{"id":"https://openalex.org/W4416033927","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.991","title":"Probing Semantic Routing in Large Mixture-of-Expert Models","display_name":"Probing Semantic Routing in Large Mixture-of-Expert Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416033927","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.991"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.991","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.991","pdf_url":"https://aclanthology.org/2025.findings-emnlp.991.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":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.991.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5064025350","display_name":"Matthew Olson","orcid":"https://orcid.org/0000-0002-4179-456X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Matthew Lyle Olson","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076851527","display_name":"Neale Ratzlaff","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Neale Ratzlaff","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5094579721","display_name":"Musashi Hinck","orcid":"https://orcid.org/0000-0001-8699-156X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Musashi Hinck","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027498859","display_name":"Man Luo","orcid":"https://orcid.org/0009-0004-8704-343X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Man Luo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055356130","display_name":"Sungduk Yu","orcid":"https://orcid.org/0000-0002-4506-3887"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sungduk Yu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103738915","display_name":"Chaogai Xue","orcid":"https://orcid.org/0009-0005-2633-2286"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chendi Xue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5087898808","display_name":"Vasudev Lal","orcid":"https://orcid.org/0000-0002-5907-9898"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vasudev Lal","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":7,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.3209994,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"18263","last_page":"18278"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.21449999511241913,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.21449999511241913,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/T10064","display_name":"Complex Network Analysis Techniques","score":0.11270000040531158,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13274","display_name":"Expert finding and Q&A systems","score":0.10480000078678131,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/routing","display_name":"Routing (electronic design automation)","score":0.38609999418258667},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.29649999737739563},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.29190000891685486},{"id":"https://openalex.org/keywords/identification","display_name":"Identification (biology)","score":0.2718999981880188}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.680400013923645},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.38609999418258667},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.33899998664855957},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3197999894618988},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.31949999928474426},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.29649999737739563},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.29190000891685486},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.27810001373291016},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.2718999981880188},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2687999904155731},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.26109999418258667}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.991","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.991","pdf_url":"https://aclanthology.org/2025.findings-emnlp.991.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"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.991","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.991","pdf_url":"https://aclanthology.org/2025.findings-emnlp.991.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":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416033927.pdf","grobid_xml":"https://content.openalex.org/works/W4416033927.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"In":[0],"the":[1,16,49,52,74,84],"past":[2],"year,":[3],"large":[4,43,107],"(>":[5],"100B":[6],"parameter)":[7],"mixture-of-expert":[8],"(MoE)":[9],"models":[10,45],"have":[11],"become":[12],"increasingly":[13],"common":[14],"in":[15,24,42,73,106],"open":[17],"domain.While":[18],"their":[19],"advantages":[20],"are":[21],"often":[22],"framed":[23],"terms":[25],"of":[26,51,103],"efficiency,":[27],"prior":[28],"work":[29],"has":[30],"also":[31],"explored":[32],"functional":[33],"differentiation":[34],"through":[35],"routing":[36,41,105],"behavior.We":[37],"investigate":[38],"whether":[39],"expert":[40,93],"MoE":[44,108],"is":[46],"influenced":[47],"by":[48],"semantics":[50],"inputs.To":[53],"test":[54],"this,":[55],"we":[56,61,79],"design":[57],"two":[58],"controlled":[59],"experiments.First,":[60],"compare":[62],"activations":[63],"on":[64],"sentence":[65],"pairs":[66],"with":[67,87],"a":[68],"shared":[69],"target":[70,85],"word":[71,86],"used":[72],"same":[75],"or":[76,90],"different":[77],"senses.Second,":[78],"fix":[80],"context":[81],"and":[82],"substitute":[83],"semantically":[88],"similar":[89],"dissimilar":[91],"alternatives.Comparing":[92],"overlap":[94],"across":[95],"these":[96],"conditions":[97],"reveals":[98],"clear,":[99],"statistically":[100],"significant":[101],"evidence":[102],"semantic":[104],"models.":[109]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-11-08T00:00:00"}