{"id":"https://openalex.org/W4406495567","doi":"https://doi.org/10.1109/bigdata62323.2024.10825538","title":"Mitigating Sycophancy in Large Language Models via Direct Preference Optimization","display_name":"Mitigating Sycophancy in Large Language Models via Direct Preference Optimization","publication_year":2024,"publication_date":"2024-12-15","ids":{"openalex":"https://openalex.org/W4406495567","doi":"https://doi.org/10.1109/bigdata62323.2024.10825538"},"language":"en","primary_location":{"id":"doi:10.1109/bigdata62323.2024.10825538","is_oa":false,"landing_page_url":"https://doi.org/10.1109/bigdata62323.2024.10825538","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE International Conference on Big Data (BigData)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5063807294","display_name":"Azal Ahmad Khan","orcid":"https://orcid.org/0009-0000-9435-5328"},"institutions":[{"id":"https://openalex.org/I4210101327","display_name":"Twin Cities Orthopedics","ror":"https://ror.org/01en4s460","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210101327"]},{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]},{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Azal Ahmad Khan","raw_affiliation_strings":["University of Minnesota-Twin Cities,Department of Computer Science and Engineering","Virginia Tech,Department of Computer Science"],"affiliations":[{"raw_affiliation_string":"University of Minnesota-Twin Cities,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I4210101327","https://openalex.org/I130238516"]},{"raw_affiliation_string":"Virginia Tech,Department of Computer Science","institution_ids":["https://openalex.org/I859038795"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067540712","display_name":"Shariful Alam","orcid":"https://orcid.org/0000-0001-8263-117X"},"institutions":[{"id":"https://openalex.org/I1317621060","display_name":"Indian Institute of Technology Guwahati","ror":"https://ror.org/0022nd079","country_code":"IN","type":"education","lineage":["https://openalex.org/I1317621060"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Sayan Alam","raw_affiliation_strings":["Indian Institute of Technology,Guwahati"],"affiliations":[{"raw_affiliation_string":"Indian Institute of Technology,Guwahati","institution_ids":["https://openalex.org/I1317621060"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100458486","display_name":"Xinran Wang","orcid":"https://orcid.org/0000-0003-1120-8248"},"institutions":[{"id":"https://openalex.org/I4210101327","display_name":"Twin Cities Orthopedics","ror":"https://ror.org/01en4s460","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210101327"]},{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Xinran Wang","raw_affiliation_strings":["University of Minnesota-Twin Cities,Department of Computer Science and Engineering"],"affiliations":[{"raw_affiliation_string":"University of Minnesota-Twin Cities,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I4210101327","https://openalex.org/I130238516"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050886354","display_name":"Ahmad Khan","orcid":"https://orcid.org/0000-0002-6955-8876"},"institutions":[{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]},{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]},{"id":"https://openalex.org/I4210101327","display_name":"Twin Cities Orthopedics","ror":"https://ror.org/01en4s460","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210101327"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ahmad Faraz Khan","raw_affiliation_strings":["University of Minnesota-Twin Cities,Department of Computer Science and Engineering","Virginia Tech,Department of Computer Science"],"affiliations":[{"raw_affiliation_string":"University of Minnesota-Twin Cities,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I4210101327","https://openalex.org/I130238516"]},{"raw_affiliation_string":"Virginia Tech,Department of Computer Science","institution_ids":["https://openalex.org/I859038795"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039364567","display_name":"Debanga Raj Neog","orcid":"https://orcid.org/0000-0002-2794-4787"},"institutions":[{"id":"https://openalex.org/I1317621060","display_name":"Indian Institute of Technology Guwahati","ror":"https://ror.org/0022nd079","country_code":"IN","type":"education","lineage":["https://openalex.org/I1317621060"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Debanga