{"id":"https://openalex.org/W7139911608","doi":"https://doi.org/10.48550/arxiv.2603.19149","title":"Optimal Splitting of Language Models from Mixtures to Specialized Domains","display_name":"Optimal Splitting of Language Models from Mixtures to Specialized Domains","publication_year":2026,"publication_date":"2026-03-19","ids":{"openalex":"https://openalex.org/W7139911608","doi":"https://doi.org/10.48550/arxiv.2603.19149"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.19149","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19149","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.19149","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5059839283","display_name":"Skyler Seto","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Seto, Skyler","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130225286","display_name":"Pierre Ablin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ablin, Pierre","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130251284","display_name":"Anastasiia Filippova","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Filippova, Anastasiia","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130217843","display_name":"Jiayuan Ye","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ye, Jiayuan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130238289","display_name":"Louis Bethune","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bethune, Louis","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031829458","display_name":"Angelos Katharopoulos","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Katharopoulos, Angelos","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5065912572","display_name":"David Grangier","orcid":"https://orcid.org/0000-0002-8847-9532"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Grangier, David","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5059839283"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"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":0.35030001401901245,"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":0.35030001401901245,"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":0.16529999673366547,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.11999999731779099,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/language-model","display_name":"Language model","score":0.7139000296592712},{"id":"https://openalex.org/keywords/variety","display_name":"Variety (cybernetics)","score":0.651199996471405},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.46650001406669617},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.421099990606308},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.30239999294281006},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.2904999852180481}],"concepts":[{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.7139000296592712},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6633999943733215},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.651199996471405},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5575000047683716},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.46810001134872437},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.46650001406669617},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.42149999737739563},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.421099990606308},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.30239999294281006},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2904999852180481},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.28850001096725464},{"id":"https://openalex.org/C2775957009","wikidata":"https://www.wikidata.org/wiki/Q399495","display_name":"Standard language","level":2,"score":0.2734000086784363},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.27070000767707825},{"id":"https://openalex.org/C76969082","wikidata":"https://www.wikidata.org/wiki/Q486902","display_name":"Mathematical model","level":2,"score":0.2696000039577484},{"id":"https://openalex.org/C2781316041","wikidata":"https://www.wikidata.org/wiki/Q1230584","display_name":"Diversity (politics)","level":2,"score":0.25839999318122864}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.19149","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19149","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.19149","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.19149","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Language":[0],"models":[1,65,83],"achieve":[2],"impressive":[3],"performance":[4,141],"on":[5,34,43,66],"a":[6,29,44,78,86,111],"variety":[7],"of":[8,20,38,46,63,110,113,131],"knowledge,":[9],"language,":[10],"and":[11,18,90,98,119,123,129,147,154],"reasoning":[12,148],"tasks":[13],"due":[14],"to":[15,71,125,134],"the":[16,35,52,56,92,108],"scale":[17],"diversity":[19],"pretraining":[21,32,62,81,88,97,100,118],"data":[22,39,50],"available.":[23],"The":[24],"standard":[25],"training":[26],"recipe":[27],"is":[28],"two-stage":[30],"paradigm:":[31],"first":[33],"full":[36,53],"corpus":[37],"followed":[40],"by":[41],"specialization":[42,121],"subset":[45],"high":[47],"quality,":[48],"specialized":[49,68],"from":[51],"corpus.":[54],"In":[55],"multi-domain":[57],"setting,":[58],"this":[59],"involves":[60],"continued":[61,99],"multiple":[64,82],"each":[67],"domain,":[69],"referred":[70],"as":[72],"split":[73],"model":[74,112,127,136,152],"training.":[75],"We":[76],"propose":[77],"method":[79],"for":[80],"independently":[84],"over":[85],"general":[87],"corpus,":[89],"determining":[91],"optimal":[93],"compute":[94,155],"allocation":[95],"between":[96],"using":[101],"scaling":[102],"laws.":[103],"Our":[104],"approach":[105,139],"accurately":[106],"predicts":[107],"loss":[109],"size":[114],"N":[115],"with":[116],"D":[117],"D'":[120],"tokens,":[122],"extrapolates":[124],"larger":[126],"sizes":[128,153],"number":[130],"tokens.":[132],"Applied":[133],"language":[135],"training,":[137],"our":[138],"improves":[140],"consistently":[142],"across":[143,150],"common":[144],"sense":[145],"knowledge":[146],"benchmarks":[149],"different":[151],"budgets.":[156]},"counts_by_year":[],"updated_date":"2026-03-21T06:36:02.116451","created_date":"2026-03-21T00:00:00"}