{"id":"https://openalex.org/W4412889880","doi":"https://doi.org/10.18653/v1/2025.acl-long.1313","title":"Exposing the Achilles\u2019 Heel: Evaluating LLMs Ability to Handle Mistakes in Mathematical Reasoning","display_name":"Exposing the Achilles\u2019 Heel: Evaluating LLMs Ability to Handle Mistakes in Mathematical Reasoning","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412889880","doi":"https://doi.org/10.18653/v1/2025.acl-long.1313"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.1313","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1313","pdf_url":"https://aclanthology.org/2025.acl-long.1313.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.1313.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101278709","display_name":"Joykirat Singh","orcid":null},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Joykirat Singh","raw_affiliation_strings":["Microsoft Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5099230427","display_name":"Akshay Nambi","orcid":null},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Akshay Nambi","raw_affiliation_strings":["Microsoft Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5099230428","display_name":"Vibhav Vineet","orcid":null},"institutions":[{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Vibhav Vineet","raw_affiliation_strings":["Microsoft Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I4210164937"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I4210164937"],"apc_list":null,"apc_paid":null,"fwci":22.5973,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.98605212,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":94,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"27044","last_page":"27065"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12884","display_name":"Educational Assessment and Pedagogy","score":0.9180999994277954,"subfield":{"id":"https://openalex.org/subfields/3304","display_name":"Education"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T12884","display_name":"Educational Assessment and Pedagogy","score":0.9180999994277954,"subfield":{"id":"https://openalex.org/subfields/3304","display_name":"Education"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10130","display_name":"Mathematics Education and Teaching Techniques","score":0.9174000024795532,"subfield":{"id":"https://openalex.org/subfields/3304","display_name":"Education"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/heel","display_name":"Heel","score":0.6175668835639954},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5083915591239929},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.2338869571685791}],"concepts":[{"id":"https://openalex.org/C2777248721","wikidata":"https://www.wikidata.org/wiki/Q174647","display_name":"Heel","level":2,"score":0.6175668835639954},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5083915591239929},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.2338869571685791},{"id":"https://openalex.org/C66938386","wikidata":"https://www.wikidata.org/wiki/Q633538","display_name":"Structural engineering","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.1313","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1313","pdf_url":"https://aclanthology.org/2025.acl-long.1313.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.1313","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1313","pdf_url":"https://aclanthology.org/2025.acl-long.1313.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412889880.pdf","grobid_xml":"https://content.openalex.org/works/W4412889880.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2423203884","https://openalex.org/W1674176876","https://openalex.org/W2411329941","https://openalex.org/W2141376040","https://openalex.org/W2329465293","https://openalex.org/W2066354772","https://openalex.org/W1967218338"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"significantly":[5],"impacted":[6],"the":[7,34,148,168,171,174,196,213,216,225,244,247,250,278,297,300,310,333,336,339,351],"field":[8],"of":[9,37,78,170,173,176,189,215,246,249,271,299,335,338],"Math":[10],"Word":[11],"Problems":[12],"(MWPs),":[13],"transforming":[14],"how":[15],"these":[16],"problems":[17],"are":[18,219,303],"approached":[19],"and":[20,50,64,72,84,92,99,108,117,134,154,161,184,205,236,289,305],"solved,":[21],"particularly":[22,103],"in":[23,96,104,123,132,143,163],"educational":[24],"contexts.However,":[25],"existing":[26],"evaluations":[27,153],"often":[28],"focus":[29],"on":[30,137],"final":[31,135],"accuracy,":[32],"neglecting":[33],"critical":[35],"aspect":[36],"reasoning":[38,52,66,133,152,266],"capabilities.This":[39],"work":[40],"addresses":[41],"that":[42],"gap":[43],"by":[44,227,312],"evaluating":[45],"LLMs'":[46],"abilities":[47],"to":[48,157],"detect":[49],"correct":[51,63],"mistakes.We":[53],"present":[54],"a":[55,177,190,272],"novel":[56,109],"dataset,":[57],"MWP-MISTAKE,":[58],"containing":[59],"MWPs":[60],"with":[61,114,180],"both":[62],"incorrect":[65],"steps":[67],"generated":[68],"through":[69],"rule-based":[70],"methods":[71],"smaller":[73],"language":[74],"models.Our":[75],"comprehensive":[76],"benchmarking":[77],"state-of-the-art":[79],"models":[80],"such":[81],"as":[82],"GPT-4o":[83,94],"GPT4":[85],"uncovers":[86],"important":[87],"insights":[88],"into":[89],"their":[90],"strengths":[91],"limitations.While":[93],"excels":[95],"mistake":[97,144],"detection":[98],"rectification,":[100],"gaps":[101],"remain,":[102],"handling":[105],"complex":[106,138],"datasets":[107],"problems.Additionally,":[110],"we":[111,222,307],"identify":[112],"concerns":[113],"data":[115],"contamination":[116],"memorization,":[118],"which":[119],"affect":[120],"LLM":[121,159],"reliability":[122],"real-world":[124],"applications.While":[125],"OpenAI'":[126],"O1":[127],"model":[128],"demonstrates":[129],"90%":[130],"accuracy":[131],"answers":[136],"tasks,":[139],"it":[140],"remains":[141],"weak":[142],"detection.Our":[145],"findings":[146],"highlight":[147],"need":[149],"for":[150],"improved":[151],"suggest":[155],"ways":[156],"enhance":[158],"generalization":[160],"robustness":[162],"math":[164],"problemsolving.":[165],"Question:What":[166],"is":[167,193,252,275],"product":[169,245,334],"coordinates":[172,248,337],"midpoint":[175,188,226,251,270,311,340],"line":[178,191,217,273,301],"segment":[179,192,218,274,302],"endpoints":[181,214,298],"at":[182,195,277],"(1,1)":[183,220,304],"(-7,5)?":[185],"Reasoning":[186,260],"Step:The":[187],"located":[194,276],"point":[197,279],"(h,k),":[198,280],"where":[199,281],"h":[200,229,282,314],"=":[201,207,230,234,238,242,254,258,283,291,315,319,321,324,328,330,343,347],"(x1":[202,284],"+":[203,209,232,240,285,293,317,326],"x2)/2":[204],"k":[206,237,290,323],"(y1":[208,292],"y2)/2.In":[210,294],"this":[211,295],"case,":[212,296],"and(-7,5).So,":[221],"can":[223,308],"find":[224,309],"calculating":[228],"(1":[231,239,316,325],"(-7))/2":[233,318],"6":[235,255],"5)/2":[241,327],"3.Therefore,":[243],"hk":[253,342],"*":[256,345],"3":[257,331,346],"18.":[259],"chain":[261,267,354],"Correct":[262,355],"(Yes/No):":[263,356],"No":[264],"Corrected":[265],"or":[268],"NA:The":[269],"x2":[286],")/":[287],"2":[288],"(-7,5).So,":[306],"calculating:":[313],"-6/2":[320],"-3":[322,344],"6/2":[329],"Therefore,":[332],"is:":[341],"-9Final":[348],"answer":[349],"(just":[350],"number):":[352],"-9Reasoning":[353]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}