{"id":"https://openalex.org/W4404783643","doi":"https://doi.org/10.18653/v1/2024.emnlp-main.164","title":"Alignment-Enhanced Decoding: Defending Jailbreaks via Token-Level Adaptive Refining of Probability Distributions","display_name":"Alignment-Enhanced Decoding: Defending Jailbreaks via Token-Level Adaptive Refining of Probability Distributions","publication_year":2024,"publication_date":"2024-01-01","ids":{"openalex":"https://openalex.org/W4404783643","doi":"https://doi.org/10.18653/v1/2024.emnlp-main.164"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2024.emnlp-main.164","doi":"https://doi.org/10.18653/v1/2024.emnlp-main.164","is_oa":true,"landing_page_url":"http://dx.doi.org/10.18653/v1/2024.emnlp-main.164","pdf_url":"https://aclanthology.org/2024.emnlp-main.164.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"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 2024 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2024.emnlp-main.164.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100414577","display_name":"Quan Liu","orcid":"https://orcid.org/0000-0002-9036-0290","observed_orcids":["https://orcid.org/0000-0002-9036-0290"]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Quan Liu","raw_affiliation_strings":["Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113818419","display_name":"Zhenhong Zhou","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhenhong Zhou","raw_affiliation_strings":["Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113269092","display_name":"Longzhu He","orcid":"https://orcid.org/0009-0001-7842-6605","observed_orcids":["https://orcid.org/0009-0001-7842-6605"]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Longzhu He","raw_affiliation_strings":["Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5115883506","display_name":"Yi Liu","orcid":"https://orcid.org/0000-0002-4978-127X","observed_orcids":["https://orcid.org/0000-0002-4978-127X","https://orcid.org/0009-0007-8292-3296"]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yi Liu","raw_affiliation_strings":["Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Wei Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Zhang","raw_affiliation_strings":["Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,","institution_ids":["https://openalex.org/I139759216"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5036865453","display_name":"Sen Su","orcid":"https://orcid.org/0000-0003-4266-7527","observed_orcids":["https://orcid.org/0000-0003-4266-7527","https://orcid.org/0009-0006-4228-8809"]},"institutions":[{"id":"https://openalex.org/I139759216","display_name":"Beijing University of Posts and Telecommunications","ror":"https://ror.org/04w9fbh59","country_code":"CN","type":"education","lineage":["https://openalex.org/I139759216"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sen Su","raw_affiliation_strings":["Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beijing University of Posts and Telecommunications {liuquan , zhouzhenhong ,","institution_ids":["https://openalex.org/I139759216"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I139759216"],"apc_list":null,"apc_paid":null,"fwci":0.4478,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.62158111,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":97},"biblio":{"volume":null,"issue":null,"first_page":"2802","last_page":"2816"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9994999766349792,"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":0.9994999766349792,"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":9.999999747378752e-05,"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/T11424","display_name":"Security and Verification in Computing","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/jailbreak-attacks","display_name":"jailbreak attacks","score":0.9440000057220459},{"id":"https://openalex.org/keywords/large-language-model-safety","display_name":"large language model safety","score":0.7720000147819519},{"id":"https://openalex.org/keywords/alignment-failure","display_name":"alignment failure","score":0.7429999709129333},{"id":"https://openalex.org/keywords/adaptive-decoding","display_name":"adaptive decoding","score":0.921999990940094},{"id":"https://openalex.org/keywords/competitive-index","display_name":"competitive index","score":0.7260000109672546},{"id":"https://openalex.org/keywords/harmful-content-generation","display_name":"harmful content generation","score":0.6370000243186951}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8616959452629089},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.7824772000312805},{"id":"https://openalex.org/C60044698","wikidata":"https://www.wikidata.org/wiki/Q1283324","display_name":"Refining (metallurgy)","level":2,"score":0.6744853258132935},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6326078176498413},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.22269073128700256},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2162759006023407},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0688830316066742},{"id":"https://openalex.org/C147789679","wikidata":"https://www.wikidata.org/wiki/Q11372","display_name":"Physical chemistry","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/v1/2024.emnlp-main.164","doi":"https://doi.org/10.18653/v1/2024.emnlp-main.164","is_oa":true,"landing_page_url":"http://dx.doi.org/10.18653/v1/2024.emnlp-main.164","pdf_url":"https://aclanthology.org/2024.emnlp-main.164.