiqhpFKQnIYeKffEyOkfHr8KMEHxoEs+qA27fvbCaaTX340qFQw39/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fd8bec66901db9b00a8d88693d898f9a7b6b076ea3d3ea4da10b0ade68cfa2fd","last_reissued_at":"2026-07-05T11:13:26.303528Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:26.303528Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TrimR: Verifier-based Training-Free Thinking Compression for Efficient Test-Time Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hui-Ling Zhen, Mingxuan Yuan, Weizhe Lin, Wulong Liu, Xianzhi Yu, Xiaojin Fu, Xiaosong Li, Xing Li, Yaoyuan Wang, Zhiyuan Yang","submitted_at":"2025-05-22T12:23:30Z","abstract_excerpt":"Large Reasoning Models (LRMs) demonstrate exceptional capability in tackling complex mathematical, logical, and coding tasks by leveraging extended Chain-of-Thought (CoT) reasoning. Test-time scaling methods, such as prolonging CoT with explicit token-level exploration, can push LRMs' accuracy boundaries, but they incur significant decoding overhead. A key inefficiency source is LRMs often generate redundant thinking CoTs, which demonstrate clear structured overthinking and underthinking patterns. Inspired by human cognitive reasoning processes and numerical optimization theories, we propose T"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.17155","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/2505.17155/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":"2505.17155","created_at":"2026-07-05T11:13:26.303584+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.17155v2","created_at":"2026-07-05T11:13:26.303584+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.17155","created_at":"2026-07-05T11:13:26.303584+00:00"},{"alias_kind":"pith_short_12","alias_value":"7WF6YZUQDW43","created_at":"2026-07-05T11:13:26.303584+00:00"},{"alias_kind":"pith_short_16","alias_value":"7WF6YZUQDW43ACUN","created_at":"2026-07-05T11:13:26.303584+00:00"},{"alias_kind":"pith_short_8","alias_value":"7WF6YZUQ","created_at":"2026-07-05T11:13:26.303584+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07108","citing_title":"DyCon: Dynamic Reasoning Control via Evolving Difficulty Modeling","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2603.05433","citing_title":"CRISP: Compressed Reasoning via Iterative Self-Policy Distillation","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10195","citing_title":"Breaking the Reward Barrier: Accelerating Tree-of-Thought Reasoning via Speculative Exploration","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2503.16419","citing_title":"Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models","ref_index":105,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10195","citing_title":"Breaking the Reward Barrier: Accelerating Tree-of-Thought Reasoning via Speculative Exploration","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06165","citing_title":"Post Reasoning: Improving the Performance of Non-Thinking Models at No Cost","ref_index":248,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07922","citing_title":"SAT: Balancing Reasoning Accuracy and Efficiency with Stepwise Adaptive Thinking","ref_index":18,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ","json":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ.json","graph_json":"https://pith.science/api/pith-number/7WF6YZUQDW43ACUNRBUT3CMPTJ/graph.json","events_json":"https://pith.science/api/pith-number/7WF6YZUQDW43ACUNRBUT3CMPTJ/events.json","paper":"https://pith.science/paper/7WF6YZUQ"},"agent_actions":{"view_html":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ","download_json":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ.json","view_paper":"https://pith.science/paper/7WF6YZUQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.17155&json=true","fetch_graph":"https://pith.science/api/pith-number/7WF6YZUQDW43ACUNRBUT3CMPTJ/graph.json","fetch_events":"https://pith.science/api/pith-number/7WF6YZUQDW43ACUNRBUT3CMPTJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ/action/storage_attestation","attest_author":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ/action/author_attestation","sign_citation":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ/action/citation_signature","submit_replication":"https://pith.science/pith/7WF6YZUQDW43ACUNRBUT3CMPTJ/action/replication_record"}},"created_at":"2026-07-05T11:13:26.303584+00:00","updated_at":"2026-07-05T11:13:26.303584+00:00"}