{"id":"https://openalex.org/W7155075127","doi":"https://doi.org/10.1109/synasc69064.2025.00031","title":"LLMic: Building a Romanian Foundation Language Model","display_name":"LLMic: Building a Romanian Foundation Language Model","publication_year":2025,"publication_date":"2025-09-22","ids":{"openalex":"https://openalex.org/W7155075127","doi":"https://doi.org/10.1109/synasc69064.2025.00031"},"language":null,"primary_location":{"id":"doi:10.1109/synasc69064.2025.00031","is_oa":false,"landing_page_url":"https://doi.org/10.1109/synasc69064.2025.00031","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5088685497","display_name":"Vlad-Andrei B\u0103doiu","orcid":"https://orcid.org/0009-0000-2890-7980"},"institutions":[{"id":"https://openalex.org/I61641377","display_name":"Universitatea Na\u021bional\u0103 de \u0218tiin\u021b\u0103 \u0219i Tehnologie Politehnica Bucure\u0219ti","ror":"https://ror.org/0558j5q12","country_code":"RO","type":"education","lineage":["https://openalex.org/I61641377"]}],"countries":["RO"],"is_corresponding":false,"raw_author_name":"Vlad-Andrei B\u0103doiu","raw_affiliation_strings":["University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania","institution_ids":["https://openalex.org/I61641377"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106953914","display_name":"Alexandru M. Gherghescu","orcid":null},"institutions":[{"id":"https://openalex.org/I61641377","display_name":"Universitatea Na\u021bional\u0103 de \u0218tiin\u021b\u0103 \u0219i Tehnologie Politehnica Bucure\u0219ti","ror":"https://ror.org/0558j5q12","country_code":"RO","type":"education","lineage":["https://openalex.org/I61641377"]}],"countries":["RO"],"is_corresponding":false,"raw_author_name":"Alexandru M. Gherghescu","raw_affiliation_strings":["University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania","institution_ids":["https://openalex.org/I61641377"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000170074","display_name":"Alexandru Agache","orcid":null},"institutions":[{"id":"https://openalex.org/I61641377","display_name":"Universitatea Na\u021bional\u0103 de \u0218tiin\u021b\u0103 \u0219i Tehnologie Politehnica Bucure\u0219ti","ror":"https://ror.org/0558j5q12","country_code":"RO","type":"education","lineage":["https://openalex.org/I61641377"]}],"countries":["RO"],"is_corresponding":false,"raw_author_name":"Alexandru Agache","raw_affiliation_strings":["University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania","institution_ids":["https://openalex.org/I61641377"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5099362896","display_name":"Mihai-Valentin Dumitru","orcid":"https://orcid.org/0009-0009-3202-6110"},"institutions":[{"id":"https://openalex.org/I61641377","display_name":"Universitatea Na\u021bional\u0103 de \u0218tiin\u021b\u0103 \u0219i Tehnologie Politehnica Bucure\u0219ti","ror":"https://ror.org/0558j5q12","country_code":"RO","type":"education","lineage":["https://openalex.org/I61641377"]}],"countries":["RO"],"is_corresponding":false,"raw_author_name":"Mihai-Valentin Dumitru","raw_affiliation_strings":["University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania","institution_ids":["https://openalex.org/I61641377"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5008493066","display_name":"Costin Raiciu","orcid":"https://orcid.org/0000-0002-5937-2162"},"institutions":[{"id":"https://openalex.org/I61641377","display_name":"Universitatea Na\u021bional\u0103 de \u0218tiin\u021b\u0103 \u0219i Tehnologie Politehnica Bucure\u0219ti","ror":"https://ror.org/0558j5q12","country_code":"RO","type":"education","lineage":["https://openalex.org/I61641377"]}],"countries":["RO"],"is_corresponding":false,"raw_author_name":"Costin Raiciu","raw_affiliation_strings":["University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University Politehnica of Bucharest,Computer Science and Engineering,Bucharest,Romania","institution_ids":["https://openalex.org/I61641377"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I61641377"],"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":"168","last_page":"175"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.4124000072479248,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.4124000072479248,"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/T13629","display_name":"Text Readability and Simplification","score":0.03480000048875809,"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/T13523","display_name":"Mathematics, Computing, and Information Processing","score":0.026900000870227814,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/foundation","display_name":"Foundation (evidence)","score":0.5662999749183655},{"id":"https://openalex.org/keywords/romanian","display_name":"Romanian","score":0.5569000244140625},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.32580000162124634},{"id":"https://openalex.org/keywords/on-language","display_name":"On Language","score":0.2727999985218048},{"id":"https://openalex.org/keywords/language-industry","display_name":"Language