{"id":"https://openalex.org/W4412945351","doi":"https://doi.org/10.18653/v1/2025.acl-long.594","title":"Safe: Enhancing Mathematical Reasoning in Large Language Models via Retrospective Step-aware Formal Verification","display_name":"Safe: Enhancing Mathematical Reasoning in Large Language Models via Retrospective Step-aware Formal Verification","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412945351","doi":"https://doi.org/10.18653/v1/2025.acl-long.594"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.594","doi":"https://doi.org/10.18653/v1/2025.acl-long.594","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.594","pdf_url":"https://aclanthology.org/2025.acl-long.594.pdf","source":{"id":"https://openalex.org/S4393915740","display_name":"Proceedings of the conference - Association for Computational Linguistics. Meeting","issn_l":"0736-587X","issn":["0736-587X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["norway-1"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://aclanthology.org/2025.acl-long.594.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5040272323","display_name":"Chengwu Liu","orcid":"https://orcid.org/0000-0001-7383-0951","observed_orcids":["https://orcid.org/0000-0001-7383-0951"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chengwu Liu","raw_affiliation_strings":["School of Computer Science , National Key Laboratory for Multimedia Information Processing ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , National Key Laboratory for Multimedia Information Processing ,","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100337347","display_name":"Ye Yuan","orcid":"https://orcid.org/0000-0003-2988-5001","observed_orcids":["https://orcid.org/0000-0003-2988-5001"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ye Yuan","raw_affiliation_strings":["School of Computer Science , National Key Laboratory for Multimedia Information Processing ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science , National Key Laboratory for Multimedia Information Processing ,","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102209620","display_name":"Yichun Yin","orcid":"https://orcid.org/0000-0001-5567-8861","observed_orcids":["https://orcid.org/0000-0001-5567-8861"]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yichun Yin","raw_affiliation_strings":["PKU-Anker LLM Lab , Peking University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"PKU-Anker LLM Lab , Peking University ,","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112127849","display_name":"Yan Mei Xu","orcid":"https://orcid.org/0000-0002-7608-7013","observed_orcids":["https://orcid.org/0000-0002-7608-7013"]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yan Xu","raw_affiliation_strings":["PKU-Anker LLM Lab , Peking University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"PKU-Anker LLM Lab , Peking University ,","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081050709","display_name":"Xin Xu","orcid":"https://orcid.org/0000-0003-2145-9198","observed_orcids":["https://orcid.org/0000-0003-2145-9198"]},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xin Xu","raw_affiliation_strings":["The Hong Kong University of Science and Technology ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology ,","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069440808","display_name":"Zaoyu Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I14243506","display_name":"Hong Kong Polytechnic University","ror":"https://ror.org/0030zas98","country_code":"HK","type":"education","lineage":["https://openalex.org/I14243506"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Zaoyu Chen","raw_affiliation_strings":["The Hong Kong Polytechnic University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong Polytechnic University ,","institution_ids":["https://openalex.org/I14243506"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5115592503","display_name":"Yasheng Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yasheng Wang","raw_affiliation_strings":["PKU-Anker LLM Lab , Peking University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"PKU-Anker LLM Lab , Peking University ,","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046228314","display_name":"Lifeng Shang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lifeng Shang","raw_affiliation_strings":["PKU-Anker LLM Lab , Peking University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"PKU-Anker LLM Lab , Peking University ,","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Qun Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20231570","display_name":"Peking University","ror":"https://ror.org/02v51f717","country_code":"CN","type":"education","lineage":["https://openalex.org/I20231570"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Qun Liu","raw_affiliation_strings":["PKU-Anker LLM Lab , Peking University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"PKU-Anker LLM Lab , Peking University ,","institution_ids":["https://openalex.org/I20231570"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100447321","display_name":"Ming Zhang","orcid":"https://orcid.org/0000-0003-2713-5004","observed_orcids":["https://orcid.org/0000-0003-2713-5004"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ming Zhang","raw_affiliation_strings":["Huawei Noah's Ark Lab ,","School of Computer Science , National Key Laboratory for Multimedia Information Processing ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah's Ark Lab ,","institution_ids":["https://openalex.org/I2250955327"]},{"raw_affiliation_string":"School of Computer Science , National Key Laboratory for Multimedia Information Processing ,","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":5.359,"has_fulltext":true,"cited_by_count":3,"citation_normalized_percentile":{"value":0.95979899,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":90,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"12171","last_page":"12186"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T13523","display_name":"Mathematics, Computing, and Information Processing","score":0.3725999891757965,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T13523","display_name":"Mathematics, Computing, and Information Processing","score":0.3725999891757965,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.21639999747276306,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10126","display_name":"Logic, programming, and type systems","score":0.08640000224113464,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8309999704360962},{"id":"https://openalex.org/keywords/chain-of-thought-prompting","display_name":"chain-of-thought