{"id":"https://openalex.org/W7166839472","doi":"https://doi.org/10.18653/v1/2026.findings-acl.877","title":"When to Trust Tools? Adaptive Tool Trust Calibration For Tool-Integrated Math Reasoning","display_name":"When to Trust Tools? Adaptive Tool Trust Calibration For Tool-Integrated Math Reasoning","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166839472","doi":"https://doi.org/10.18653/v1/2026.findings-acl.877"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.877","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.877","pdf_url":"https://aclanthology.org/2026.findings-acl.877.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.877.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5037627177","display_name":"Ruotao Xu","orcid":"https://orcid.org/0000-0002-5277-9859"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruotao Xu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139823019","display_name":"Yixin Ji","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yixin Ji","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139832151","display_name":"Yu Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu Luo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139797770","display_name":"Jinpeng Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinpeng Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139819746","display_name":"Dong Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dong Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139844027","display_name":"Peifeng Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peifeng Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139842492","display_name":"Juntao Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Juntao Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139762730","display_name":"Min Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Min Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.82905096,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"17691","last_page":"17706"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11424","display_name":"Security and Verification in Computing","score":0.1589999943971634,"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"}},"topics":[{"id":"https://openalex.org/T11424","display_name":"Security and Verification in Computing","score":0.1589999943971634,"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/T10927","display_name":"Access Control and Trust","score":0.09880000352859497,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.06270000338554382,"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/calibration","display_name":"Calibration","score":0.5587000250816345},{"id":"https://openalex.org/keywords/non-monotonic-logic","display_name":"Non-monotonic logic","score":0.31139999628067017},{"id":"https://openalex.org/keywords/automated-reasoning","display_name":"Automated reasoning","score":0.266400009393692},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.26429998874664307}],"concepts":[{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.5587000250816345},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5037000179290771},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4496999979019165},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3813999891281128},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.31299999356269836},{"id":"https://openalex.org/C159032336","wikidata":"https://www.wikidata.org/wiki/Q2488768","display_name":"Non-monotonic logic","level":2,"score":0.31139999628067017},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.27480000257492065},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.266400009393692},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.26429998874664307},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.23839999735355377}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.877","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.877","pdf_url":"https://aclanthology.org/2026.findings-acl.877.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.877","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.877","pdf_url":"https://aclanthology.org/2026.findings-acl.877.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.47699838876724243}],"awards":[],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320324720","display_name":"Soochow University","ror":"https://ror.org/05t8y2r12"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166839472.pdf","grobid_xml":"https://content.openalex.org/works/W7166839472.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"reasoning":[1,55,81],"models":[2,71,180],"(LRMs)":[3],"have":[4,27,59],"achieved":[5],"strong":[6],"performance":[7,200],"enhancement":[8],"through":[9],"scaling":[10],"test":[11],"time":[12],"computation,":[13],"but":[14,108],"due":[15],"to":[16,93,132,155,158,204],"the":[17,21,54,80,83,87,90,103,112,126,136,153,162,167,193],"inherent":[18],"limitations":[19],"of":[20,82,170,181,202],"underlying":[22],"language":[23],"models,":[24,65],"they":[25],"still":[26,72],"shortcomings":[28],"in":[29,95,115,198],"tasks":[30],"that":[31,47,69,78,125,151,189],"require":[32],"precise":[33],"computation":[34],"and":[35,51,184],"extensive":[36],"knowledge":[37],"reserves.Tool-Integrated":[38],"Reasoning":[39],"(TIR)":[40],"has":[41],"emerged":[42],"as":[43,121],"a":[44,148,199],"promising":[45],"paradigm":[46],"incorporates":[48],"tool":[49,88,104,163],"call":[50],"execution":[52],"within":[53],"trajectory.Although":[56],"recent":[57],"works":[58],"released":[60],"some":[61],"powerful":[62],"open-source":[63,178],"TIR":[64,179],"our":[66],"analysis":[67],"reveals":[68],"these":[70,139],"suffer":[73],"from":[74,176],"critical":[75],"deficiencies.We":[76],"find":[77],"when":[79,131],"model":[84,91,127,154],"conflicts":[85],"with":[86],"results,":[89],"tends":[92],"believe":[94],"its":[96],"own":[97],"reasoning.And":[98],"there":[99],"are":[100,106,109,207],"cases":[101],"where":[102],"results":[105,164,175],"correct":[107],"ignored":[110],"by":[111],"model,":[113],"resulting":[114,197],"incorrect":[116],"answers,":[117],"which":[118],"we":[119],"define":[120],"\"Tool":[122,194],"Ignored\".This":[123],"indicates":[124],"does":[128],"not":[129],"know":[130],"trust":[133,159],"or":[134,160],"ignore":[135,161],"tool.To":[137],"overcome":[138],"limitations,":[140],"We":[141],"introduce":[142],"Adaptive":[143],"Tool":[144],"Trust":[145],"Calibration":[146],"(ATTC),":[147],"novel":[149],"framework":[150],"guides":[152],"adaptively":[156],"choose":[157],"based":[165],"on":[166],"confidence":[168],"score":[169],"generated":[171],"code":[172],"blocks.The":[173],"experimental":[174],"various":[177],"different":[182],"sizes":[183],"across":[185],"multiple":[186],"datasets":[187],"demonstrate":[188],"ATTC":[190],"effectively":[191],"reduces":[192],"Ignored\"":[195],"issue,":[196],"increase":[201],"4.1%":[203],"7.5%.Our":[205],"codes":[206],"available":[208],"at":[209]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}