{"id":"https://openalex.org/W7140102726","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.34","title":"Complexity-aware fine-tuning","display_name":"Complexity-aware fine-tuning","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7140102726","doi":"https://doi.org/10.18653/v1/2026.findings-eacl.34"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-eacl.34","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.34","pdf_url":"https://aclanthology.org/2026.findings-eacl.34.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: EACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-eacl.34.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Andrey Goncharov","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Andrey Goncharov","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Daniil Vyazhev","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Daniil Vyazhev","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Petr Sychev","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Petr Sychev","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Edvard Khalafyan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Edvard Khalafyan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Alexey Zaytsev","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alexey Zaytsev","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"682","last_page":"696"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.06710000336170197,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.06710000336170197,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10720","display_name":"Complexity and Algorithms in Graphs","score":0.04019999876618385,"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"}},{"id":"https://openalex.org/T10142","display_name":"Formal Methods in Verification","score":0.0364999994635582,"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/language-model","display_name":"Language model","score":0.4846999943256378},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.4830000102519989},{"id":"https://openalex.org/keywords/entropy","display_name":"Entropy (arrow of time)","score":0.40310001373291016},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.35499998927116394},{"id":"https://openalex.org/keywords/phrase","display_name":"Phrase","score":0.31220000982284546},{"id":"https://openalex.org/keywords/principle-of-maximum-entropy","display_name":"Principle of maximum entropy","score":0.3102000057697296},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.30550000071525574},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.29190000891685486}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7031000256538391},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4846999943256378},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.4830000102519989},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44279998540878296},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.42739999294281006},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.40310001373291016},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.36059999465942383},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.35499998927116394},{"id":"https://openalex.org/C2776224158","wikidata":"https://www.wikidata.org/wiki/Q187931","display_name":"Phrase","level":2,"score":0.31220000982284546},{"id":"https://openalex.org/C9679016","wikidata":"https://www.wikidata.org/wiki/Q1417473","display_name":"Principle of maximum entropy","level":2,"score":0.3102000057697296},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.30550000071525574},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.29190000891685486},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.28929999470710754},{"id":"https://openalex.org/C155911762","wikidata":"https://www.wikidata.org/wiki/Q422321","display_name":"Blueprint","level":2,"score":0.28610000014305115},{"id":"https://openalex.org/C41458344","wikidata":"https://www.wikidata.org/wiki/Q732577","display_name":"Publication","level":2,"score":0.27730000019073486},{"id":"https://openalex.org/C3020493868","wikidata":"https://www.wikidata.org/wiki/Q55631277","display_name":"Real world data","level":2,"score":0.266400009393692},{"id":"https://openalex.org/C84525736","wikidata":"https://www.wikidata.org/wiki/Q831366","display_name":"Decision tree","level":2,"score":0.2628999948501587},{"id":"https://openalex.org/C2993776861","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Open domain","level":3,"score":0.25769999623298645},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.25589999556541443},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.25360000133514404},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.25}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-eacl.34","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.34","pdf_url":"https://aclanthology.org/2026.findings-eacl.34.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: EACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-eacl.34","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-eacl.34","pdf_url":"https://aclanthology.org/2026.findings-eacl.34.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: EACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7140102726.pdf","grobid_xml":"https://content.openalex.org/works/W7140102726.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"General-purpose":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4,89],"are":[5],"frequently":[6],"fine-tuned":[7],"through":[8],"supervised":[9],"fine-tuning":[10],"(SFT)":[11],"to":[12,132],"enhance":[13],"performance":[14],"in":[15,136],"specific":[16],"domains.Better":[17],"results":[18],"can":[19],"be":[20],"achieved":[21],"by":[22,59,76],"distilling":[23],"the":[24,31,70,101,112],"chain-of-thought":[25],"of":[26,33,42],"a":[27,38,45,77],"larger":[28],"model":[29],"at":[30],"cost":[32],"numerous":[34],"expensive":[35],"calls":[36],"and":[37,92,94,110,129],"much":[39],"greater":[40],"amount":[41],"data.We":[43,124],"propose":[44],"novel":[46],"blueprint":[47],"for":[48,55],"efficient":[49],"finetuning":[50],"that":[51,96],"uses":[52],"reasoning":[53],"only":[54],"complex":[56],"data":[57,72,130],"identified":[58],"entropy.Specifically,":[60],"across":[61],"three":[62],"small":[63],"open":[64],"models":[65,88],"(":[66],"3B)":[67],"we":[68],"split":[69],"training":[71],"into":[73],"complexity":[74],"categories":[75],"single":[78],"token":[79],"answer":[80],"entropy":[81],"(ROC":[82],"AUC":[83],"0.73),":[84],"fine-tune":[85],"large":[86],"language":[87],"via":[90],"SFT":[91,103],"distillation,":[93],"show":[95],"our":[97,126],"pipeline":[98],"significantly":[99],"outperforms":[100,111],"standard":[102],"approach":[104,114],"(0.58":[105,115],"vs":[106,116],"0.45":[107],"average":[108,118],"accuracy)":[109,119],"distillation":[113],"0.56":[117],"while":[120],"using":[121],"81%":[122],"less":[123],"publish":[125],"code":[127],"1":[128],"2":[131,141],"facilitate":[133],"further":[134],"research":[135],"this":[137],"direction.1":[138],"https://github.com/LabARSS/":[139],"complexity-aware-fine-tuning":[140],"https:":[142]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-02-27T00:00:00"}