{"id":"https://openalex.org/W4412888589","doi":"https://doi.org/10.18653/v1/2025.findings-acl.401","title":"Chain-of-Scrutiny: Detecting Backdoor Attacks for Large Language Models","display_name":"Chain-of-Scrutiny: Detecting Backdoor Attacks for Large Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888589","doi":"https://doi.org/10.18653/v1/2025.findings-acl.401"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.401","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.401","pdf_url":"https://aclanthology.org/2025.findings-acl.401.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 2025","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.401.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5111256771","display_name":"Xi Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xi Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Ruofan Mao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruofan Mao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038357504","display_name":"Yusen Zhang","orcid":"https://orcid.org/0000-0003-3842-1153"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yusen Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109731561","display_name":"Renze Lou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Renze Lou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100834917","display_name":"Chen Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen Wu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5108147441","display_name":"Jiaqi Wang","orcid":"https://orcid.org/0009-0007-7824-3394"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiaqi Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":10.5526,"has_fulltext":true,"cited_by_count":6,"citation_normalized_percentile":{"value":0.97990815,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"7705","last_page":"7727"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9846000075340271,"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":0.9846000075340271,"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":0.9386000037193298,"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/backdoor","display_name":"Backdoor","score":0.9780949354171753},{"id":"https://openalex.org/keywords/scrutiny","display_name":"Scrutiny","score":0.8341939449310303},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.649357795715332},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.5642199516296387},{"id":"https://openalex.org/keywords/law","display_name":"Law","score":0.166835218667984},{"id":"https://openalex.org/keywords/political-science","display_name":"Political science","score":0.14450204372406006}],"concepts":[{"id":"https://openalex.org/C2781045450","wikidata":"https://www.wikidata.org/wiki/Q254569","display_name":"Backdoor","level":2,"score":0.9780949354171753},{"id":"https://openalex.org/C2776050585","wikidata":"https://www.wikidata.org/wiki/Q7439360","display_name":"Scrutiny","level":2,"score":0.8341939449310303},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.649357795715332},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.5642199516296387},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.166835218667984},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.14450204372406006}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.401","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.401","pdf_url":"https://aclanthology.org/2025.findings-acl.401.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 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.401","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.401","pdf_url":"https://aclanthology.org/2025.findings-acl.401.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 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888589.pdf","grobid_xml":"https://content.openalex.org/works/W4412888589.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/W4320031223","https://openalex.org/W4200629851","https://openalex.org/W4281902577","https://openalex.org/W4309417370","https://openalex.org/W4292107232","https://openalex.org/W3009072493","https://openalex.org/W4401407399"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs),":[3],"especially":[4],"those":[5],"accessed":[6],"via":[7],"APIs,":[8],"have":[9],"demonstrated":[10],"impressive":[11],"capabilities":[12],"across":[13],"various":[14,164],"domains.However,":[15],"users":[16,48],"without":[17],"technical":[18],"expertise":[19],"often":[20],"turn":[21],"to":[22,30,37,47,70,93,100,147],"(untrustworthy)":[23],"third-party":[24],"services,":[25],"such":[26],"as":[27],"prompt":[28],"engineering,":[29],"enhance":[31],"their":[32],"LLM":[33,99,128],"experience,":[34],"creating":[35],"vulnerabilities":[36],"adversarial":[38],"threats":[39],"like":[40],"backdoor":[41,95],"attacks.Backdoor-compromised":[42],"LLMs":[43,68],"generate":[44,101],"malicious":[45],"outputs":[46],"when":[49],"inputs":[50],"contain":[51],"specific":[52],"\"triggers\"":[53],"set":[54],"by":[55,141],"attackers.Traditional":[56],"defense":[57,150],"strategies,":[58],"originally":[59],"designed":[60],"for":[61,66,104,110,124,173],"small-scale":[62],"models,":[63],"are":[64],"impractical":[65],"APIaccessible":[67],"due":[69],"limited":[71],"model":[72],"access,":[73],"high":[74],"computational":[75],"costs,":[76],"and":[77,108,135,139,166],"data":[78],"requirements.To":[79],"address":[80],"these":[81],"limitations,":[82],"we":[83],"propose":[84],"Chain-of-Scrutiny":[85],"(CoS)":[86],"which":[87],"leverages":[88],"LLMs'":[89],"unique":[90],"reasoning":[91,102],"abilities":[92],"mitigate":[94],"attacks.It":[96],"guides":[97],"the":[98,113,125,149,156],"steps":[103],"a":[105,119],"given":[106],"input":[107],"scrutinizes":[109],"consistency":[111],"with":[112,136,168],"final":[114],"output":[115],"-any":[116],"inconsistencies":[117],"indicating":[118],"potential":[120],"attack.It":[121],"is":[122],"well-suited":[123],"popular":[126],"API-only":[127],"deployments,":[129],"enabling":[130],"detection":[131],"at":[132],"minimal":[133],"cost":[134],"little":[137],"data.User-friendly":[138],"driven":[140],"natural":[142],"language,":[143],"it":[144],"allows":[145],"non-experts":[146],"perform":[148],"independently":[151],"while":[152],"maintaining":[153],"transparency.We":[154],"validate":[155],"effectiveness":[157],"of":[158],"CoS":[159],"through":[160],"extensive":[161],"experiments":[162],"on":[163],"tasks":[165],"LLMs,":[167],"results":[169],"showing":[170],"greater":[171],"benefits":[172],"more":[174],"powerful":[175],"LLMs.":[176]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":4}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}