{"id":"https://openalex.org/W4400103834","doi":"https://doi.org/10.48550/arxiv.2406.18118","title":"SafeAligner: Safety Alignment against Jailbreak Attacks via Response Disparity Guidance","display_name":"SafeAligner: Safety Alignment against Jailbreak Attacks via Response Disparity Guidance","publication_year":2024,"publication_date":"2024-06-26","ids":{"openalex":"https://openalex.org/W4400103834","doi":"https://doi.org/10.48550/arxiv.2406.18118"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2406.18118","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2406.18118","pdf_url":"https://arxiv.org/pdf/2406.18118","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,"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"},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2406.18118","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5016693603","display_name":"Caishuang Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Caishuang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101388158","display_name":"Wanxu Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Wanxu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103144065","display_name":"Rui Zheng","orcid":"https://orcid.org/0000-0002-5334-2640"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Rui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104337938","display_name":"Huijie Lv","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lv, Huijie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Zhan, Wenyu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhan, Wenyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5083173421","display_name":"Shihan Dou","orcid":"https://orcid.org/0009-0002-6013-3035"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dou, Shihan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000525969","display_name":"Sixian Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Sixian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112719601","display_name":"Xiao Wang","orcid":"https://orcid.org/0000-0002-0008-0659"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Xiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087033501","display_name":"Enyu Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Enyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100777718","display_name":"Junjie Ye","orcid":"https://orcid.org/0009-0004-0921-6323"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ye, Junjie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100958962","display_name":"Yuming Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Yuming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058353652","display_name":"Tao Gui","orcid":"https://orcid.org/0000-0002-6154-0751"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gui, Tao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100360194","display_name":"Qi Zhang","orcid":"https://orcid.org/0000-0001-5303-9804"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Huang, Xuanjing","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Xuanjing","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":2,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10734","display_name":"Information and Cyber Security","score":0.993399977684021,"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":0.993399977684021,"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"}},{"id":"https://openalex.org/T10400","display_name":"Network Security and Intrusion Detection","score":0.988099992275238,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T11241","display_name":"Advanced Malware Detection Techniques","score":0.9830999970436096,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/computer-science","display_name":"Computer science","score":0.4586515426635742},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.4546918272972107}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4586515426635742},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.4546918272972107}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2406.18118","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2406.18118","pdf_url":"https://arxiv.org/pdf/2406.18118","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,"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.2406.18118","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2406.18118","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2406.18118","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2406.18118","pdf_url":"https://arxiv.org/pdf/2406.18118","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,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4400103834.pdf","grobid_xml":"https://content.openalex.org/works/W4400103834.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052","https://openalex.org/W2382290278","https://openalex.org/W4395014643"],"abstract_inverted_index":{"As":[0],"the":[1,60,76,86,96,102,116,121,126,136,143],"development":[2],"of":[3,23,125,138,145],"large":[4],"language":[5],"models":[6,12,106],"(LLMs)":[7],"rapidly":[8],"advances,":[9],"securing":[10],"these":[11,51,105],"effectively":[13,114],"without":[14],"compromising":[15],"their":[16],"utility":[17],"has":[18],"become":[19],"a":[20,56],"pivotal":[21],"area":[22],"research.":[24],"However,":[25],"current":[26],"defense":[27],"strategies":[28],"against":[29,66],"jailbreak":[30,67],"attacks":[31],"(i.e.,":[32],"efforts":[33],"to":[34,63,82,90,107,155],"bypass":[35],"security":[36,99],"protocols)":[37],"often":[38],"suffer":[39],"from":[40,104],"limited":[41],"adaptability,":[42],"restricted":[43],"general":[44],"capability,":[45],"and":[46,85,111],"high":[47],"cost.":[48],"To":[49],"address":[50],"challenges,":[52],"we":[53],"introduce":[54],"SafeAligner,":[55],"methodology":[57],"implemented":[58],"at":[59],"decoding":[61],"stage":[62],"fortify":[64],"defenses":[65],"attacks.":[68],"We":[69],"begin":[70],"by":[71,119],"developing":[72],"two":[73],"specialized":[74],"models:":[75],"Sentinel":[77],"Model,":[78,88],"which":[79],"is":[80],"trained":[81],"foster":[83],"safety,":[84],"Intruder":[87],"designed":[89],"generate":[91],"riskier":[92],"responses.":[93],"SafeAligner":[94,133],"leverages":[95],"disparity":[97],"in":[98],"levels":[100],"between":[101,109],"responses":[103],"differentiate":[108],"harmful":[110,146],"beneficial":[112,139],"tokens,":[113,140],"guiding":[115],"safety":[117],"alignment":[118,151],"altering":[120],"output":[122],"token":[123],"distribution":[124],"target":[127],"model.":[128],"Extensive":[129],"experiments":[130],"show":[131],"that":[132],"can":[134],"increase":[135],"likelihood":[137],"while":[141],"reducing":[142],"occurrence":[144],"ones,":[147],"thereby":[148],"ensuring":[149],"secure":[150],"with":[152],"minimal":[153],"loss":[154],"generality.":[156]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2026-08-26T07:47:46.906454","created_date":"2025-10-10T00:00:00"}
