{"id":"https://openalex.org/W7137927306","doi":"https://doi.org/10.1609/aaai.v40i37.40372","title":"Model Editing as a Double-Edged Sword: Steering Agent Behavior Toward Beneficence or Harm","display_name":"Model Editing as a Double-Edged Sword: Steering Agent Behavior Toward Beneficence or Harm","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7137927306","doi":"https://doi.org/10.1609/aaai.v40i37.40372"},"language":"en","primary_location":{"id":"doi:10.1609/aaai.v40i37.40372","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i37.40372","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1609/aaai.v40i37.40372","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129746883","display_name":"Baixiang Huang","orcid":null},"institutions":[{"id":"https://openalex.org/I150468666","display_name":"Emory University","ror":"https://ror.org/03czfpz43","country_code":"US","type":"education","lineage":["https://openalex.org/I150468666"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Baixiang Huang","raw_affiliation_strings":["Emory University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Emory University","institution_ids":["https://openalex.org/I150468666"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129738799","display_name":"Zhen Tan","orcid":null},"institutions":[{"id":"https://openalex.org/I55732556","display_name":"Arizona State University","ror":"https://ror.org/03efmqc40","country_code":"US","type":"education","lineage":["https://openalex.org/I55732556"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhen Tan","raw_affiliation_strings":["Arizona State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Arizona State University","institution_ids":["https://openalex.org/I55732556"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129738882","display_name":"Haoran Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I150468666","display_name":"Emory University","ror":"https://ror.org/03czfpz43","country_code":"US","type":"education","lineage":["https://openalex.org/I150468666"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Haoran Wang","raw_affiliation_strings":["Emory University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Emory University","institution_ids":["https://openalex.org/I150468666"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129736858","display_name":"Zijie Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I114027177","display_name":"University of North Carolina at Chapel Hill","ror":"https://ror.org/0130frc33","country_code":"US","type":"education","lineage":["https://openalex.org/I114027177"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zijie Liu","raw_affiliation_strings":["University of North Carolina at Chapel Hill"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of North Carolina at Chapel Hill","institution_ids":["https://openalex.org/I114027177"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129748132","display_name":"Dawei Li","orcid":null},"institutions":[{"id":"https://openalex.org/I55732556","display_name":"Arizona State University","ror":"https://ror.org/03efmqc40","country_code":"US","type":"education","lineage":["https://openalex.org/I55732556"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dawei Li","raw_affiliation_strings":["Arizona State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Arizona State University","institution_ids":["https://openalex.org/I55732556"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129693663","display_name":"Ali Payani","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ali Payani","raw_affiliation_strings":["Cisco"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Cisco","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129662930","display_name":"Huan Liu","orcid":null},"institutions":[{"id":"https://openalex.org/I55732556","display_name":"Arizona State University","ror":"https://ror.org/03efmqc40","country_code":"US","type":"education","lineage":["https://openalex.org/I55732556"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Huan Liu","raw_affiliation_strings":["Arizona State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Arizona State University","institution_ids":["https://openalex.org/I55732556"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129694069","display_name":"Tianlong Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I114027177","display_name":"University of North Carolina at Chapel Hill","ror":"https://ror.org/0130frc33","country_code":"US","type":"education","lineage":["https://openalex.org/I114027177"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tianlong Chen","raw_affiliation_strings":["University of North Carolina at Chapel Hill"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of North Carolina at Chapel Hill","institution_ids":["https://openalex.org/I114027177"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129711308","display_name":"Kai Shu","orcid":null},"institutions":[{"id":"https://openalex.org/I150468666","display_name":"Emory University","ror":"https://ror.org/03czfpz43","country_code":"US","type":"education","lineage":["https://openalex.org/I150468666"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Kai Shu","raw_affiliation_strings":["Emory University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Emory University","institution_ids":["https://openalex.org/I150468666"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"37","first_page":"31113","last_page":"31121"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.2639999985694885,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.2639999985694885,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.1331000030040741,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"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":0.10570000112056732,"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/harm","display_name":"Harm","score":0.6140000224113464},{"id":"https://openalex.org/keywords/variety","display_name":"Variety (cybernetics)","score":0.5192999839782715},{"id":"https://openalex.org/keywords/beneficence","display_name":"Beneficence","score":0.5001000165939331},{"id":"https://openalex.org/keywords/frame","display_name":"Frame (networking)","score":0.46380001306533813},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.41339999437332153},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.36000001430511475}],"concepts":[{"id":"https://openalex.org/C2777363581","wikidata":"https://www.wikidata.org/wiki/Q15098235","display_name":"Harm","level":2,"score":0.6140000224113464},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5306000113487244},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.5192999839782715},{"id":"https://openalex.org/C117438556","wikidata":"https://www.wikidata.org/wiki/Q4887413","display_name":"Beneficence","level":3,"score":0.5001000165939331},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.46380001306533813},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.45890000462532043},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.41339999437332153},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.36000001430511475},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.33309999108314514},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.313400000333786},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.3000999987125397},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.2953999936580658},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.2840000092983246},{"id":"https://openalex.org/C2780873155","wikidata":"https://www.wikidata.org/wiki/Q392811","display_name":"Agent-based model","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.2653000056743622},{"id":"https://openalex.org/C2777834853","wikidata":"https://www.wikidata.org/wiki/Q96776939","display_name":"Liability","level":2,"score":0.2615000009536743},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.26030001044273376},{"id":"https://openalex.org/C2776035688","wikidata":"https://www.wikidata.org/wiki/Q1606558","display_name":"Affect (linguistics)","level":2,"score":0.2599000036716461},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.25870001316070557},{"id":"https://openalex.org/C14262774","wikidata":"https://www.wikidata.org/wiki/Q4880695","display_name":"Behavior change","level":2,"score":0.25769999623298645},{"id":"https://openalex.org/C144501496","wikidata":"https://www.wikidata.org/wiki/Q5533489","display_name":"Genome editing","level":4,"score":0.25380000472068787}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1609/aaai.v40i37.40372","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i37.40372","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},{"id":"pmh:oai:ojs.aaai.org:article/40372","is_oa":false,"landing_page_url":"https://ojs.aaai.org/index.php/AAAI/article/view/40372","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"2159-5399","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i37.40372","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i37.40372","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[{"score":0.4128245413303375,"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Agents":[0],"based":[1],"on":[2,201],"Large":[3],"Language":[4],"Models":[5],"(LLMs)":[6],"have":[7],"demonstrated":[8],"strong":[9],"capabilities":[10],"across":[11,122,211],"a":[12,63,103,123,212,225],"wide":[13,213],"range":[14,214],"of":[15,55,78,119,125,198,208,215,238],"tasks.":[16],"However,":[17],"deploying":[18],"LLM-based":[19],"agents":[20,35,145,199],"in":[21,39,107,167],"high-stakes":[22],"domains":[23],"comes":[24],"with":[25,127],"significant":[26],"safety":[27],"and":[28,46,83,96,117,133,184,217,236],"ethical":[29,53,183],"risks.":[30],"Unethical":[31],"behavior":[32,54,60,149,186,209],"by":[33],"these":[34],"can":[36,142,178],"directly":[37],"result":[38],"serious":[40],"real-world":[41],"consequences,":[42],"including":[43],"physical":[44],"harm":[45],"financial":[47],"loss.":[48],"To":[49,93],"efficiently":[50],"steer":[51,144],"the":[52,115,147,206,234],"agents,":[56],"we":[57,68,100],"frame":[58],"agent":[59,120,230],"steering":[61,229],"as":[62],"model":[64],"editing":[65,73,118,210],"task,":[66],"which":[67],"term":[69],"Behavior":[70,140,154,176,239],"Editing.":[71,240],"Model":[72],"is":[74],"an":[75,168],"emerging":[76],"area":[77],"research":[79],"that":[80,139,175],"enables":[81,156],"precise":[82],"efficient":[84],"modifications":[85],"to":[86,181,189],"LLMs":[87],"while":[88],"preserving":[89],"their":[90],"overall":[91],"capabilities.":[92],"systematically":[94],"study":[95],"evaluate":[97],"this":[98],"approach,":[99],"introduce":[101],"BehaviorBench,":[102],"multi-tier":[104],"benchmark":[105,112],"grounded":[106],"psychological":[108],"moral":[109,171],"theories.":[110],"This":[111],"supports":[113],"both":[114,233],"evaluation":[116],"behaviors":[121],"variety":[124],"scenarios,":[126],"each":[128],"tier":[129],"introducing":[130],"more":[131,164],"complex":[132],"ambiguous":[134],"scenarios.":[135,152,218],"We":[136,173],"first":[137],"demonstrate":[138,174],"Editing":[141,155,177],"dynamically":[143],"toward":[146],"target":[148],"within":[150],"specific":[151],"Moreover,":[153],"not":[157],"only":[158],"scenario-specific":[159],"local":[160],"adjustments":[161],"but":[162],"also":[163],"extensive":[165,196],"shifts":[166],"agent\u2019s":[169],"global":[170],"alignment.":[172],"be":[179],"used":[180],"promote":[182],"benevolent":[185],"or,":[187],"conversely,":[188],"induce":[190],"harmful":[191],"or":[192],"malicious":[193],"behavior.":[194],"Through":[195],"evaluations":[197],"built":[200],"frontier":[202],"LLMs,":[203],"BehaviorBench":[204],"validates":[205],"effectiveness":[207],"models":[216],"Our":[219],"findings":[220],"offer":[221],"key":[222],"insights":[223],"into":[224],"new":[226],"paradigm":[227],"for":[228],"behavior,":[231],"highlighting":[232],"promise":[235],"perils":[237]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-03-18T00:00:00"}