{"id":"https://openalex.org/W4416034303","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1003","title":"From n-gram to Attention: How Model Architectures Learn and Propagate Bias in Language Modeling","display_name":"From n-gram to Attention: How Model Architectures Learn and Propagate Bias in Language Modeling","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416034303","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1003"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1003","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1003","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1003.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: EMNLP 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.1003.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5058398007","display_name":"Mohsinul Kabir","orcid":"https://orcid.org/0000-0001-5899-6355"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mohsinul Kabir","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5083526755","display_name":"Tasfia Tahsin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tasfia Tahsin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5077976343","display_name":"Sophia Ananiadou","orcid":"https://orcid.org/0000-0002-4097-9191"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sophia Ananiadou","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":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"18478","last_page":"18498"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.382999986410141,"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":0.382999986410141,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.10849999636411667,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T13910","display_name":"Computational and Text Analysis Methods","score":0.07880000025033951,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5737000107765198},{"id":"https://openalex.org/keywords/architecture","display_name":"Architecture","score":0.4171999990940094},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.40290001034736633},{"id":"https://openalex.org/keywords/tracing","display_name":"Tracing","score":0.37700000405311584},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.33799999952316284},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.323199987411499}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7200999855995178},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5737000107765198},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45829999446868896},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.4171999990940094},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.40290001034736633},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.38429999351501465},{"id":"https://openalex.org/C138673069","wikidata":"https://www.wikidata.org/wiki/Q322229","display_name":"Tracing","level":2,"score":0.37700000405311584},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.37610000371932983},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.33799999952316284},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.32910001277923584},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32429999113082886},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.323199987411499},{"id":"https://openalex.org/C183322885","wikidata":"https://www.wikidata.org/wiki/Q17007702","display_name":"Context model","level":3,"score":0.3149999976158142},{"id":"https://openalex.org/C74672266","wikidata":"https://www.wikidata.org/wiki/Q815859","display_name":"Language acquisition","level":2,"score":0.2921999990940094},{"id":"https://openalex.org/C78639753","wikidata":"https://www.wikidata.org/wiki/Q3318160","display_name":"Behavioral modeling","level":2,"score":0.27950000762939453},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.25780001282691956},{"id":"https://openalex.org/C2983427547","wikidata":"https://www.wikidata.org/wiki/Q93200","display_name":"Gender bias","level":2,"score":0.25769999623298645},{"id":"https://openalex.org/C179603123","wikidata":"https://www.wikidata.org/wiki/Q1941921","display_name":"Modeling language","level":3,"score":0.2500999867916107}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.1003","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1003","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1003.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: EMNLP 2025","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2505.12381","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.12381","pdf_url":"https://arxiv.org/pdf/2505.12381","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1003","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1003","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1003.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: EMNLP 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416034303.pdf","grobid_xml":"https://content.openalex.org/works/W4416034303.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Current":[0],"research":[1],"on":[2,10,61],"bias":[3,56,81,97,123,146],"in":[4,40,55,96],"language":[5,59,133],"models":[6,134],"(LMs)":[7],"predominantly":[8],"focuses":[9],"data":[11,51,110,152],"quality,":[12],"with":[13,125],"significantly":[14,111],"less":[15],"attention":[16],"paid":[17],"to":[18,44,67,92,121,147,159],"model":[19,53,74,117,154],"architecture":[20,54],"and":[21,52,77,114,153],"temporal":[22,78,106],"influences":[23],"of":[24,34,108],"data.Even":[25],"more":[26],"critically,":[27],"few":[28],"studies":[29],"systematically":[30],"investigate":[31],"the":[32,46,105,140],"origins":[33,149],"bias.We":[35],"propose":[36],"a":[37,143],"methodology":[38],"grounded":[39],"comparative":[41],"behavioral":[42],"theory":[43],"interpret":[45],"complex":[47],"interaction":[48],"between":[49],"training":[50,109],"propagation":[57],"during":[58],"modeling.Building":[60],"recent":[62],"work":[63],"that":[64],"relates":[65],"transformers":[66,100],"n-gram":[68,87],"LMs,":[69],"we":[70],"evaluate":[71],"how":[72],"data,":[73],"design":[75],"choices,":[76],"dynamics":[79],"affect":[80],"propagation.Our":[82],"findings":[83,138],"reveal":[84],"that:":[85],"(1)":[86],"LMs":[88],"are":[89],"highly":[90],"sensitive":[91],"context":[93],"window":[94],"size":[95],"propagation,":[98],"while":[99],"demonstrate":[101],"architectural":[102],"robustness;":[103],"(2)":[104],"provenance":[107],"affects":[112],"bias;":[113],"(3)":[115],"different":[116],"architectures":[118],"respond":[119],"differentially":[120],"controlled":[122],"injection,":[124],"certain":[126],"biases":[127],"(e.g.sexual":[128],"orientation)":[129],"being":[130],"disproportionately":[131],"amplified.As":[132],"become":[135],"ubiquitous,":[136],"our":[137],"highlight":[139],"need":[141],"for":[142],"holistic":[144],"approach-tracing":[145],"its":[148],"across":[150],"both":[151],"dimensions,":[155],"not":[156],"just":[157],"symptoms,":[158],"mitigate":[160],"harm.":[161]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-18T07:39:51.176621","created_date":"2025-11-08T00:00:00"}