{"id":"https://openalex.org/W7166804848","doi":"https://doi.org/10.18653/v1/2026.acl-long.1571","title":"ART: Attention Replacement Technique to Improve Factuality in LLMs","display_name":"ART: Attention Replacement Technique to Improve Factuality in LLMs","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166804848","doi":"https://doi.org/10.18653/v1/2026.acl-long.1571"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.acl-long.1571","doi":"https://doi.org/10.18653/v1/2026.acl-long.1571","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.1571","pdf_url":"https://aclanthology.org/2026.acl-long.1571.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":"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"},"volume":null,"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2026.acl-long.1571.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5037234614","display_name":"Ziqin Luo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I24943067","display_name":"Fudan University","ror":"https://ror.org/013q1eq08","country_code":"CN","type":"education","lineage":["https://openalex.org/I24943067"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ziqin Luo","raw_affiliation_strings":["Fudan University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Fudan University","institution_ids":["https://openalex.org/I24943067"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114202283","display_name":"Yihao Quan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yihao Quan","raw_affiliation_strings":["Department of Automation and Intelligent Sensing , Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Automation and Intelligent Sensing , Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139763499","display_name":"Xiaofeng Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183067930","display_name":"Shanghai Jiao Tong University","ror":"https://ror.org/0220qvk04","country_code":"CN","type":"education","lineage":["https://openalex.org/I183067930"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiaofeng Zhang","raw_affiliation_strings":["Department of Automation and Intelligent Sensing , Shanghai Jiao Tong University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Automation and Intelligent Sensing , Shanghai Jiao Tong University","institution_ids":["https://openalex.org/I183067930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139837184","display_name":"Xiaosong Yuan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xiaosong Yuan","raw_affiliation_strings":["Alibaba Cloud Computing"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Cloud Computing","institution_ids":["https://openalex.org/I45928872"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5139834999","display_name":"Chen Shen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I45928872","display_name":"Alibaba Group (China)","ror":"https://ror.org/00k642b80","country_code":"CN","type":"company","lineage":["https://openalex.org/I45928872"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen Shen","raw_affiliation_strings":["Alibaba Cloud Computing"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Alibaba Cloud Computing","institution_ids":["https://openalex.org/I45928872"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":8.9348,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.97253664,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":93,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"34068","last_page":"34078"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9160000085830688,"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/T10028","display_name":"Topic Modeling","score":0.9160000085830688,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.03849999979138374,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.027000000700354576,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/hallucination-mitigation","display_name":"hallucination mitigation","score":0.843999981880188},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9660000205039978},{"id":"https://openalex.org/keywords/attention-patterns","display_name":"attention patterns","score":0.8629999756813049},{"id":"https://openalex.org/keywords/local-attention","display_name":"local attention","score":0.8659999966621399},{"id":"https://openalex.org/keywords/shallow-layers","display_name":"shallow layers","score":0.9509999752044678},{"id":"https://openalex.org/keywords/training-free-methods","display_name":"training-free methods","score":0.7059999704360962},{"id":"https://openalex.org/keywords/factuality","display_name":"factuality","score":0.6470000147819519},{"id":"https://openalex.org/keywords/question-answering","display_name":"question answering","score":0.9340000152587891}],"concepts":[{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.3885999917984009},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.319599986076355},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.2694999873638153},{"id":"https://openalex.org/C119767625","wikidata":"https://www.wikidata.org/wiki/Q618211","display_name":"Optometry","level":1,"score":0.2628999948501587},{"id":"https://openalex.org/C12174686","wikidata":"https://www.wikidata.org/wiki/Q1058438","display_name":"Risk assessment","level":2,"score":0.2587999999523163},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.24500000476837158},{"id":"https://openalex.org/C95457728","wikidata":"https://www.wikidata.org/wiki/Q309","display_name":"History","level":0,"score":0.24420000612735748},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.2418999969959259},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.23800000548362732},{"id":"https://openalex.org/C138816342","wikidata":"https://www.wikidata.org/wiki/Q189603","display_name":"Public health","level":2,"score":0.23520000278949738}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.1571","doi":"https://doi.org/10.18653/v1/2026.acl-long.1571","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.1571","pdf_url":"https://aclanthology.org/2026.acl-long.1571.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":"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.1571","doi":"https://doi.org/10.18653/v1/2026.acl-long.1571","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.1571","pdf_url":"https://aclanthology.org/2026.acl-long.1571.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":"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":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166804848.pdf","grobid_xml":"https://content.openalex.org/works/W7166804848.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4400703223","https://openalex.org/W7213865576","https://openalex.org/W4415249784","https://openalex.org/W4409347972","https://openalex.org/W4414677042","https://openalex.org/W4404569270","https://openalex.org/W7213497804","https://openalex.org/W4416036038","https://openalex.org/W4414591368","https://openalex.org/W4417515922"],"abstract_inverted_index":{"Hallucination":[0],"in":[1,13,129,157],"large":[2],"language":[3],"models":[4,19],"(LLMs)":[5],"continues":[6],"to":[7,33,97,103,141],"be":[8],"a":[9,66,115],"significant":[10,155],"issue,":[11,112],"particularly":[12],"tasks":[14],"like":[15],"question":[16],"answering,":[17],"where":[18,81],"often":[20],"generate":[21],"plausible":[22],"yet":[23],"incorrect":[24],"or":[25,171],"irrelevant":[26],"information.Although":[27],"various":[28],"methods":[29],"have":[30],"been":[31,45],"proposed":[32],"mitigate":[34,110],"hallucinations,":[35,98],"the":[36,52,82,89,100,106,130,139,145],"relationship":[37],"between":[38],"attention":[39,55,61,79,86,93,127,135],"patterns":[40,128],"and":[41,60,68,166],"hallucinations":[42,158],"has":[43],"not":[44],"fully":[46],"explored.In":[47],"this":[48,111],"paper,":[49],"we":[50,113],"analyze":[51],"distribution":[53],"of":[54,63,73],"scores":[56],"across":[57,88,159],"each":[58],"layer":[59],"head":[62],"LLMs,":[64],"revealing":[65],"common":[67],"intriguing":[69],"phenomenon:":[70],"Shallow":[71],"layers":[72,132],"LLMs":[74],"primarily":[75],"rely":[76],"on":[77,105,144],"uniform":[78,92,126],"patterns,":[80],"model":[83,101,140],"distributes":[84],"its":[85,164],"evenly":[87],"entire":[90],"sequence.This":[91],"pattern":[94],"can":[95],"lead":[96],"as":[99],"fails":[102],"focus":[104,142],"most":[107],"relevant":[108,146],"information.To":[109],"propose":[114],"trainingfree":[116],"method":[117],"called":[118],"Attention":[119],"Replacement":[120],"Technique":[121],"(ART),":[122],"which":[123],"replaces":[124],"these":[125],"shallow":[131],"with":[133],"local":[134],"patterns.This":[136],"change":[137],"directs":[138],"more":[143],"contexts,":[147],"thus":[148],"reducing":[149],"hallucinations.Through":[150],"extensive":[151],"experiments,":[152],"ART":[153],"demonstrates":[154],"reductions":[156],"multiple":[160],"LLM":[161],"architectures,":[162],"proving":[163],"effectiveness":[165],"generalizability":[167],"without":[168],"requiring":[169],"fine-tuning":[170],"additional":[172],"training":[173],"data.":[174]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-10-29T09:56:32.521200","created_date":"2026-07-02T00:00:00"}