{"id":"https://openalex.org/W7148522013","doi":"https://doi.org/10.1109/asru65441.2025.11434665","title":"Evaluation of LLMs in Speech is Often Flawed: Test Set Contamination in Large Language Models for Speech Recognition","display_name":"Evaluation of LLMs in Speech is Often Flawed: Test Set Contamination in Large Language Models for Speech Recognition","publication_year":2025,"publication_date":"2025-12-06","ids":{"openalex":"https://openalex.org/W7148522013","doi":"https://doi.org/10.1109/asru65441.2025.11434665"},"language":null,"primary_location":{"id":"doi:10.1109/asru65441.2025.11434665","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru65441.2025.11434665","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","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/A5132813852","display_name":"Yuan Tseng","orcid":null},"institutions":[{"id":"https://openalex.org/I4210117523","display_name":"Samsung (United Kingdom)","ror":"https://ror.org/01w6gjq94","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210117523"]}],"countries":["GB"],"is_corresponding":true,"raw_author_name":"Yuan Tseng","raw_affiliation_strings":["AI Center Cambridge, Samsung,United Kingdom"],"affiliations":[{"raw_affiliation_string":"AI Center Cambridge, Samsung,United Kingdom","institution_ids":["https://openalex.org/I4210117523"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089505434","display_name":"Titouan Parcollet","orcid":"https://orcid.org/0000-0003-0672-1346"},"institutions":[{"id":"https://openalex.org/I4210117523","display_name":"Samsung (United Kingdom)","ror":"https://ror.org/01w6gjq94","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210117523"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Titouan Parcollet","raw_affiliation_strings":["AI Center Cambridge, Samsung,United Kingdom"],"affiliations":[{"raw_affiliation_string":"AI Center Cambridge, Samsung,United Kingdom","institution_ids":["https://openalex.org/I4210117523"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132807857","display_name":"Rogier Van Dalen","orcid":null},"institutions":[{"id":"https://openalex.org/I4210117523","display_name":"Samsung (United Kingdom)","ror":"https://ror.org/01w6gjq94","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210117523"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Rogier Van Dalen","raw_affiliation_strings":["AI Center Cambridge, Samsung,United Kingdom"],"affiliations":[{"raw_affiliation_string":"AI Center Cambridge, Samsung,United Kingdom","institution_ids":["https://openalex.org/I4210117523"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003030422","display_name":"Shucong Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210117523","display_name":"Samsung (United Kingdom)","ror":"https://ror.org/01w6gjq94","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210117523"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Shucong Zhang","raw_affiliation_strings":["AI Center Cambridge, Samsung,United Kingdom"],"affiliations":[{"raw_affiliation_string":"AI Center Cambridge, Samsung,United Kingdom","institution_ids":["https://openalex.org/I4210117523"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5087701164","display_name":"Sourav Bhattacharya","orcid":"https://orcid.org/0000-0001-9670-5264"},"institutions":[{"id":"https://openalex.org/I4210117523","display_name":"Samsung (United Kingdom)","ror":"https://ror.org/01w6gjq94","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250650973","https://openalex.org/I4210117523"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Sourav Bhattacharya","raw_affiliation_strings":["AI Center Cambridge, Samsung,United Kingdom"],"affiliations":[{"raw_affiliation_string":"AI Center Cambridge, Samsung,United Kingdom","institution_ids":["https://openalex.org/I4210117523"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5132813852"],"corresponding_institution_ids":["https://openalex.org/I4210117523"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.87574262,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"8"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.8294000029563904,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.8294000029563904,"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/T12380","display_name":"Authorship Attribution and Profiling","score":0.0471000000834465,"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/T10028","display_name":"Topic Modeling","score":0.021299999207258224,"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/set","display_name":"Set (abstract data type)","score":0.6026999950408936},{"id":"https://openalex.org/keywords/reliability","display_name":"Reliability (semiconductor)","score":0.5774999856948853},{"id":"https://openalex.org/keywords/test","display_name":"Test (biology)","score":0.4927999973297119},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.45509999990463257},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3612000048160553},{"id":"https://openalex.org/keywords/measure","display_name":"Measure (data warehouse)","score":0.34689998626708984}],"concepts":[{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6026999950408936},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.5774999856948853},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.569599986076355},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5317000150680542},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.4927999973297119},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.45509999990463257},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.4212000072002411},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.39160001277923584},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3612000048160553},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.34689998626708984},{"id":"https://openalex.org/C522180918","wikidata":"https://www.wikidata.org/wiki/Q18342738","display_name":"Public speaking","level":2,"score":0.3343999981880188},{"id":"https://openalex.org/C169903167","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Test set","level":2,"score":0.31450000405311584},{"id":"https://openalex.org/C204201278","wikidata":"https://www.wikidata.org/wiki/Q1332614","display_name":"Voice activity detection","level":3,"score":0.31119999289512634},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.3102000057697296},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.30090001225471497},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.299699991941452},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.29510000348091125},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.2937000095844269},{"id":"https://openalex.org/C112758219","wikidata":"https://www.wikidata.org/wiki/Q16038819","display_name":"Duration (music)","level":2,"score":0.27059999108314514}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/asru65441.2025.11434665","is_oa":false,"landing_page_url":"https://doi.org/10.1109/asru65441.2025.11434665","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"score":0.6370218396186829,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":20,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W2936774411","https://openalex.org/W3209984917","https://openalex.org/W4285107714","https://openalex.org/W4285225959","https://openalex.org/W4300601563","https://openalex.org/W4389518953","https://openalex.org/W4391021505","https://openalex.org/W4391021666","https://openalex.org/W4392903956","https://openalex.org/W4392931626","https://openalex.org/W4402111222","https://openalex.org/W4402112329","https://openalex.org/W4402112335","https://openalex.org/W4402671236","https://openalex.org/W4402671286","https://openalex.org/W4404781855","https://openalex.org/W4406461293","https://openalex.org/W4408345654","https://openalex.org/W4411630291"],"related_works":[],"abstract_inverted_index":{"Recent":[0],"work":[1,33],"suggests":[2],"that":[3,35,124],"large":[4],"language":[5],"models":[6],"(LLMs)":[7],"can":[8,127],"improve":[9],"performance":[10],"of":[11,39,59,133,139],"speech":[12,92,142],"tasks":[13],"compared":[14],"to":[15,82,116],"existing":[16],"systems.":[17],"To":[18,66],"support":[19],"their":[20],"claims,":[21],"results":[22],"on":[23,95],"LibriSpeech":[24,41],"and":[25,42],"Common":[26,43],"Voice":[27,44],"are":[28,74,97],"often":[29],"quoted.":[30],"However,":[31],"this":[32],"finds":[34],"a":[36],"substantial":[37],"amount":[38],"the":[40,57,107,137],"evaluation":[45],"sets":[46],"appear":[47],"in":[48],"public":[49],"LLM":[50,78,108,120,125],"pretraining":[51],"corpora.":[52],"This":[53],"calls":[54],"into":[55],"question":[56],"reliability":[58],"findings":[60],"drawn":[61],"from":[62],"these":[63],"two":[64],"datasets.":[65],"measure":[67],"contamination":[68,73],"impact,":[69],"LLMs":[70,96],"trained":[71],"with/without":[72],"compared.":[75,98],"A":[76],"contaminated":[77],"is":[79,109],"more":[80],"likely":[81],"generate":[83],"test":[84],"sentences":[85],"it":[86],"has":[87],"seen":[88,118],"during":[89,119],"training.":[90,121],"Then,":[91],"recognisers":[93],"based":[94],"They":[99],"show":[100,123],"only":[101],"subtle":[102],"error":[103],"rate":[104],"differences":[105],"if":[106],"contaminated,":[110],"but":[111],"assign":[112],"significantly":[113],"higher":[114],"probabilities":[115],"transcriptions":[117],"Results":[122],"outputs":[126],"be":[128],"biased":[129],"by":[130],"tiny":[131],"amounts":[132],"data":[134],"contamination,":[135],"highlighting":[136],"importance":[138],"evaluating":[140],"LLM-based":[141],"systems":[143],"with":[144],"held-out":[145],"data.":[146]},"counts_by_year":[],"updated_date":"2026-04-03T16:44:17.987007","created_date":"2026-04-03T00:00:00"}