{"id":"https://openalex.org/W7140114123","doi":"https://doi.org/10.18653/v1/2026.eacl-long.161","title":"Breach in the Shield: Unveiling the Vulnerabilities of Large Language Models","display_name":"Breach in the Shield: Unveiling the Vulnerabilities of Large Language Models","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140114123","doi":"https://doi.org/10.18653/v1/2026.eacl-long.161"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.eacl-long.161","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.161","pdf_url":"https://aclanthology.org/2026.eacl-long.161.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 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.eacl-long.161.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130411633","display_name":"Runpeng Dai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Runpeng Dai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130366854","display_name":"Run Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Run Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130375034","display_name":"Fan Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan Zhou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130367322","display_name":"Hongtu Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hongtu Zhu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"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":0.39749266,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3509","last_page":"3521"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.2085999995470047,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.2085999995470047,"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/T13910","display_name":"Computational and Text Analysis Methods","score":0.05570000037550926,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"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":0.048700001090765,"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/data-breach","display_name":"Data breach","score":0.383899986743927},{"id":"https://openalex.org/keywords/vulnerability","display_name":"Vulnerability (computing)","score":0.3375999927520752},{"id":"https://openalex.org/keywords/government","display_name":"Government (linguistics)","score":0.3034999966621399},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.24889999628067017}],"concepts":[{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.559499979019165},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4975000023841858},{"id":"https://openalex.org/C165609540","wikidata":"https://www.wikidata.org/wiki/Q1172486","display_name":"Data breach","level":2,"score":0.383899986743927},{"id":"https://openalex.org/C108827166","wikidata":"https://www.wikidata.org/wiki/Q175975","display_name":"Internet privacy","level":1,"score":0.34389999508857727},{"id":"https://openalex.org/C95713431","wikidata":"https://www.wikidata.org/wiki/Q631425","display_name":"Vulnerability (computing)","level":2,"score":0.3375999927520752},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.3070000112056732},{"id":"https://openalex.org/C2778137410","wikidata":"https://www.wikidata.org/wiki/Q2732820","display_name":"Government (linguistics)","level":2,"score":0.3034999966621399},{"id":"https://openalex.org/C190253527","wikidata":"https://www.wikidata.org/wiki/Q295354","display_name":"Law and economics","level":1,"score":0.2492000013589859},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.24889999628067017},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.23340000212192535}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.eacl-long.161","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.161","pdf_url":"https://aclanthology.org/2026.eacl-long.161.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 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.eacl-long.161","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.161","pdf_url":"https://aclanthology.org/2026.eacl-long.161.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 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140114123.pdf","grobid_xml":"https://content.openalex.org/works/W7140114123.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2,5],"and":[3,38,70,77,84],"Vision-Language":[4],"have":[6],"achieved":[7],"impressive":[8],"performance":[9],"across":[10,82],"a":[11,55],"wide":[12],"range":[13],"of":[14,32,74,97,114],"tasks,":[15],"yet":[16],"they":[17],"remain":[18],"vulnerable":[19,116],"to":[20,28,48,88,108,122],"carefully":[21],"crafted":[22],"perturbations.In":[23],"this":[24,33,51],"study,":[25],"we":[26,53],"seek":[27],"pinpoint":[29],"the":[30,72,112],"sources":[31],"fragility":[34],"by":[35],"identifying":[36],"parameters":[37,98,117],"input":[39,78,100],"dimensions":[40,101],"(pixels":[41],"or":[42,99],"token":[43],"embeddings)":[44],"that":[45],"are":[46],"susceptible":[47],"such":[49],"perturbations.To":[50],"end,":[52],"propose":[54],"stability":[56],"measure":[57],"called":[58],"FI,":[59],"First":[60],"order":[61],"local":[62],"Influence,":[63],"which":[64],"is":[65],"rooted":[66],"in":[67],"information":[68],"geometry":[69],"quantifies":[71],"sensitivity":[73],"individual":[75],"parameter":[76],"dimensions.Our":[79],"extensive":[80],"analysis":[81],"LLMs":[83],"VLMs":[85],"(from":[86],"1.5B":[87],"13B":[89],"parameters)":[90],"reveals":[91],"that:":[92],"(I)":[93],"A":[94],"small":[95],"subset":[96],"with":[102],"high":[103],"FI":[104],"values":[105],"disproportionately":[106],"contribute":[107],"model":[109,119],"brittleness.(II)":[110],"Mitigating":[111],"influence":[113],"these":[115],"during":[118],"merging":[120],"leads":[121],"improved":[123],"performance.":[124]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-03-24T00:00:00"}