{"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/arxiv-2601-23184","title":"ReGuLaR: Variational Latent Reasoning Guided by Rendered Chain-of-Thought","arxiv_id":"2601.23184","date":"2026-01-30","proceeding":null,"authors":["Fanmeng Wang","Haotian Liu","Guojiang Zhao","Hongteng Xu","Zhifeng Gao"],"abstract":"While Chain-of-Thought (CoT) significantly enhances the performance of Large Language Models (LLMs), explicit reasoning chains introduce substantial computational redundancy. Recent latent reasoning methods attempt to mitigate this by compressing reasoning processes into latent space, but often suffer from severe performance degradation due to the lack of appropriate compression guidance. In this study, we propose Rendered CoT-Guided variational Latent Reasoning (ReGuLaR), a simple yet novel latent learning paradigm resolving this issue. Fundamentally, we formulate latent reasoning within the Variational Auto-Encoding (VAE) framework, sampling the current latent reasoning state from the posterior distribution conditioned on previous ones. Specifically, when learning this variational latent reasoning model, we render explicit reasoning chains as images, from which we extract dense visual-semantic representations to regularize the posterior distribution, thereby achieving efficient compression with minimal information loss. Extensive experiments demonstrate that ReGuLaR significantly outperforms existing latent reasoning methods across both computational efficiency and reasoning effectiveness, and even surpasses CoT through multi-modal reasoning, providing a new and insightful solution to latent reasoning. Code: https://github.com/FanmengWang/ReGuLaR.","url_abs":"https://arxiv.org/abs/2601.23184","url_pdf":"https://arxiv.org/pdf/2601.23184","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2601.23184","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2601.23184"}},"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":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/FanmengWang/ReGuLaR","reach":null}],"summary":{"ran":3,"unverified":7},"by_repo_kind":{"found_in_text":{"samples":10,"ran":3,"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":10,"samples":[{"code_sha256_prefix":"efa042d231a33f48","entry":"extract_coordinates_and_label","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"models/modeling_deepseekocr.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/models/modeling_deepseekocr.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"efa042d231a33f48"}},{"code_sha256_prefix":"423c3753d415a8e4","entry":"load_image","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"models/modeling_deepseekocr.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/models/modeling_deepseekocr.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"423c3753d415a8e4"}},{"code_sha256_prefix":"5f6c894866c688de","entry":"re_match","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"models/modeling_deepseekocr.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/models/modeling_deepseekocr.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"5f6c894866c688de"}},{"code_sha256_prefix":"aa233c91849bbb72","entry":"get_precompute_path","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/datasets/qsa.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/datasets/qsa.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"aa233c91849bbb72"}},{"code_sha256_prefix":"d61594eb95b26795","entry":"get_sincos_pe","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/modules/embeddings.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/modules/embeddings.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"d61594eb95b26795"}},{"code_sha256_prefix":"edc8ba96bd8d9220","entry":"group_advantages","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/modules/grpo.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/modules/grpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"edc8ba96bd8d9220"}},{"code_sha256_prefix":"609b422c9cdf8797","entry":"join_experience_batch","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/modules/grpo.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/modules/grpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"609b422c9cdf8797"}},{"code_sha256_prefix":"832fe1e0fa149cb7","entry":"setup_logger","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/utils/log.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/utils/log.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"832fe1e0fa149cb7"}},{"code_sha256_prefix":"745626aeb45c996a","entry":"split_experience_batch","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/modules/grpo.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/modules/grpo.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"745626aeb45c996a"}},{"code_sha256_prefix":"3f573803bf191901","entry":"steps_collate_fn","repo":"FanmengWang/ReGuLaR","repo_kind":"found_in_text","path":"src/datasets/qsa.py","file_url":"https://github.com/FanmengWang/ReGuLaR/blob/HEAD/src/datasets/qsa.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"GPL-3.0","inline_ok":false,"mcp_get_code":{"code_sha256":"3f573803bf191901"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}