{"id":"https://openalex.org/W4380136091","doi":"https://doi.org/10.48550/arxiv.2306.04735","title":"Soft-prompt Tuning for Large Language Models to Evaluate Bias","display_name":"Soft-prompt Tuning for Large Language Models to Evaluate Bias","publication_year":2023,"publication_date":"2023-06-07","ids":{"openalex":"https://openalex.org/W4380136091","doi":"https://doi.org/10.48550/arxiv.2306.04735"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2306.04735","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2306.04735","pdf_url":"https://arxiv.org/pdf/2306.04735","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2306.04735","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5104317088","display_name":"Jacob-Junqi Tian","orcid":"https://orcid.org/0000-0002-2922-8063","observed_orcids":["https://orcid.org/0000-0002-2922-8063"]},"institutions":[{"id":"https://openalex.org/I4210127509","display_name":"Vector Institute","ror":"https://ror.org/03kqdja62","country_code":"CA","type":"facility","lineage":["https://openalex.org/I4210127509"]},{"id":"https://openalex.org/I5023651","display_name":"McGill University","ror":"https://ror.org/01pxwe438","country_code":"CA","type":"education","lineage":["https://openalex.org/I5023651"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Tian, Jacob-Junqi","raw_affiliation_strings":["McGill University","Vector Institute for AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"McGill University","institution_ids":["https://openalex.org/I5023651"]},{"raw_affiliation_string":"Vector Institute for AI","institution_ids":["https://openalex.org/I4210127509"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090800503","display_name":"D. B. Emerson","orcid":"https://orcid.org/0000-0002-4630-7003","observed_orcids":["https://orcid.org/0000-0002-4630-7003"]},"institutions":[{"id":"https://openalex.org/I4210127509","display_name":"Vector Institute","ror":"https://ror.org/03kqdja62","country_code":"CA","type":"facility","lineage":["https://openalex.org/I4210127509"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Emerson, David","raw_affiliation_strings":["Vector Institute for AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Vector Institute for AI","institution_ids":["https://openalex.org/I4210127509"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092130651","display_name":"Sevil Zanjani Miyandoab","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I158248296","display_name":"Amirkabir University of Technology","ror":"https://ror.org/04gzbav43","country_code":"IR","type":"education","lineage":["https://openalex.org/I158248296"]}],"countries":["IR"],"is_corresponding":false,"raw_author_name":"Miyandoab, Sevil Zanjani","raw_affiliation_strings":["Amirkabir University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Amirkabir University","institution_ids":["https://openalex.org/I158248296"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108445535","display_name":"Deval A. Pandya","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210127509","display_name":"Vector Institute","ror":"https://ror.org/03kqdja62","country_code":"CA","type":"facility","lineage":["https://openalex.org/I4210127509"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Pandya, Deval","raw_affiliation_strings":["Vector Institute for AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Vector Institute for AI","institution_ids":["https://openalex.org/I4210127509"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004971542","display_name":"Laleh Seyyed-Kalantari","orcid":"https://orcid.org/0000-0002-1059-7125","observed_orcids":["https://orcid.org/0000-0002-1059-7125"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Seyyed-Kalantari, Laleh","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5009939236","display_name":"Faiza Khan Khattak","orcid":"https://orcid.org/0000-0002-9902-0583","observed_orcids":["https://orcid.org/0000-0002-9902-0583"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Khattak, Faiza Khan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.9160000085830688,"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"}},"topics":[{"id":"https://openalex.org/T10883","display_name":"Ethics and Social Impacts of AI","score":0.9160000085830688,"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":0.029100000858306885,"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/T10664","display_name":"Sentiment Analysis and Opinion Mining","score":0.008500000461935997,"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/soft-prompt-tuning","display_name":"soft prompt tuning","score":0.9070000052452087},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9309999942779541},{"id":"https://openalex.org/keywords/algorithmic-bias","display_name":"algorithmic bias","score":0.8519999980926514},{"id":"https://openalex.org/keywords/bias-assessment","display_name":"bias