{"id":"https://openalex.org/W4412888198","doi":"https://doi.org/10.18653/v1/2025.findings-acl.760","title":"Beyond Surface-Level Patterns: An Essence-Driven Defense Framework Against Jailbreak Attacks in LLMs","display_name":"Beyond Surface-Level Patterns: An Essence-Driven Defense Framework Against Jailbreak Attacks in LLMs","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888198","doi":"https://doi.org/10.18653/v1/2025.findings-acl.760"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.760","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.760","pdf_url":"https://aclanthology.org/2025.findings-acl.760.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":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.760.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119181362","display_name":"Shiyu Xiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shiyu Xiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Ansen Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ansen Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yanfei Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yanfei Cao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100346596","display_name":"Fan Yang","orcid":"https://orcid.org/0000-0001-6707-4344"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Ronghao Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ronghao Chen","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":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"14727","last_page":"14742"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10734","display_name":"Information and Cyber Security","score":0.9196000099182129,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10734","display_name":"Information and Cyber Security","score":0.9196000099182129,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/attack-surface","display_name":"Attack surface","score":0.6731893420219421},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.46294817328453064},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.4328475296497345},{"id":"https://openalex.org/keywords/business","display_name":"Business","score":0.3308258056640625}],"concepts":[{"id":"https://openalex.org/C2776576444","wikidata":"https://www.wikidata.org/wiki/Q303569","display_name":"Attack surface","level":2,"score":0.6731893420219421},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.46294817328453064},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4328475296497345},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.3308258056640625}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.760","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.760","pdf_url":"https://aclanthology.org/2025.findings-acl.760.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.760","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.760","pdf_url":"https://aclanthology.org/2025.findings-acl.760.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":[{"display_name":"Peace, Justice and strong institutions","score":0.5400000214576721,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888198.pdf","grobid_xml":"https://content.openalex.org/works/W4412888198.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W4382046111","https://openalex.org/W2536179434","https://openalex.org/W1929087753","https://openalex.org/W3152571439","https://openalex.org/W3083937200","https://openalex.org/W3080294124","https://openalex.org/W3041847967","https://openalex.org/W2976636665"],"abstract_inverted_index":{"Although":[0],"Aligned":[1],"Large":[2],"Language":[3],"Models":[4],"(LLMs)":[5],"are":[6],"trained":[7],"to":[8,15,88],"refuse":[9],"harmful":[10,141],"requests,":[11],"they":[12],"remain":[13],"vulnerable":[14],"jailbreak":[16,132],"attacks.Unfortunately,":[17],"existing":[18,115],"methods":[19,116],"often":[20],"focus":[21],"on":[22],"surface-level":[23],"patterns,":[24],"overlooking":[25],"the":[26,40,45,90,119,136],"deeper":[27],"attack":[28,35,99],"essences.As":[29],"a":[30,63,94],"result,":[31],"defenses":[32],"fail":[33],"when":[34],"prompts":[36],"change,":[37],"even":[38],"though":[39],"underlying":[41],"\"attack":[42,91],"essence\"":[43,92],"remains":[44],"same.To":[46],"address":[47],"this":[48],"issue,":[49],"we":[50],"introduce":[51],"EDDF,":[52],"an":[53,105],"Essence-Driven":[54],"Defense":[55],"Framework":[56],"Against":[57],"Jailbreak":[58],"Attacks":[59],"in":[60,69,104],"LLMs.EDDF":[61],"is":[62,87],"plugand-play":[64],"input-filtering":[65],"method":[66],"and":[67,77,101],"operates":[68],"two":[70],"stages:":[71],"1)":[72],"offline":[73,106],"essence":[74],"database":[75],"construction,":[76],"2)":[78],"online":[79],"adversarial":[80],"query":[81],"detection.The":[82],"key":[83],"idea":[84],"behind":[85],"EDDF":[86,112],"extract":[89],"from":[93],"diverse":[95],"set":[96],"of":[97,135],"known":[98],"instances":[100],"store":[102],"it":[103],"vector":[107],"database.Experimental":[108],"results":[109],"demonstrate":[110],"that":[111],"significantly":[113],"outperforms":[114],"by":[117,123],"reducing":[118],"Attack":[120],"Success":[121],"Rate":[122],"at":[124],"least":[125],"20%,":[126],"underscoring":[127],"its":[128],"superior":[129],"robustness":[130],"against":[131],"attacks.warning:":[133],"Some":[134],"examples":[137],"may":[138],"contain":[139],"potentially":[140],"content!":[142]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}