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"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 2024 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2408.07663","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2408.07663","pdf_url":"https://arxiv.org/pdf/2408.07663","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2408.07663","doi":"https://doi.org/10.48550/arxiv.2408.07663","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2408.07663","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/v1/2024.emnlp-main.164","doi":"https://doi.org/10.18653/v1/2024.emnlp-main.164","is_oa":true,"landing_page_url":"http://dx.doi.org/10.18653/v1/2024.emnlp-main.164","pdf_url":"https://aclanthology.org/2024.emnlp-main.164.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"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 2024 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G1755690126","display_name":"\u9690\u79c1\u4fdd\u62a4\u7684\u6570\u636e\u96c6\u6210\u4e0e\u5171\u4eab\u5173\u952e\u6280\u672f\u7814\u7a76","funder_award_id":"62072052","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G2993257473","display_name":null,"funder_award_id":"61921003","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G4318698815","display_name":null,"funder_award_id":"61921003","funder_id":"https://openalex.org/F4320327720","funder_display_name":"Foundation for Innovative Research Groups of the National Natural Science Foundation of China"},{"id":"https://openalex.org/G7838274900","display_name":null,"funder_award_id":"62072052","funder_id":"https://openalex.org/F4320327720","funder_display_name":"Foundation for Innovative Research Groups of the National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320327720","display_name":"Foundation for Innovative Research Groups of the National Natural Science Foundation of China","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4404783643.pdf","grobid_xml":"https://content.openalex.org/works/W4404783643.grobid-xml"},"referenced_works_count":29,"referenced_works":["https://openalex.org/W2896457183","https://openalex.org/W3047185145","https://openalex.org/W3083410900","https://openalex.org/W3195577433","https://openalex.org/W3210904070","https://openalex.org/W4223908421","https://openalex.org/W4226146865","https://openalex.org/W4298181573","https://openalex.org/W4309395891","https://openalex.org/W4366850566","https://openalex.org/W4378506863","https://openalex.org/W4379468930","https://openalex.org/W4384918448","https://openalex.org/W4385374425","https://openalex.org/W4386270999","https://openalex.org/W4386436506","https://openalex.org/W4386555960","https://openalex.org/W4387436631","https://openalex.org/W4387559087","https://openalex.org/W4387595881","https://openalex.org/W4390041933","https://openalex.org/W4392822465","https://openalex.org/W4396945171","https://openalex.org/W4402671718","https://openalex.org/W4404658388","https://openalex.org/W4404781557","https://openalex.org/W4410609100","https://openalex.org/W4415798066","https://openalex.org/W7127179963"],"related_works":["https://openalex.org/W4402427566","https://openalex.org/W4400103834","https://openalex.org/W7124456009","https://openalex.org/W7169804890","https://openalex.org/W4411120028","https://openalex.org/W4391871555","https://openalex.org/W4404782026","https://openalex.org/W4407425033","https://openalex.org/W7208759636","https://openalex.org/W7155087623"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,107],"are":[3],"susceptible":[4],"to":[5,51,64,73,89],"jailbreak":[6,57],"attacks,":[7],"which":[8],"can":[9],"result":[10],"in":[11],"the":[12,31,53,61,86,113,116],"generation":[13],"of":[14,34,56,118],"harmful":[15],"content.While":[16],"prior":[17],"defenses":[18],"mitigate":[19],"these":[20],"risks":[21],"by":[22],"perturbing":[23],"or":[24],"inspecting":[25],"inputs,":[26],"they":[27],"ignore":[28],"competing":[29],"objectives,":[30],"underlying":[32],"cause":[33],"alignment":[35,66,99],"failures.In":[36],"this":[37],"paper,":[38],"we":[39],"propose":[40],"Alignment-Enhanced":[41],"Decoding":[42],"(AED),":[43],"a":[44],"novel":[45],"defense":[46],"that":[47],"employs":[48],"adaptive":[49],"decoding":[50],"address":[52],"root":[54],"causes":[55],"issues.We":[58],"first":[59],"define":[60],"Competitive":[62,80],"Index":[63,81],"quantify":[65],"failures":[67],"and":[68,82,92,108],"utilize":[69],"feedback":[70],"from":[71],"self-evaluation":[72],"compute":[74],"postalignment":[75],"logits.Then,":[76],"AED":[77],"adaptively":[78],"combines":[79],"post-alignment":[83],"logits":[84,88],"with":[85,112],"original":[87],"obtain":[90],"harmless":[91],"helpful":[93],"distributions.Consequently,":[94],"our":[95,119],"method":[96],"enhances":[97],"safety":[98],"while":[100],"maintaining":[101],"helpfulness.We":[102],"conduct":[103],"experiments":[104],"across":[105],"five":[106],"four":[109],"common":[110],"jailbreaks,":[111],"results":[114],"validating":[115],"effectiveness":[117],"approach.Code":[120],"is":[121],"available":[122],"at":[123],"https://github.com/GIGABaozi/AED.":[124]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2024-11-28T00:00:00"}