{"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/cumulative-reasoning-with-large-language","title":"Cumulative Reasoning with Large Language Models","arxiv_id":"2308.04371","date":"2023-08-08","proceeding":null,"authors":["Yifan Zhang","Jingqin Yang","Yang Yuan","Andrew Chi-Chih Yao"],"abstract":"Despite the recent advancements in language models (LMs), their ability to solve complex problems remains limited. This paper introduces Cumulative Reasoning (CR), a novel approach that utilizes LMs cumulatively and iteratively, mirroring human thought processes for problem-solving. CR decomposes tasks into smaller, manageable components and leverages previous propositions for effective composition, significantly enhancing problem-solving capabilities. We demonstrate CR's superiority through several complex reasoning tasks: it outperforms existing methods in logical inference tasks with up to a 9.3% improvement, achieving 98.04% accuracy on the curated FOLIO wiki dataset. In the Game of 24, it achieves 98% accuracy, marking a 24% improvement over the prior state-of-the-art. Additionally, CR sets new state-of-the-art on the MATH dataset, achieving a 4.2% increase from previous methods and a 43% relative improvement in the most challenging problems. By extending CR to incorporate a code environment without external aids like retrieval or web browsing, we further harness the computational and logical reasoning capabilities of LMs, achieving a remarkable 72.2% accuracy on the MATH dataset and outperforming the PAL/PoT method by 38.8%. Our work not only sets new state-of-the-art but also paves the way toward more sophisticated AI reasoning methods. The code is available at https://github.com/iiis-ai/cumulative-reasoning.","url_abs":"https://arxiv.org/abs/2308.04371v6","url_pdf":"https://arxiv.org/pdf/2308.04371v6.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":"cumulative-reasoning-with-large-language","repo_url":"https://github.com/iiis-ai/cumulative-reasoning","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decision-making","task_name":"Decision Making"},{"task_slug":"logical-reasoning","task_name":"Logical Reasoning"},{"task_slug":"math","task_name":"Math"},{"task_slug":"math-word-problem-solving","task_name":"Math Word Problem Solving"}],"methods":[{"method_slug":"fail","method_name":"fail"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"CR (GPT-4-turbo model, w/ code)","rank_in_archive_order":14,"of":135,"metrics":{"Accuracy":"72.2"},"uses_additional_data":false},{"leaderboard":"/sota/math-word-problem-solving-on-math","task":"Math Word Problem Solving","dataset":"MATH","model":"CR (GPT-4 model, w/o code)","rank_in_archive_order":30,"of":135,"metrics":{"Accuracy":"58.0"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2308.04371","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2308.04371"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+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/iiis-ai/cumulative-reasoning","reach":{"status":"ok"}}],"summary":{"ran_draft_wrong":2,"ran":3,"unverified":1},"by_repo_kind":{"official":{"samples":6,"ran":5,"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":6,"samples":[{"code_sha256_prefix":"bc1a96ec28a11a5f","entry":"encode_with_messages_format","repo":"iiis-ai/cumulative-reasoning","repo_kind":"official","path":"CR-Agent/train/finetune.py","file_url":"https://github.com/iiis-ai/cumulative-reasoning/blob/HEAD/CR-Agent/train/finetune.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"bc1a96ec28a11a5f"}},{"code_sha256_prefix":"1906977026be9a86","entry":"encode_with_prompt_completion_format","repo":"iiis-ai/cumulative-reasoning","repo_kind":"official","path":"CR-Agent/train/finetune.py","file_url":"https://github.com/iiis-ai/cumulative-reasoning/blob/HEAD/CR-Agent/train/finetune.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"1906977026be9a86"}},{"code_sha256_prefix":"da8f1e8f740ab7e8","entry":"get_time_str","repo":"iiis-ai/cumulative-reasoning","repo_kind":"official","path":"MATH/math-complex-cot.py","file_url":"https://github.com/iiis-ai/cumulative-reasoning/blob/HEAD/MATH/math-complex-cot.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"da8f1e8f740ab7e8"}},{"code_sha256_prefix":"8e901d4768f46792","entry":"is_sure","repo":"iiis-ai/cumulative-reasoning","repo_kind":"official","path":"Game24/game24-openai.py","file_url":"https://github.com/iiis-ai/cumulative-reasoning/blob/HEAD/Game24/game24-openai.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8e901d4768f46792"}},{"code_sha256_prefix":"ae7597e49bb0a65a","entry":"run","repo":"iiis-ai/cumulative-reasoning","repo_kind":"official","path":"Game24/game24-openai.py","file_url":"https://github.com/iiis-ai/cumulative-reasoning/blob/HEAD/Game24/game24-openai.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ae7597e49bb0a65a"}},{"code_sha256_prefix":"bcfffe8211f7aac9","entry":"try_wrapper","repo":"iiis-ai/cumulative-reasoning","repo_kind":"official","path":"MATH/math-complex-cot.py","file_url":"https://github.com/iiis-ai/cumulative-reasoning/blob/HEAD/MATH/math-complex-cot.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"bcfffe8211f7aac9"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}