prompting","score":0.8709999918937683},{"id":"https://openalex.org/keywords/hallucination-mitigation","display_name":"hallucination mitigation","score":0.8080000281333923},{"id":"https://openalex.org/keywords/formal-verification","display_name":"formal verification","score":0.9330000281333923},{"id":"https://openalex.org/keywords/lean-4","display_name":"Lean 4","score":0.9330000281333923},{"id":"https://openalex.org/keywords/mathematical-reasoning","display_name":"mathematical reasoning","score":0.6399999856948853},{"id":"https://openalex.org/keywords/formal-proof","display_name":"formal proof","score":0.753000020980835},{"id":"https://openalex.org/keywords/theorem-proving","display_name":"theorem proving","score":0.6119999885559082},{"id":"https://openalex.org/keywords/process-reward-models","display_name":"process reward models","score":0.6990000009536743},{"id":"https://openalex.org/keywords/self-consistency","display_name":"self-consistency","score":0.9610000252723694}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7441960573196411},{"id":"https://openalex.org/C111498074","wikidata":"https://www.wikidata.org/wiki/Q173326","display_name":"Formal verification","level":2,"score":0.6120060682296753},{"id":"https://openalex.org/C75606506","wikidata":"https://www.wikidata.org/wiki/Q1049183","display_name":"Formal methods","level":2,"score":0.552117109298706},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.5124857425689697},{"id":"https://openalex.org/C146072743","wikidata":"https://www.wikidata.org/wiki/Q192161","display_name":"Formal language","level":2,"score":0.4596083164215088}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.594","doi":"https://doi.org/10.18653/v1/2025.acl-long.594","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.594","pdf_url":"https://aclanthology.org/2025.acl-long.594.pdf","source":{"id":"https://openalex.org/S4393915740","display_name":"Proceedings of the conference - Association for Computational Linguistics. Meeting","issn_l":"0736-587X","issn":["0736-587X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["norway-1"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2506.04592","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2506.04592","pdf_url":"https://arxiv.org/pdf/2506.04592","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-170435","doi":null,"is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-170435","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Conference paper"},{"id":"doi:10.48550/arxiv.2506.04592","doi":"https://doi.org/10.48550/arxiv.2506.04592","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2506.04592","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.594","doi":"https://doi.org/10.18653/v1/2025.acl-long.594","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.594","pdf_url":"https://aclanthology.org/2025.acl-long.594.pdf","source":{"id":"https://openalex.org/S4393915740","display_name":"Proceedings of the conference - Association for Computational Linguistics. Meeting","issn_l":"0736-587X","issn":["0736-587X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["norway-1"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G1504557152","display_name":null,"funder_award_id":"2023YFC3341203","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G2997987028","display_name":null,"funder_award_id":"62276002","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412945351.pdf","grobid_xml":"https://content.openalex.org/works/W4412945351.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4416035882","https://openalex.org/W4410601367","https://openalex.org/W7106708571","https://openalex.org/W4407185029","https://openalex.org/W4417536095","https://openalex.org/W4417493510","https://openalex.org/W7164182582","https://openalex.org/W4415331377","https://openalex.org/W7155081826","https://openalex.org/W4406890586"],"abstract_inverted_index":{"Chain-of-Thought":[0],"(CoT)":[1],"prompting":[2],"has":[3],"become":[4],"the":[5,63,157,165,185,194],"de":[6],"facto":[7],"method":[8],"to":[9,18,27,75,97,115,168,197],"elicit":[10],"reasoning":[11,109],"capabilities":[12],"from":[13,62],"large":[14],"language":[15,104,124,172,178],"models":[16,35,125],"(LLMs).However,":[17],"mitigate":[19],"hallucinations":[20],"in":[21,101,193],"CoT":[22],"that":[23,65],"are":[24],"notoriously":[25],"difficult":[26],"detect,":[28],"current":[29],"methods":[30],"such":[31],"as":[32,40,144],"process":[33],"reward":[34],"(PRMs)":[36],"or":[37],"self-consistency":[38],"operate":[39],"opaque":[41],"boxes":[42],"and":[43,111,126,138],"do":[44],"not":[45],"provide":[46,76,112,198],"checkable":[47],"evidence":[48],"for":[49,69,147,175,202],"their":[50,54],"judgments,":[51],"possibly":[52],"limiting":[53],"effectiveness.To":[55],"address":[56],"this":[57],"issue,":[58],"we":[59,95],"draw":[60],"inspiration":[61],"idea":[64],"\"the":[66],"gold":[67],"standard":[68],"supporting":[70],"a":[71,77,84,131,145,199],"mathematical":[72,99,103,128,171,189],"claim":[73],"is":[74],"proof\"":[78],"(Avigad":[79],"et":[80],"al.,":[81],"2021).We":[82],"propose":[83,142],"retrospective,":[85],"step-aware":[86],"formal":[87,102,113,155,170,188],"verification":[88],"framework":[89,120],"Safe.Rather":[90],"than":[91],"assigning":[92],"arbitrary":[93],"scores,":[94],"strive":[96],"articulate":[98],"claims":[100],"Lean":[105,173],"4":[106,174],"at":[107],"each":[108],"step":[110,148],"proofs":[114],"identify":[116],"hallucinations.We":[117],"evaluate":[118],"our":[119,160,162],"Safe":[121],"across":[122],"multiple":[123],"various":[127],"datasets,":[129],"demonstrating":[130],"significant":[132],"performance":[133],"improvement":[134],"while":[135],"offering":[136],"interpretable":[137],"verifiable":[139],"evidence.We":[140],"also":[141],"FormalStep":[143],"benchmark":[146],"correctness":[149],"theorem":[150],"proving":[151],"with":[152,184],"30,":[153],"809":[154],"statements.To":[156],"best":[158],"of":[159],"knowledge,":[161],"work":[163],"represents":[164],"first":[166,195],"endeavor":[167],"utilize":[169],"verifying":[176],"natural":[177],"content":[179],"generated":[180],"by":[181],"LLMs,":[182],"aligning":[183],"reason":[186],"why":[187],"languages":[190],"were":[191],"created":[192],"place:":[196],"robust":[200],"foundation":[201],"hallucination-prone":[203],"human-written":[204],"proofs.":[205]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":1}],"updated_date":"2026-10-29T09:56:32.521200","created_date":"2025-10-10T00:00:00"}
