{"id":"https://openalex.org/W7135211585","doi":"https://doi.org/10.48550/arxiv.2603.11881","title":"Bielik-Minitron-7B: Compressing Large Language Models via Structured Pruning and Knowledge Distillation for the Polish Language","display_name":"Bielik-Minitron-7B: Compressing Large Language Models via Structured Pruning and Knowledge Distillation for the Polish Language","publication_year":2026,"publication_date":"2026-03-12","ids":{"openalex":"https://openalex.org/W7135211585","doi":"https://doi.org/10.48550/arxiv.2603.11881"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.11881","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11881","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2603.11881","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128987726","display_name":"Remigiusz Kinas","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Kinas, Remigiusz","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129087589","display_name":"Pawe\u0142 Kiszczak","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kiszczak, Pawe\u0142","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057207205","display_name":"Sergio P. Perez","orcid":"https://orcid.org/0000-0001-8485-609X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Perez, Sergio P.","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128955900","display_name":"Krzysztof Ociepa","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ociepa, Krzysztof","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128965847","display_name":"\u0141ukasz Flis","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Flis, \u0141ukasz","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128934725","display_name":"Krzysztof Wr\u00f3bel","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wr\u00f3bel, Krzysztof","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5114618217","display_name":"Adrian Gwo\u017adziej","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gwo\u017adziej, Adrian","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5128987726"],"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/T10181","display_name":"Natural Language Processing Techniques","score":0.5598000288009644,"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":0.5598000288009644,"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/T10028","display_name":"Topic Modeling","score":0.08380000293254852,"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/T14347","display_name":"Big Data and Digital Economy","score":0.043800000101327896,"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/pruning","display_name":"Pruning","score":0.800599992275238},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.6913999915122986},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6065000295639038},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5954999923706055},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.5745999813079834},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.44690001010894775},{"id":"https://openalex.org/keywords/baseline","display_name":"Baseline (sea)","score":0.39660000801086426}],"concepts":[{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.800599992275238},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7976999878883362},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6913999915122986},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.614799976348877},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6065000295639038},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5954999923706055},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.5745999813079834},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5379999876022339},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.44690001010894775},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.39660000801086426},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.3589000105857849},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.35670000314712524},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.35359999537467957},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.3319000005722046},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.3292999863624573},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.28110000491142273},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.27790001034736633},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C179603123","wikidata":"https://www.wikidata.org/wiki/Q1941921","display_name":"Modeling language","level":3,"score":0.2685999870300293}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.11881","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11881","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.11881","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.11881","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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":{"This":[0,114],"report":[1],"details":[2],"the":[3,13,29,43,55,63,75,103,128],"creation":[4],"of":[5,12,83,102],"Bielik-Minitron-7B,":[6],"a":[7,23,78],"compressed":[8],"7.35B":[9],"parameter":[10,45],"version":[11],"Bielik-11B-v3.0":[14],"model,":[15],"specifically":[16],"optimized":[17],"for":[18,59,67,70,124],"European":[19],"languages.":[20],"By":[21],"leveraging":[22],"two-stage":[24],"compression":[25],"methodology":[26],"inspired":[27],"by":[28,47],"NVIDIA":[30,56,64],"Minitron":[31],"approach,":[32],"we":[33],"combined":[34],"structured":[35],"hybrid":[36],"pruning":[37,61],"and":[38,62,91],"knowledge":[39],"distillation":[40,69],"to":[41,51,110,120],"reduce":[42],"model's":[44,105],"count":[46],"33.4%,":[48],"from":[49],"11.04B":[50],"7.35B.":[52],"We":[53],"utilized":[54],"Model":[57],"Optimizer":[58],"structural":[60],"NeMo":[65],"Framework":[66],"logit-based":[68],"quality":[71,131],"recovery.":[72],"Following":[73],"distillation,":[74],"model":[76,97,130],"underwent":[77],"rigorous":[79],"alignment":[80],"pipeline":[81],"consisting":[82],"Supervised":[84],"Fine-Tuning":[85],"(SFT),":[86],"Direct":[87],"Preference":[88],"Optimization":[89],"(DPO-P),":[90],"Reinforcement":[92],"Learning":[93],"(GRPO).":[94],"Our":[95],"final":[96],"successfully":[98],"recovered":[99],"approximately":[100],"90%":[101],"baseline":[104],"performance":[106],"while":[107,132],"providing":[108],"up":[109],"50%":[111],"inference":[112,134],"speedup.":[113],"approach":[115],"demonstrates":[116],"an":[117],"efficient":[118],"pathway":[119],"create":[121],"language":[122],"models":[123],"less-represented":[125],"languages,":[126],"preserving":[127],"original":[129],"reducing":[133],"deployment":[135],"costs.":[136]},"counts_by_year":[],"updated_date":"2026-03-14T06:46:50.379900","created_date":"2026-03-14T00:00:00"}