{"id":"https://openalex.org/W7166809914","doi":"https://doi.org/10.18653/v1/2026.findings-acl.933","title":"Be Your Own Red Teamer: Safety Alignment via Self-Play and Reflective Experience Replay","display_name":"Be Your Own Red Teamer: Safety Alignment via Self-Play and Reflective Experience Replay","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166809914","doi":"https://doi.org/10.18653/v1/2026.findings-acl.933"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.933","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.933","pdf_url":"https://aclanthology.org/2026.findings-acl.933.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 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.933.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139793910","display_name":"Hao Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hao Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139820809","display_name":"Yanting Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yanting Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139836816","display_name":"Hao Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hao Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139770245","display_name":"Rui Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rui Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139734966","display_name":"Lei Sha","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lei Sha","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.848434,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"18700","last_page":"18716"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T13219","display_name":"Mind wandering and attention","score":0.11079999804496765,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},"topics":[{"id":"https://openalex.org/T13219","display_name":"Mind wandering and attention","score":0.11079999804496765,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10525","display_name":"Human-Automation Interaction and Safety","score":0.0608999989926815,"subfield":{"id":"https://openalex.org/subfields/3207","display_name":"Social Psychology"},"field":{"id":"https://openalex.org/fields/32","display_name":"Psychology"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T12607","display_name":"Personal Information Management and User Behavior","score":0.04610000178217888,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.30709999799728394},{"id":"https://openalex.org/keywords/troubleshooting","display_name":"Troubleshooting","score":0.2872999906539917},{"id":"https://openalex.org/keywords/focus","display_name":"Focus (optics)","score":0.2759000062942505},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.2745000123977661},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.26010000705718994}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5432000160217285},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.33970001339912415},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.30709999799728394},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.29350000619888306},{"id":"https://openalex.org/C147494362","wikidata":"https://www.wikidata.org/wiki/Q2078905","display_name":"Troubleshooting","level":2,"score":0.2872999906539917},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.2759000062942505},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2745000123977661},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.266400009393692},{"id":"https://openalex.org/C49774154","wikidata":"https://www.wikidata.org/wiki/Q131765","display_name":"Multimedia","level":1,"score":0.26429998874664307},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.26010000705718994},{"id":"https://openalex.org/C65682993","wikidata":"https://www.wikidata.org/wiki/Q1056451","display_name":"Reflection (computer programming)","level":2,"score":0.2574000060558319},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.25209999084472656}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.933","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.933","pdf_url":"https://aclanthology.org/2026.findings-acl.933.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 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.933","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.933","pdf_url":"https://aclanthology.org/2026.findings-acl.933.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 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4458316040","display_name":null,"funder_award_id":"KZ37117501","funder_id":"https://openalex.org/F4320333688","funder_display_name":"National Outstanding Youth Science Fund Project of National Natural Science Foundation of China"},{"id":"https://openalex.org/G5831522889","display_name":null,"funder_award_id":"62306024","funder_id":"https://openalex.org/F4320321001","funder_display_name":"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/F4320329860","display_name":"National Science and Technology Major Project","ror":null},{"id":"https://openalex.org/F4320333688","display_name":"National Outstanding Youth Science Fund Project of National Natural Science Foundation of China","ror":null},{"id":"https://openalex.org/F4320335787","display_name":"Fundamental Research Funds for the Central Universities","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166809914.pdf","grobid_xml":"https://content.openalex.org/works/W7166809914.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"achieved":[5],"remarkable":[6],"capabilities":[7],"but":[8],"remain":[9],"vulnerable":[10],"to":[11,16,39,49,51,64,90,116,161],"adversarial":[12,36,76,199],"\"jailbreak\"":[13],"attacks":[14],"designed":[15],"bypass":[17],"safety":[18,20,130,208],"guardrails.Current":[19],"alignment":[21],"methods":[22],"depend":[23],"heavily":[24],"on":[25,163,197],"static":[26,198],"external":[27],"red":[28,68],"teaming,":[29],"utilizing":[30],"fixed":[31],"defense":[32,42,122,191],"prompts":[33],"or":[34],"pre-collected":[35],"datasets.This":[37],"leads":[38],"a":[40,83,87,106,154,203],"rigid":[41],"that":[43,85,184],"overfits":[44],"known":[45],"patterns":[46],"and":[47,74,99,180,201],"fails":[48],"generalize":[50],"novel,":[52],"sophisticated":[53],"threats.To":[54],"address":[55],"this":[56],"critical":[57,129],"limitation,":[58],"we":[59,78,135],"propose":[60],"empowering":[61],"the":[62,95,100,125,133,150,170],"model":[63,171],"be":[65],"its":[66],"own":[67],"teamer,":[69],"capable":[70],"of":[71],"achieving":[72],"autonomous":[73],"evolving":[75,113],"attacks.Specifically,":[77],"introduce":[79,136],"Safety":[80],"Self-Play":[81],"(SSP),":[82],"system":[84],"utilizes":[86],"single":[88],"LLM":[89],"act":[91],"concurrently":[92],"as":[93],"both":[94],"Attacker":[96],"(generating":[97],"jailbreaks)":[98],"Defender":[101,126],"(refusing":[102],"harmful":[103],"requests)":[104],"within":[105],"unified":[107],"Reinforcement":[108],"Learning":[109],"(RL)":[110],"loop,":[111],"dynamically":[112],"attack":[114],"strategies":[115],"uncover":[117],"vulnerabilities":[118],"while":[119,177],"simultaneously":[120],"strengthening":[121],"mechanisms.To":[123],"ensure":[124],"effectively":[127],"addresses":[128],"issues":[131],"during":[132],"selfplay,":[134],"an":[137,145],"advanced":[138],"Reflective":[139],"Experience":[140],"Replay":[141],"Mechanism,":[142],"which":[143],"uses":[144],"experience":[146],"pool":[147],"accumulated":[148],"throughout":[149],"process.The":[151],"mechanism":[152],"employs":[153],"Upper":[155],"Confidence":[156],"Bound":[157],"(UCB)":[158],"sampling":[159],"strategy":[160],"focus":[162],"failure":[164],"cases":[165],"with":[166],"low":[167],"rewards,":[168],"helping":[169],"learn":[172],"from":[173],"past":[174],"hard":[175],"mistakes":[176],"balancing":[178],"exploration":[179],"exploitation.Extensive":[181],"experiments":[182],"demonstrate":[183],"our":[185],"SSP":[186],"approach":[187],"autonomously":[188],"evolves":[189],"robust":[190],"capabilities,":[192],"significantly":[193],"outperforming":[194],"baselines":[195],"trained":[196],"datasets":[200],"establishing":[202],"new":[204],"benchmark":[205],"for":[206],"proactive":[207],"alignment.":[209]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}