assessment","score":0.5619999766349792},{"id":"https://openalex.org/keywords/sentiment-classification","display_name":"sentiment classification","score":0.949999988079071},{"id":"https://openalex.org/keywords/group-fairness","display_name":"group fairness","score":0.968999981880188},{"id":"https://openalex.org/keywords/sensitive-attributes","display_name":"sensitive attributes","score":0.8539999723434448}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7514002323150635},{"id":"https://openalex.org/C2780586970","wikidata":"https://www.wikidata.org/wiki/Q1357284","display_name":"Popularity","level":2,"score":0.6154666543006897},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5513331294059753},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5037803053855896},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4552229642868042},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4205728769302368},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.41384726762771606},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4032255709171295},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.3851393163204193},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3265308141708374},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.16538739204406738},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.09751087427139282},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.08752888441085815},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2306.04735","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2306.04735","pdf_url":"https://arxiv.org/pdf/2306.04735","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2306.04735","doi":"https://doi.org/10.48550/arxiv.2306.04735","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2306.04735","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2306.04735","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2306.04735","pdf_url":"https://arxiv.org/pdf/2306.04735","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[{"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10","score":0.4000000059604645}],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4380136091.pdf","grobid_xml":"https://content.openalex.org/works/W4380136091.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7128690443","https://openalex.org/W7160709534","https://openalex.org/W7160745628","https://openalex.org/W4224031624","https://openalex.org/W7152934385","https://openalex.org/W7165040045","https://openalex.org/W4417064358","https://openalex.org/W4416280661","https://openalex.org/W4406123638","https://openalex.org/W7126376559"],"abstract_inverted_index":{"Prompting":[0],"large":[1,60],"language":[2,61,72],"models":[3,62,76,164],"has":[4],"gained":[5],"immense":[6],"popularity":[7],"in":[8,148,151,165],"recent":[9],"years":[10],"due":[11],"to":[12,31,37,55,86,96,103,157,175,179],"the":[13,21,46,57,108,113,126,134,149,159],"advantage":[14,110],"of":[15,48,59,91,111],"producing":[16],"good":[17],"results":[18],"even":[19],"without":[20],"need":[22],"for":[23],"labelled":[24],"data.":[25],"However,":[26],"this":[27,42],"requires":[28],"prompt":[29],"tuning":[30,50],"get":[32],"optimal":[33],"prompts":[34],"that":[35,82,116],"lead":[36],"better":[38],"model":[39,127],"performances.":[40],"In":[41],"paper,":[43],"we":[44],"explore":[45],"use":[47,181],"soft-prompt":[49],"on":[51,79,129],"sentiment":[52],"classification":[53],"task":[54],"quantify":[56],"biases":[58,128,160],"(LLMs)":[63],"such":[64],"as":[65],"Open":[66],"Pre-trained":[67],"Transformers":[68],"(OPT)":[69],"and":[70,138,171],"Galactica":[71],"model.":[73],"Since":[74,143],"these":[75,98,163],"are":[77],"trained":[78],"real-world":[80],"data":[81],"could":[83],"be":[84,118],"prone":[85],"bias":[87,105,141],"toward":[88],"certain":[89],"groups":[90],"populations,":[92],"it":[93,154],"is":[94,155],"important":[95],"identify":[97,158],"underlying":[99],"issues.":[100],"Using":[101],"soft-prompts":[102],"evaluate":[104],"gives":[106],"us":[107],"extra":[109],"avoiding":[112],"human-bias":[114],"injection":[115],"can":[117],"caused":[119],"by":[120],"manually":[121],"designed":[122],"prompts.":[123],"We":[124,167],"check":[125],"different":[130],"sensitive":[131],"attributes":[132],"using":[133],"group":[135],"fairness":[136],"(bias)":[137],"find":[139],"interesting":[140],"patterns.":[142],"LLMs":[144],"have":[145],"been":[146],"used":[147],"industry":[150,173],"various":[152],"applications,":[153],"crucial":[156],"before":[161],"deploying":[162],"practice.":[166],"open-source":[168],"our":[169,177],"pipeline":[170],"encourage":[172],"researchers":[174],"adapt":[176],"work":[178],"their":[180],"cases.":[182]},"counts_by_year":[{"year":2024,"cited_by_count":2}],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2023-06-10T00:00:00"}