Raj Neog","raw_affiliation_strings":["Indian Institute of Technology,Guwahati"],"affiliations":[{"raw_affiliation_string":"Indian Institute of Technology,Guwahati","institution_ids":["https://openalex.org/I1317621060"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5054645319","display_name":"Ali Anwar","orcid":"https://orcid.org/0000-0003-4487-2436"},"institutions":[{"id":"https://openalex.org/I4210101327","display_name":"Twin Cities Orthopedics","ror":"https://ror.org/01en4s460","country_code":"US","type":"healthcare","lineage":["https://openalex.org/I4210101327"]},{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ali Anwar","raw_affiliation_strings":["University of Minnesota-Twin Cities,Department of Computer Science and Engineering"],"affiliations":[{"raw_affiliation_string":"University of Minnesota-Twin Cities,Department of Computer Science and Engineering","institution_ids":["https://openalex.org/I4210101327","https://openalex.org/I130238516"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5063807294"],"corresponding_institution_ids":["https://openalex.org/I130238516","https://openalex.org/I4210101327","https://openalex.org/I859038795"],"apc_list":null,"apc_paid":null,"fwci":1.0142,"has_fulltext":false,"cited_by_count":3,"citation_normalized_percentile":{"value":0.82045985,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1664","last_page":"1671"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":1.0,"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/T10028","display_name":"Topic Modeling","score":1.0,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9994000196456909,"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/T12031","display_name":"Speech and dialogue systems","score":0.9976000189781189,"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/computer-science","display_name":"Computer science","score":0.6812616586685181},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.6187500357627869},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.09428280591964722},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.06896263360977173}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6812616586685181},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.6187500357627869},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09428280591964722},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.06896263360977173}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/bigdata62323.2024.10825538","is_oa":false,"landing_page_url":"https://doi.org/10.1109/bigdata62323.2024.10825538","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE International Conference on Big Data (BigData)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":104,"referenced_works":["https://openalex.org/W2896457183","https://openalex.org/W2952357537","https://openalex.org/W2963026768","https://openalex.org/W2970352191","https://openalex.org/W2970771982","https://openalex.org/W2973379954","https://openalex.org/W2990138404","https://openalex.org/W3001279689","https://openalex.org/W3034238904","https://openalex.org/W3034408878","https://openalex.org/W3168867926","https://openalex.org/W3174770825","https://openalex.org/W3177813494","https://openalex.org/W3195577433","https://openalex.org/W4224308101","https://openalex.org/W4287674181","https://openalex.org/W4307079201","https://openalex.org/W4308902180","https://openalex.org/W4311991106","https://openalex.org/W4322718191","https://openalex.org/W4362515116","https://openalex.org/W4362707064","https://openalex.org/W4365211596","https://openalex.org/W4378509449","https://openalex.org/W4378765257","https://openalex.org/W4378771755","https://openalex.org/W4384918448","https://openalex.org/W4385373622","https://openalex.org/W4385571158","https://openalex.org/W4385714610","https://openalex.org/W4385894687","https://openalex.org/W4386556219","https://openalex.org/W4386557569","https://openalex.org/W4386793209","https://openalex.org/W4387156634","https://openalex.org/W4387165392","https://openalex.org/W4387801352","https://openalex.org/W4387892136","https://openalex.org/W4388717471","https://openalex.org/W4388747879","https://openalex.org/W4388787482","https://openalex.org/W4389520524","https://openalex.org/W4390047809","https://openalex.org/W4390833061","https://openalex.org/W4390833718","https://openalex.org/W4391555989","https://openalex.org/W4391835799","https://openalex.org/W4392240262","https://openalex.org/W4393160410","https://openalex.org/W4402670807","https://openalex.org/W4402670859","https://openalex.org/W4404792953","https://openalex.org/W6739585900","https://openalex.org/W6745026867","https://openalex.org/W6755207826","https://openalex.org/W6759579507","https://openalex.org/W6764456104","https://openalex.org/W6767858076","https://openalex.org/W6772383348","https://openalex.org/W6778883912","https://openalex.org/W6782465632","https://openalex.org/W6796581206","https://openalex.org/W6798182279","https://openalex.org/W6800751262","https://openalex.org/W6804664148","https://openalex.org/W6810081322","https://openalex.org/W6810738896","https://openalex.org/W6846930601","https://openalex.org/W6846937427","https://openalex.org/W6847076894","https://openalex.org/W6847753483","https://openalex.org/W6850625674","https://openalex.org/W6851009429","https://openalex.org/W6851775633","https://openalex.org/W6851960618","https://openalex.org/W6852065897","https://openalex.org/W6852222651","https://openalex.org/W6852418670","https://openalex.org/W6852800892","https://openalex.org/W6852927592","https://openalex.org/W6853251322","https://openalex.org/W6853986968","https://openalex.org/W6854866820","https://openalex.org/W6854948896","https://openalex.org/W6855129433","https://openalex.org/W6855723013","https://openalex.org/W6856273322","https://openalex.org/W6856535631","https://openalex.org/W6856587086","https://openalex.org/W6856729674","https://openalex.org/W6856860495","https://openalex.org/W6857010465","https://openalex.org/W6857273126","https://openalex.org/W6857294608","https://openalex.org/W6858023062","https://openalex.org/W6858493252","https://openalex.org/W6858708973","https://openalex.org/W6858731683","https://openalex.org/W6859347454","https://openalex.org/W6859529408","https://openalex.org/W6860318346","https://openalex.org/W6860465829","https://openalex.org/W6861324011","https://openalex.org/W6861440952"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Large":[0],"language":[1,149],"models":[2,150],"(LLMs)":[3],"have":[4],"demonstrated":[5],"remarkable":[6],"capabilities,":[7],"yet":[8],"they":[9],"occasionally":[10],"exhibit":[11],"sycophantic":[12,37,70,110,137],"behavior,":[13],"generating":[14],"responses":[15,73,89,113],"that":[16,90,151],"align":[17,91],"with":[18,21,69,92,109,159],"or":[19,26,34],"agree":[20],"a":[22,50,62,104],"user\u2019s":[23],"stated":[24],"opinions":[25,31],"preferences,":[27],"even":[28],"when":[29],"those":[30],"are":[32],"incorrect":[33],"biased.":[35],"This":[36,47],"tendency":[38],"can":[39,152],"undermine":[40],"the":[41,93,142],"trustworthiness":[42],"and":[43,71,111,128,147,155],"reliability":[44],"of":[45,106,123,136],"LLMs.":[46,116],"work":[48],"proposes":[49],"novel":[51],"approach":[52,118],"to":[53,87,114],"mitigate":[54],"sycophancy":[55],"in":[56,125,130],"LLMs":[57,86],"by":[58],"fine-tuning":[59],"them":[60],"on":[61],"carefully":[63],"curated":[64],"dataset":[65,105],"comprising":[66],"prompts":[67,108],"paired":[68],"non-sycophantic":[72,112],"<sup":[74],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[75],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">1</sup>.":[76],"Our":[77,117,139],"method":[78],"leverages":[79],"Direct":[80],"Preference":[81],"Optimization":[82],"(DPO),":[83],"which":[84],"optimizes":[85],"generate":[88],"preferred":[94],"(non-sycophantic)":[95],"outputs":[96],"without":[97],"requiring":[98],"explicit":[99],"reward":[100],"modeling.":[101],"We":[102],"develop":[103],"1000":[107],"fine-tune":[115],"achieves":[119],"an":[120],"average":[121],"reduction":[122],"85%":[124],"persona-based":[126],"tests":[127],"84%":[129],"preference-driven":[131],"tests,":[132],"demonstrating":[133],"significant":[134],"mitigation":[135],"behaviors.":[138],"findings":[140],"pave":[141],"way":[143],"for":[144],"more":[145],"trustworthy":[146],"reliable":[148],"provide":[153],"objective":[154],"unbiased":[156],"responses,":[157],"aligning":[158],"human":[160],"preferences":[161],"while":[162],"maintaining":[163],"factual":[164],"accuracy.":[165]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":2}],"updated_date":"2026-04-14T08:04:32.555800","created_date":"2025-10-10T00:00:00"}