{"id":"https://openalex.org/W7140101800","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.338","title":"DeVisE: Towards the Behavioral Testing of Medical Large Language Models","display_name":"DeVisE: Towards the Behavioral Testing of Medical Large Language Models","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140101800","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.338"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.findings-eacl.338","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.338","pdf_url":"https://aclanthology.org/2026.findings-eacl.338.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: EACL 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-eacl.338.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5120057183","display_name":"Camila Zurdo Tagliabue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Camila Zurdo Tagliabue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5052637330","display_name":"Helo\u00edsa Oss Boll","orcid":"https://orcid.org/0000-0001-8121-2002"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Heloisa Oss Boll","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004736226","display_name":"Aykut Erdem","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Aykut Erdem","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130329548","display_name":"Erkut Erdem","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Erkut Erdem","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130403881","display_name":"Iacer Calixto","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Iacer Calixto","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":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"6427","last_page":"6441"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T13702","display_name":"Machine Learning in Healthcare","score":0.2630999982357025,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.2630999982357025,"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":0.23280000686645508,"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/T10028","display_name":"Topic Modeling","score":0.12319999933242798,"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/behavioral-analysis","display_name":"Behavioral analysis","score":0.3237999975681305},{"id":"https://openalex.org/keywords/natural-language","display_name":"Natural language","score":0.2777999937534332},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.2773999869823456},{"id":"https://openalex.org/keywords/test","display_name":"Test (biology)","score":0.25369998812675476}],"concepts":[{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.4787999987602234},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.46810001134872437},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.38440001010894775},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.34779998660087585},{"id":"https://openalex.org/C2989277270","wikidata":"https://www.wikidata.org/wiki/Q168338","display_name":"Behavioral analysis","level":2,"score":0.3237999975681305},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.321399986743927},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2773999869823456},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.25369998812675476},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.25049999356269836}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2026.findings-eacl.338","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.338","pdf_url":"https://aclanthology.org/2026.findings-eacl.338.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: EACL 2026","raw_type":"proceedings-article"},{"id":"pmh:oai:pure.amsterdamumc.nl:openaire/33bba462-faa4-4eb5-b2cb-4fc30e09df32","is_oa":true,"landing_page_url":"https://pure.amsterdamumc.nl/en/publications/33bba462-faa4-4eb5-b2cb-4fc30e09df32","pdf_url":null,"source":{"id":"https://openalex.org/S7407055222","display_name":"Pure Amsterdam UMC","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":null,"raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-eacl.338","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.338","pdf_url":"https://aclanthology.org/2026.findings-eacl.338.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: EACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.6050875186920166,"id":"https://metadata.un.org/sdg/4"}],"awards":[{"id":"https://openalex.org/G44630794","display_name":null,"funder_award_id":"2247-A","funder_id":"https://openalex.org/F4320322626","funder_display_name":"T\u00fcrkiye Bilimsel ve Teknolojik Ara\u015ft\u0131rma Kurumu"},{"id":"https://openalex.org/G5336964267","display_name":"CaRe-NLP: Human-Centric and Responsible NLP methods for Dutch healthcare","funder_award_id":"NGF.1607.22.014","funder_id":"https://openalex.org/F4320321800","funder_display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek"}],"funders":[{"id":"https://openalex.org/F4320321800","display_name":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek","ror":"https://ror.org/04jsz6e67"},{"id":"https://openalex.org/F4320322626","display_name":"T\u00fcrkiye Bilimsel ve Teknolojik Ara\u015ft\u0131rma Kurumu","ror":"https://ror.org/04w9kkr77"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140101800.pdf","grobid_xml":"https://content.openalex.org/works/W7140101800.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,125],"(LLMs)":[3],"are":[4],"increasingly":[5],"applied":[6],"in":[7,65,121,128],"clinical":[8,40],"decision":[9],"support,":[10],"yet":[11],"current":[12],"evaluations":[13],"rarely":[14],"reveal":[15],"whether":[16],"their":[17,102],"outputs":[18],"reflect":[19],"genuine":[20],"medical":[21,80],"reasoning":[22],"or":[23],"superficial":[24],"correlations.We":[25],"introduce":[26],"DeVisE":[27],"(Demographics":[28],"and":[29,58,70,79,97,108,131],"Vital":[30],"signs":[31],"Evaluation),":[32],"a":[33],"behavioral":[34],"testing":[35],"framework":[36],"that":[37,113],"probes":[38],"finegrained":[39],"understanding":[41],"through":[42,88],"controlled":[43],"counterfactuals.Using":[44],"intensive":[45],"care":[46],"unit":[47],"(ICU)":[48],"discharge":[49],"notes":[50],"from":[51],"MIMIC-IV,":[52],"we":[53],"construct":[54],"both":[55],"raw":[56],"(real-world)":[57],"templatebased":[59],"(synthetic)":[60],"variants":[61],"with":[62,124],"single-variable":[63],"perturbations":[64],"demographic":[66],"(age,":[67],"gender,":[68],"ethnicity)":[69],"vital":[71],"sign":[72],"attributes.We":[73],"evaluate":[74],"eight":[75],"LLMs,":[76],"spanning":[77],"general-purpose":[78],"variants,":[81],"under":[82],"zero-shot":[83],"setting.Model":[84],"behavior":[85],"is":[86],"analyzed":[87],"(1)":[89],"inputlevel":[90],"sensitivity,":[91],"capturing":[92],"how":[93,129,153],"counterfactuals":[94,142,151],"alter":[95],"perplexity,":[96],"(2)":[98],"downstream":[99,148],"reasoning,":[100],"measuring":[101],"effect":[103],"on":[104],"predicted":[105],"ICU":[106],"lengthof-stay":[107],"mortality.Overall,":[109],"our":[110],"results":[111],"show":[112],"standard":[114],"task":[115],"metrics":[116],"obscure":[117],"clinically":[118],"relevant":[119],"differences":[120],"model":[122],"behavior,":[123],"differing":[126],"substantially":[127],"consistently":[130],"proportionally":[132],"they":[133],"adjust":[134],"predictions":[135],"to":[136],"counterfactual":[137],"perturbations.":[138],"1":[139],"How":[140],"do":[141,150],"change":[143,152],"the":[144,155],"probability":[145],"distribution":[146],"of":[147],"tasks?How":[149],"likely":[154],"patient":[156],"is?OpenBioLLM":[157],"70B":[158,161],"LLaMA-3.3-Instruct70B":[159],"DeepSeek-R1-Distill":[160],"Qwen-2.5-Instruct72B":[162],"GPT-OSS":[163],"120B":[164],"GPT-4.1-mini?":[165]},"counts_by_year":[],"updated_date":"2026-07-19T07:52:34.831488","created_date":"2026-03-24T00:00:00"}