{"id":"https://openalex.org/W7166811476","doi":"https://doi.org/10.18653/v1/2026.acl-long.1473","title":"Do LLMs Know Tool Irrelevance? Demystifying Structural Alignment Bias in Tool Invocations","display_name":"Do LLMs Know Tool Irrelevance? Demystifying Structural Alignment Bias in Tool Invocations","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166811476","doi":"https://doi.org/10.18653/v1/2026.acl-long.1473"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.acl-long.1473","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.1473","pdf_url":"https://aclanthology.org/2026.acl-long.1473.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 64th 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/2026.acl-long.1473.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139783670","display_name":"Yilong Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yilong Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101641900","display_name":"Xixun Lin","orcid":"https://orcid.org/0009-0004-6645-0597"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xixun Lin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139775434","display_name":"Pengfei Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pengfei Cao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139788833","display_name":"Ge Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ge Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139777599","display_name":"Fang Fang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang Fang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139829822","display_name":"Yanan Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yanan Cao","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":0.80149484,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"31934","last_page":"31958"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.21780000627040863,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.21780000627040863,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.14319999516010284,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"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/T10028","display_name":"Topic Modeling","score":0.10509999841451645,"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/key","display_name":"Key (lock)","score":0.28040000796318054},{"id":"https://openalex.org/keywords/government","display_name":"Government (linguistics)","score":0.2705000042915344},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.26190000772476196},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.258899986743927}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4577000141143799},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.40380001068115234},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.3896999955177307},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.3093999922275543},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.2874999940395355},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.28040000796318054},{"id":"https://openalex.org/C56739046","wikidata":"https://www.wikidata.org/wiki/Q192060","display_name":"Knowledge management","level":1,"score":0.2797999978065491},{"id":"https://openalex.org/C2778137410","wikidata":"https://www.wikidata.org/wiki/Q2732820","display_name":"Government (linguistics)","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C39549134","wikidata":"https://www.wikidata.org/wiki/Q133080","display_name":"Public relations","level":1,"score":0.26440000534057617},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.26190000772476196},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.258899986743927},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.1473","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.1473","pdf_url":"https://aclanthology.org/2026.acl-long.1473.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 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.acl-long.1473","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.1473","pdf_url":"https://aclanthology.org/2026.acl-long.1473.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 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Reduced inequalities","score":0.4023086130619049,"id":"https://metadata.un.org/sdg/10"}],"awards":[{"id":"https://openalex.org/G4789092901","display_name":null,"funder_award_id":"62402491","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5077752208","display_name":null,"funder_award_id":"U2336202","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8692561774","display_name":null,"funder_award_id":"2025M771524","funder_id":"https://openalex.org/F4320321543","funder_display_name":"China Postdoctoral Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320321543","display_name":"China Postdoctoral Science Foundation","ror":"https://ror.org/0426zh255"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166811476.pdf","grobid_xml":"https://content.openalex.org/works/W7166811476.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"have":[4],"demonstrated":[5,169],"impressive":[6],"capabilities":[7],"in":[8,27,48,110,117],"utilizing":[9],"external":[10],"tools.In":[11],"practice,":[12],"however,":[13],"LLMs":[14,67],"are":[15,21,181],"often":[16],"exposed":[17],"to":[18,23,34,62,70,80],"tools":[19],"that":[20,93,102,162],"irrelevant":[22],"the":[24,30,64,121],"user's":[25,65],"query,":[26],"which":[28,51,132],"case":[29],"desired":[31],"behavior":[32],"is":[33],"refrain":[35],"from":[36,97],"invocations.In":[37],"this":[38,85,125],"work,":[39],"we":[40,52,87,127,156],"identify":[41],"a":[42,59,90,159],"widespread":[43],"yet":[44,112],"overlooked":[45],"mechanistic":[46],"flaw":[47],"tool":[49,60,81,150],"refusal,":[50],"term":[53],"structural":[54,95,103,141,165],"alignment":[55,96,104,166],"bias:":[56],"Even":[57],"when":[58],"fails":[61],"serve":[63],"goal,":[66],"still":[68],"tend":[69],"invoke":[71],"it":[72],"whenever":[73],"query":[74],"attributes":[75],"can":[76],"be":[77],"validly":[78],"assigned":[79],"parameters.To":[82],"systematically":[83],"study":[84],"bias,":[86,126,167],"introduce":[88,158],"SABEval,":[89],"new":[91],"dataset":[92,180],"decouples":[94],"semantic":[98,138],"relevance.Our":[99],"analysis":[100],"shows":[101],"bias":[105],"induces":[106],"severe":[107],"tool-invocation":[108],"errors":[109],"LLMs,":[111],"remains":[113],"largely":[114],"unaccounted":[115],"for":[116,137],"existing":[118],"evaluations.To":[119],"investigate":[120],"internal":[122],"mechanisms":[123],"underlying":[124],"propose":[128],"Contrastive":[129],"Attention":[130],"Attribution,":[131],"reveals":[133],"two":[134],"competing":[135],"pathways":[136,147],"checking":[139],"and":[140,179],"matching.The":[142],"relative":[143],"strength":[144],"of":[145],"these":[146,154],"drives":[148],"LLMs'":[149],"invocation":[151],"decisions.Based":[152],"on":[153],"findings,":[155],"further":[157],"rebalancing":[160],"strategy":[161],"effectively":[163],"mitigates":[164],"as":[168],"by":[170],"extensive":[171],"experiments,":[172],"without":[173],"degrading":[174],"general":[175],"tool-use":[176],"capabilities.Our":[177],"code":[178],"available":[182],"at":[183],"https://github.":[184],"com/along-l/irrelevant-tool.":[185]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}