industry","score":0.26190000772476196}],"concepts":[{"id":"https://openalex.org/C2780966255","wikidata":"https://www.wikidata.org/wiki/Q5474306","display_name":"Foundation (evidence)","level":2,"score":0.5662999749183655},{"id":"https://openalex.org/C129400051","wikidata":"https://www.wikidata.org/wiki/Q7913","display_name":"Romanian","level":2,"score":0.5569000244140625},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.39070001244544983},{"id":"https://openalex.org/C55587333","wikidata":"https://www.wikidata.org/wiki/Q1133029","display_name":"Engineering ethics","level":1,"score":0.35499998927116394},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.352400004863739},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.32580000162124634},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.31610000133514404},{"id":"https://openalex.org/C144024400","wikidata":"https://www.wikidata.org/wiki/Q21201","display_name":"Sociology","level":0,"score":0.2768999934196472},{"id":"https://openalex.org/C2779313563","wikidata":"https://www.wikidata.org/wiki/Q17072565","display_name":"On Language","level":2,"score":0.2727999985218048},{"id":"https://openalex.org/C1813318","wikidata":"https://www.wikidata.org/wiki/Q2465887","display_name":"Language industry","level":4,"score":0.26190000772476196},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2590000033378601},{"id":"https://openalex.org/C9652623","wikidata":"https://www.wikidata.org/wiki/Q190109","display_name":"Field (mathematics)","level":2,"score":0.2547999918460846}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/synasc69064.2025.00031","is_oa":false,"landing_page_url":"https://doi.org/10.1109/synasc69064.2025.00031","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F1100377486","display_name":"European High Performance Computing Joint Undertaking","ror":null},{"id":"https://openalex.org/F4320316785","display_name":"VMware","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":22,"referenced_works":["https://openalex.org/W2512924740","https://openalex.org/W2962834855","https://openalex.org/W2979826702","https://openalex.org/W3001434439","https://openalex.org/W3002330681","https://openalex.org/W3035016936","https://openalex.org/W3173954987","https://openalex.org/W3212156874","https://openalex.org/W3213241618","https://openalex.org/W4229679462","https://openalex.org/W4285225959","https://openalex.org/W4386768656","https://openalex.org/W4388979610","https://openalex.org/W4389518760","https://openalex.org/W4402671236","https://openalex.org/W4402671286","https://openalex.org/W4402671717","https://openalex.org/W4404792848","https://openalex.org/W4413169625","https://openalex.org/W4415795709","https://openalex.org/W4415798387","https://openalex.org/W7133204064"],"related_works":[],"abstract_inverted_index":{"Recent":[0],"advances":[1],"in":[2,54,58,62,153,171],"Large":[3],"Language":[4],"Models":[5],"(LLMs)":[6],"have":[7],"shown":[8],"remarkable":[9],"capabilities":[10,61],"across":[11],"various":[12],"tasks,":[13,125],"with":[14],"commercial":[15],"models":[16,22,122],"leading":[17],"the":[18,49,70,146,154,165,168,172],"way.":[19],"While":[20],"open":[21,55,133],"usually":[23],"operate":[24],"at":[25],"smaller":[26],"scales":[27],"due":[28],"to":[29,119,129,163],"constraints":[30],"on":[31],"available":[32],"corpora":[33],"and":[34,42,88,134],"hardware":[35],"resources,":[36],"they":[37],"maintain":[38],"competitiveness":[39],"through":[40,160],"specialization":[41],"fine-tuning.":[43],"However,":[44],"a":[45,75,80,99,107],"significant":[46],"challenge":[47],"persists:":[48],"under-representation":[50],"of":[51,73,93,167,174],"low-resource":[52,81],"languages":[53],"datasets":[56],"results":[57,127],"weak":[59],"model":[60,77],"many":[63],"languages.":[64],"In":[65],"this":[66,94,161],"paper,":[67],"we":[68,96],"document":[69],"complete":[71],"process":[72],"pretraining":[74,148],"foundation":[76],"for":[78,123,142],"Romanian,":[79],"language,":[82],"including":[83],"corpus":[84,102],"construction,":[85],"architecture":[86],"selection,":[87],"hyperparameter":[89],"optimization.":[90],"As":[91],"part":[92],"work,":[95],"introduce":[97],"FuLG,":[98],"hundred-fifty-billion-token":[100],"Romanian":[101,169],"extracted":[103],"from":[104],"CommonCrawl,":[105],"alongside":[106],"3-billion-parameter":[108],"bilingual":[109],"model,":[110],"LLMic.":[111],"Our":[112],"evaluation":[113],"shows":[114],"that":[115,139],"it":[116],"is":[117],"worthwhile":[118],"train":[120],"language-specific":[121],"specialized":[124],"achieving":[126],"comparable":[128],"other":[130],"much":[131],"larger":[132],"closed":[135],"models.":[136],"We":[137,158],"show":[138],"fine-tuning":[140],"LLMic":[141],"language":[143,170],"translation":[144,156],"after":[145],"initial":[147],"phase":[149],"outperforms":[150],"existing":[151],"solutions":[152],"English-to-Romanian":[155],"task.":[157],"hope":[159],"work":[162],"advance":[164],"standing":[166],"world":[173],"LLMs.":[175]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-04-22T00:00:00"}
