{"id":"https://openalex.org/W4416033999","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.989","title":"LlmFixer: Fix the Helpfulness of Defensive Large Language Models","display_name":"LlmFixer: Fix the Helpfulness of Defensive Large Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416033999","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.989"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.989","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.989","pdf_url":"https://aclanthology.org/2025.findings-emnlp.989.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: EMNLP 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-emnlp.989.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113146102","display_name":"Zelong Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zelong Yu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100462591","display_name":"Xiaoming Zhang","orcid":"https://orcid.org/0000-0002-6662-4102"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaoming Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101599254","display_name":"Litian Zhang","orcid":"https://orcid.org/0000-0002-6981-3873"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Litian Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100444868","display_name":"Yuan Yu","orcid":"https://orcid.org/0000-0002-3148-6600"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu Yuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5037831162","display_name":"Chaozhuo Li","orcid":"https://orcid.org/0000-0002-9867-1712"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chaozhuo Li","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":"18233","last_page":"18247"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.241799995303154,"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":0.241799995303154,"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/T10028","display_name":"Topic Modeling","score":0.22169999778270721,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.08269999921321869,"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/helpfulness","display_name":"Helpfulness","score":0.5052000284194946},{"id":"https://openalex.org/keywords/perception","display_name":"Perception","score":0.31119999289512634},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.24770000576972961},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.22859999537467957}],"concepts":[{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.6323999762535095},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.5133000016212463},{"id":"https://openalex.org/C2781265381","wikidata":"https://www.wikidata.org/wiki/Q5710255","display_name":"Helpfulness","level":2,"score":0.5052000284194946},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.31119999289512634},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.2558000087738037},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.250900000333786},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.24770000576972961},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.23180000483989716},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.22859999537467957},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2176000028848648}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.989","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.989","pdf_url":"https://aclanthology.org/2025.findings-emnlp.989.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: EMNLP 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.989","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.989","pdf_url":"https://aclanthology.org/2025.findings-emnlp.989.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: EMNLP 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/W4416033999.pdf","grobid_xml":"https://content.openalex.org/works/W4416033999.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Defense":[0],"strategies":[1,29,145],"of":[2,24,33,61,79,121,149],"large":[3,34,47,84,108,124,136],"language":[4,35,48,85,109,125,137],"models":[5,49,86],"besides":[6],"alignment":[7],"are":[8],"introduced":[9],"to":[10,19,55,87,91,97],"defend":[11],"against":[12],"jailbreak":[13,25,140],"attacks,":[14,141],"and":[15,66,94,142],"they":[16],"have":[17],"managed":[18],"decrease":[20],"the":[21,31,77,119,147],"success":[22],"rate":[23],"attacks.However,":[26],"these":[27],"defense":[28,53,144],"weakened":[30],"helpfulness":[32,130],"models.In":[36],"this":[37],"work,":[38],"we":[39],"propose":[40],"a":[41,67,73,103,122],"universal":[42],"framework":[43],"LlmFixer":[44,127],"acting":[45],"on":[46,134],"equipped":[50],"with":[51],"any":[52],"strategy":[54],"recover":[56],"their":[57],"original":[58],"helpfulness.LlmFixer":[59],"consists":[60],"an":[62],"input":[63,80],"prompt":[64,70],"re-writer":[65,71],"logic":[68,100],"patch.The":[69],"is":[72,102],"pre-model":[74],"for":[75],"clarifying":[76],"intention":[78],"prompts,":[81],"which":[82],"promotes":[83],"be":[88],"more":[89,95],"helpful":[90],"benign":[92],"inputs":[93],"rejective":[96],"malicious":[98],"inputs.The":[99],"patch":[101],"lightweight":[104],"structure":[105],"that":[106],"enhances":[107],"models'":[110],"comprehension":[111],"capacity":[112],"by":[113],"supplementing":[114],"certain":[115],"logical":[116],"relationships.Without":[117],"updating":[118],"parameters":[120],"defensive":[123],"model,":[126],"fixes":[128],"its":[129],"while":[131],"preserving":[132],"safety.Experiments":[133],"three":[135],"models,":[138],"five":[139],"four":[143],"show":[146],"effectiveness":[148],"LlmFixer.":[150]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-11-08T00:00:00"}