{"id":"https://openalex.org/W7140127682","doi":"https://doi.org/10.18653/v1/2026.eacl-long.117","title":"NeuronMoE: Efficient Cross-Lingual Extension via Neuron-Guided Mixture-of-Experts","display_name":"NeuronMoE: Efficient Cross-Lingual Extension via Neuron-Guided Mixture-of-Experts","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140127682","doi":"https://doi.org/10.18653/v1/2026.eacl-long.117"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2026.eacl-long.117","doi":"https://doi.org/10.18653/v1/2026.eacl-long.117","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.117","pdf_url":"https://aclanthology.org/2026.eacl-long.117.pdf","source":{"id":"https://openalex.org/S4306418011","display_name":"Conference of the European Chapter of the Association for Computational Linguistics","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2026.eacl-long.117.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128233966","display_name":"Rongzhi Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210110652","display_name":"RIKEN","ror":"https://ror.org/01sjwvz98","country_code":"JP","type":"facility","lineage":["https://openalex.org/I4210110652"]},{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Rongzhi Li","raw_affiliation_strings":["Riken","The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Riken","institution_ids":["https://openalex.org/I4210110652"]},{"raw_affiliation_string":"The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5126349274","display_name":"Hitomi Yanaka","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hitomi Yanaka","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.7486,"has_fulltext":true,"cited_by_count":2,"citation_normalized_percentile":{"value":0.91764408,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"2573","last_page":"2586"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.43380001187324524,"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":0.43380001187324524,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.41040000319480896,"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":0.12409999966621399,"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/mixture-of-experts","display_name":"mixture of experts","score":0.9340000152587891},{"id":"https://openalex.org/keywords/low-resource-languages","display_name":"low-resource languages","score":0.871999979019165},{"id":"https://openalex.org/keywords/multilingual-large-language-models","display_name":"multilingual large language models","score":0.7789999842643738},{"id":"https://openalex.org/keywords/parameter-efficiency","display_name":"parameter efficiency","score":0.781000018119812},{"id":"https://openalex.org/keywords/transformer-layers","display_name":"transformer layers","score":0.5479999780654907},{"id":"https://openalex.org/keywords/llama-3-2","display_name":"Llama 3.2","score":0.6679999828338623},{"id":"https://openalex.org/keywords/qwen","display_name":"Qwen","score":0.6740000247955322},{"id":"https://openalex.org/keywords/greeks","display_name":"Greeks","score":0.5379999876022339},{"id":"https://openalex.org/keywords/turkish-language","display_name":"Turkish language","score":0.9990000128746033},{"id":"https://openalex.org/keywords/hungarian-language","display_name":"Hungarian language","score":1.0}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.48730000853538513},{"id":"https://openalex.org/C2778029271","wikidata":"https://www.wikidata.org/wiki/Q5421931","display_name":"Extension (predicate logic)","level":2,"score":0.4302999973297119},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.36559998989105225},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.3264000117778778},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.325300008058548},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2558000087738037},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2556999921798706},{"id":"https://openalex.org/C25343380","wikidata":"https://www.wikidata.org/wiki/Q277521","display_name":"Relation (database)","level":2,"score":0.25130000710487366},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.24150000512599945},{"id":"https://openalex.org/C136119220","wikidata":"https://www.wikidata.org/wiki/Q1000660","display_name":"Algebra over a field","level":2,"score":0.2303999960422516}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/v1/2026.eacl-long.117","doi":"https://doi.org/10.18653/v1/2026.eacl-long.117","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.117","pdf_url":"https://aclanthology.org/2026.eacl-long.117.pdf","source":{"id":"https://openalex.org/S4306418011","display_name":"Conference of the European Chapter of the Association for Computational Linguistics","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2603.05046","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2603.05046","pdf_url":"https://arxiv.org/pdf/2603.05046","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.2603.05046","doi":"https://doi.org/10.48550/arxiv.2603.05046","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05046","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":"doi:10.18653/v1/2026.eacl-long.117","doi":"https://doi.org/10.18653/v1/2026.eacl-long.117","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.eacl-long.117","pdf_url":"https://aclanthology.org/2026.eacl-long.117.pdf","source":{"id":"https://openalex.org/S4306418011","display_name":"Conference of the European Chapter of the Association for Computational Linguistics","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320338075","display_name":"Core Research for Evolutional Science and Technology","ror":"https://ror.org/00097mb19"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140127682.pdf","grobid_xml":"https://content.openalex.org/works/W7140127682.grobid-xml"},"referenced_works_count":1,"referenced_works":["https://openalex.org/W4416035658"],"related_works":["https://openalex.org/W7134072965","https://openalex.org/W7161691122","https://openalex.org/W4411119569","https://openalex.org/W7125229514","https://openalex.org/W7151245152","https://openalex.org/W7162653711","https://openalex.org/W7166252858","https://openalex.org/W4414972183","https://openalex.org/W4412944946","https://openalex.org/W4385572336"],"abstract_inverted_index":{"Extending":[0],"large":[1],"language":[2,17,50,111],"models":[3,15,138],"to":[4,70,83],"lowresource":[5,110],"languages":[6,87],"is":[7,18],"essential":[8],"for":[9,85],"global":[10],"accessibility,":[11],"but":[12,30],"training":[13],"separate":[14],"per":[16,74],"prohibitively":[19],"expensive.Mixtureof-Experts":[20],"(MoE)":[21],"architectures":[22],"address":[23],"this":[24,92],"by":[25],"adding":[26],"sparse":[27],"language-specific":[28,64],"parameters,":[29],"determining":[31],"how":[32,136],"many":[33],"experts":[34,44,112],"each":[35],"layer":[36,75],"needs":[37],"remains":[38],"an":[39],"open":[40],"question.Current":[41],"approaches":[42],"allocate":[43],"based":[45,76],"on":[46,77,148],"layerlevel":[47],"similarity,":[48],"yet":[49],"processing":[51],"exhibits":[52],"fine-grained":[53],"specialization":[54,116],"at":[55],"individual":[56],"neurons.We":[57],"propose":[58],"NeuronMoE,":[59],"a":[60],"method":[61],"that":[62,109,152],"analyzes":[63],"neurons":[65],"across":[66,144],"all":[67],"transformer":[68],"components":[69],"guide":[71],"expert":[72,159],"allocation":[73,153],"empirically":[78],"measured":[79],"cross-lingual":[80],"neuron":[81,115],"diversity.Applied":[82],"Llama-3.2-3B":[84],"low-resource":[86],"(Greek,":[88],"Turkish,":[89],"and":[90,127,150],"Hungarian),":[91],"approach":[93,142],"achieves":[94],"approximately":[95],"40%":[96],"average":[97],"parameter":[98],"reduction":[99],"while":[100],"matching":[101],"the":[102,105,119],"performance":[103],"of":[104],"LayerMoE":[106],"baseline.We":[107],"find":[108],"independently":[113],"develop":[114],"patterns":[117],"mirroring":[118],"high-resource":[120],"language,":[121],"which":[122],"are":[123],"concentrated":[124],"in":[125,135],"early":[126],"late":[128],"layers.This":[129],"reveals":[130],"potential":[131],"universal":[132],"architectural":[133],"principles":[134],"multilingual":[137],"organize":[139],"linguistic":[140],"knowledge.Our":[141],"generalizes":[143],"architectures,":[145],"as":[146],"validated":[147],"Qwen,":[149],"shows":[151],"strategy":[154],"matters":[155],"more":[156],"than":[157],"total":[158],"count.":[160],"1":[161]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-03-24T00:00:00"}