{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/recut-balancing-reasoning-length-and-accuracy","title":"ReCUT: Balancing Reasoning Length and Accuracy in LLMs via Stepwise Trails and Preference Optimization","arxiv_id":"2506.10822","date":"2025-06-12","proceeding":null,"authors":["Zhensheng Jin","Xinze Li","Yifan Ji","Chunyi Peng","Zhenghao Liu","Qi Shi","Yukun Yan","Shuo Wang","Furong Peng","Ge Yu"],"abstract":"Recent advances in Chain-of-Thought (CoT) prompting have substantially improved the reasoning capabilities of Large Language Models (LLMs). However, these methods often suffer from overthinking, leading to unnecessarily lengthy or redundant reasoning traces. Existing approaches attempt to mitigate this issue through curating multiple reasoning chains for training LLMs, but their effectiveness is often constrained by the quality of the generated data and prone to overfitting. To address the challenge, we propose Reasoning Compression ThroUgh Stepwise Trials (ReCUT), a novel method aimed at balancing the accuracy and length of reasoning trajectory. Specifically, ReCUT employs a stepwise exploration mechanism and a long-short switched sampling strategy, enabling LLMs to incrementally generate diverse reasoning paths. These paths are evaluated and used to construct preference pairs to train two specialized models (Gemini LLMs)-one optimized for reasoning accuracy, the other for shorter reasoning. A final integrated model is obtained by interpolating the parameters of these two models. Experimental results across multiple math reasoning datasets and backbone models demonstrate that ReCUT significantly reduces reasoning lengths by approximately 30-50%, while maintaining or improving reasoning accuracy compared to various baselines. All codes and data will be released via https://github.com/NEUIR/ReCUT.","url_abs":"https://arxiv.org/abs/2506.10822v1","url_pdf":"https://arxiv.org/pdf/2506.10822v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"recut-balancing-reasoning-length-and-accuracy","repo_url":"https://github.com/neuir/recut","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"math","task_name":"Math"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2506.10822","atlas_url":"https://app.syntology.ai/?focus=2506.10822","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2506.10822"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/neuir/recut","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":8},"by_repo_kind":{"official":{"samples":8,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"5394b27a305c456c","entry":"analyze_test_results","repo":"neuir/recut","repo_kind":"official","path":"evaluate/evaluate_length.py","file_url":"https://github.com/neuir/recut/blob/HEAD/evaluate/evaluate_length.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"5394b27a305c456c"}},{"code_sha256_prefix":"8bc54889033168e8","entry":"extract_final_answer","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_reward_generate.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_reward_generate.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8bc54889033168e8"}},{"code_sha256_prefix":"bbae4c79ee31ac2a","entry":"find_generation_by_length","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_dpo.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_dpo.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bbae4c79ee31ac2a"}},{"code_sha256_prefix":"53ad32b801cb7bea","entry":"generate_output_paths","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_split_optimal.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_split_optimal.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"53ad32b801cb7bea"}},{"code_sha256_prefix":"c426029e35919853","entry":"merge_text","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_dpo.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_dpo.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c426029e35919853"}},{"code_sha256_prefix":"5565601a35e6f32a","entry":"process_data","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_dpo.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_dpo.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"5565601a35e6f32a"}},{"code_sha256_prefix":"b509dd8e750ca813","entry":"process_math_problems","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_sft.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_sft.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b509dd8e750ca813"}},{"code_sha256_prefix":"1bb412e3dea61577","entry":"split_by_correctness","repo":"neuir/recut","repo_kind":"official","path":"src/long2short_split_optimal.py","file_url":"https://github.com/neuir/recut/blob/HEAD/src/long2short_split_optimal.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1bb412e3dea61577"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}