{"id":"https://openalex.org/W3024936740","doi":"https://doi.org/10.48550/arxiv.2005.07647","title":"Finding Experts in Transformer Models","display_name":"Finding Experts in Transformer Models","publication_year":2020,"publication_date":"2020-05-15","ids":{"openalex":"https://openalex.org/W3024936740","doi":"https://doi.org/10.48550/arxiv.2005.07647","mag":"3024936740"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2005.07647","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2005.07647","pdf_url":"https://arxiv.org/pdf/2005.07647","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2005.07647","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5046208400","display_name":"Xavier Suau","orcid":"https://orcid.org/0000-0002-4052-7048","observed_orcids":["https://orcid.org/0000-0002-4052-7048"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Suau, Xavier","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043365442","display_name":"Luca Zappella","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zappella, Luca","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5071541814","display_name":"Nicholas Apostoloff","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Apostoloff, Nicholas","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":29,"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/T10028","display_name":"Topic Modeling","score":0.6934000253677368,"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":0.6934000253677368,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.2709999978542328,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.01489999983459711,"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/transformer","display_name":"transformer models","score":0.7390000224113464},{"id":"https://openalex.org/keywords/concept-classification","display_name":"concept classification","score":0.5},{"id":"https://openalex.org/keywords/generalization","display_name":"generalization","score":0.6779999732971191},{"id":"https://openalex.org/keywords/explainability","display_name":"explainability","score":0.8379999995231628},{"id":"https://openalex.org/keywords/self-conditioning","display_name":"self-conditioning","score":0.8349999785423279},{"id":"https://openalex.org/keywords/pre-trained-language-models","display_name":"pre-trained language models","score":0.8029999732971191}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7740198373794556},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6484348177909851},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6035932302474976},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5959872603416443},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5814337730407715},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.4733426868915558},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4486694931983948},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.43829068541526794},{"id":"https://openalex.org/C66905080","wikidata":"https://www.wikidata.org/wiki/Q17005494","display_name":"Binary classification","level":3,"score":0.43810051679611206},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.4378678798675537},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.11988908052444458},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.10385039448738098},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.08116558194160461},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07971367239952087},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","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},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C12267149","wikidata":"https://www.wikidata.org/wiki/Q282453","display_name":"Support vector machine","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2005.07647","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2005.07647","pdf_url":"https://arxiv.org/pdf/2005.07647","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2005.07647","doi":"https://doi.org/10.48550/arxiv.2005.07647","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2005.07647","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2005.07647","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2005.07647","pdf_url":"https://arxiv.org/pdf/2005.07647","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3024936740.pdf","grobid_xml":"https://content.openalex.org/works/W3024936740.grobid-xml"},"referenced_works_count":31,"referenced_works":["https://openalex.org/W2079182758","https://openalex.org/W2132339004","https://openalex.org/W2184268318","https://openalex.org/W2187089797","https://openalex.org/W2557449848","https://openalex.org/W2610018085","https://openalex.org/W2735642330","https://openalex.org/W2784070736","https://openalex.org/W2798812533","https://openalex.org/W2885396331","https://openalex.org/W2889056219","https://openalex.org/W2891612330","https://openalex.org/W2901107321","https://openalex.org/W2908854766","https://openalex.org/W2922565841","https://openalex.org/W2931212643","https://openalex.org/W2938704169","https://openalex.org/W2950813464","https://openalex.org/W2952984539","https://openalex.org/W2953337107","https://openalex.org/W2962776659","https://openalex.org/W2963310665","https://openalex.org/W2963323070","https://openalex.org/W2963341956","https://openalex.org/W2964204621","https://openalex.org/W2965373594","https://openalex.org/W2970049541","https://openalex.org/W2970242004","https://openalex.org/W2973049837","https://openalex.org/W2980282514","https://openalex.org/W2988660047"],"related_works":["https://openalex.org/W3201808024","https://openalex.org/W7164853426","https://openalex.org/W3205972749","https://openalex.org/W4415914267","https://openalex.org/W7161591390","https://openalex.org/W7124517177","https://openalex.org/W7160898125","https://openalex.org/W4417308017","https://openalex.org/W4390548145","https://openalex.org/W4410087697"],"abstract_inverted_index":{"In":[0],"this":[1],"work":[2],"we":[3,65],"study":[4],"the":[5,53,58,102,180,189],"presence":[6,94],"of":[7,50,69,95,105,116,139,158],"expert":[8,24,75,85,96],"units":[9,25,76,86,97],"in":[10,77,89],"pre-trained":[11,168],"Transformer":[12],"Models":[13],"(TM),":[14],"and":[15],"how":[16,127,164],"they":[17],"impact":[18],"a":[19,34,37,42,47,67,145,175],"model's":[20],"performance.":[21],"We":[22,82,119,162],"define":[23],"to":[26,32,79,125,133,148,165,171,183],"be":[27,80,132,154,184],"neurons":[28],"that":[29,72,84],"are":[30,87],"able":[31],"classify":[33],"concept":[35,43,54,150,177],"with":[36,101,174],"given":[38,176],"average":[39],"precision,":[40],"where":[41],"is":[44,98],"represented":[45],"by":[46,178],"binary":[48],"set":[49],"sentences":[51],"containing":[52],"(or":[55],"not).":[56],"Leveraging":[57],"OneSec":[59],"dataset":[60,68],"(Scarlini":[61],"et":[62],"al.,":[63],"2019),":[64],"compile":[66],"1641":[70],"concepts":[71,143],"allows":[73,108],"diverse":[74],"TM":[78,110],"discovered.":[81],"show":[83,163],"important":[88],"several":[90],"ways:":[91],"(1)":[92],"The":[93,137],"correlated":[99],"($r^2=0.833$)":[100],"generalization":[103],"power":[104],"TM,":[106],"which":[107,152],"ranking":[109],"without":[111,186],"requiring":[112,187],"fine-tuning":[113],"on":[114],"suites":[115],"downstream":[117],"tasks.":[118],"further":[120],"propose":[121],"an":[122],"empirical":[123],"method":[124],"decide":[126],"accurate":[128],"such":[129],"experts":[130,141,182],"should":[131],"evaluate":[134],"generalization.":[135],"(2)":[136],"overlap":[138],"top":[140,181],"between":[142],"provides":[144],"sensible":[146],"way":[147],"quantify":[149],"co-learning,":[151],"can":[153],"used":[155],"for":[156],"explainability":[157],"unknown":[159],"concepts.":[160],"(3)":[161],"self-condition":[166],"off-the-shelf":[167],"language":[169],"models":[170],"generate":[172],"text":[173],"forcing":[179],"active,":[185],"re-training":[188],"model":[190],"or":[191],"using":[192],"additional":[193],"parameters.":[194]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":6},{"year":2022,"cited_by_count":9},{"year":2021,"cited_by_count":5},{"year":2020,"cited_by_count":2}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2020-05-21T00:00:00"}