{"id":"https://openalex.org/W7163040351","doi":"https://doi.org/10.48550/arxiv.2605.30857","title":"MADS: Model-Aware Diverse Core Set Selection for Instruction Tuning","display_name":"MADS: Model-Aware Diverse Core Set Selection for Instruction Tuning","publication_year":2026,"publication_date":"2026-05-29","ids":{"openalex":"https://openalex.org/W7163040351","doi":"https://doi.org/10.48550/arxiv.2605.30857"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.30857","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.30857","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.30857","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137550467","display_name":"Yi Bai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bai, Yi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137576920","display_name":"Wenhao Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Wenhao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137565729","display_name":"Yao Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Yao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137609036","display_name":"Jiao Xue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xue, Jiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137589151","display_name":"Zhumin Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Zhumin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137585058","display_name":"Pengjie Ren","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ren, Pengjie","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":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":0.3813000023365021,"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":0.3813000023365021,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.10480000078678131,"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":0.07590000331401825,"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/set","display_name":"Set (abstract data type)","score":0.6812999844551086},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.6658999919891357},{"id":"https://openalex.org/keywords/core","display_name":"Core (optical fiber)","score":0.6608999967575073},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.5411999821662903},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.499099999666214},{"id":"https://openalex.org/keywords/data-set","display_name":"Data set","score":0.42579999566078186},{"id":"https://openalex.org/keywords/base","display_name":"Base (topology)","score":0.3865000009536743},{"id":"https://openalex.org/keywords/model-selection","display_name":"Model selection","score":0.3635999858379364}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7516000270843506},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6812999844551086},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.6658999919891357},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.6608999967575073},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.5411999821662903},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.499099999666214},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4629000127315521},{"id":"https://openalex.org/C58489278","wikidata":"https://www.wikidata.org/wiki/Q1172284","display_name":"Data set","level":2,"score":0.42579999566078186},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.40290001034736633},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.3865000009536743},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.3635999858379364},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3508000075817108},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.34209999442100525},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3312000036239624},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3154999911785126},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.30230000615119934},{"id":"https://openalex.org/C100808899","wikidata":"https://www.wikidata.org/wiki/Q1192100","display_name":"Set cover problem","level":3,"score":0.30149999260902405},{"id":"https://openalex.org/C78766204","wikidata":"https://www.wikidata.org/wiki/Q555032","display_name":"Multi-core processor","level":2,"score":0.28630000352859497},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.27250000834465027},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.26100000739097595}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.30857","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.30857","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.30857","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.30857","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":"Preprint"},"sustainable_development_goals":[{"score":0.8039178848266602,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Instruction":[0],"fine-tuning":[1,19,185],"is":[2],"employed":[3],"to":[4,105,141],"enhance":[5],"the":[6,15,23,32,35,51,63,84,107,110,129,134,154,164,171,194],"instruction-following":[7],"ability":[8],"of":[9,17,34,62,98,170,182],"large":[10],"language":[11],"models":[12,144,189],"(LLMs).":[13],"As":[14],"amount":[16],"instruction":[18],"data":[20,48,80,213],"increases,":[21],"selecting":[22],"optimal":[24],"core":[25,36,111,130,165],"set":[26,37,131],"becomes":[27],"particularly":[28],"important.":[29],"However,":[30],"ensuring":[31],"diversity":[33,108],"remains":[38],"a":[39,71],"significant":[40],"challenge.":[41],"Existing":[42],"methods":[43],"predominantly":[44],"distinguish":[45],"different":[46],"training":[47,192],"based":[49,82],"on":[50,83,118,153,193,207],"text":[52],"features":[53,81,104],"themselves,":[54],"decoupled":[55],"from":[56],"LLMs'":[57],"own":[58],"understanding":[59],"and":[60,148,174],"representation":[61],"data.":[64],"To":[65],"address":[66],"this":[67],"issue,":[68],"we":[69],"propose":[70],"Model-Aware":[72],"Diverse":[73],"Core":[74],"Set":[75],"Selection":[76],"method,":[77,128],"which":[78,157],"distinguishes":[79],"neural":[85],"activation":[86,103],"states":[87],"during":[88],"LLM":[89,136],"inference.":[90],"This":[91],"approach":[92],"serves":[93],"as":[94],"an":[95,179],"efficient":[96],"instantiation":[97],"coverage-based":[99],"selection":[100],"using":[101],"model-intrinsic":[102],"ensure":[106],"in":[109],"set.":[112],"We":[113],"extensively":[114],"evaluate":[115],"our":[116,127,202],"method":[117,203],"six":[119],"benchmarks":[120],"that":[121,163,201],"cover":[122],"five":[123],"distinct":[124],"tasks.":[125],"In":[126],"selected":[132,175],"by":[133,176],"3B-parameter":[135],"performs":[137],"effectively":[138],"when":[139,184],"utilized":[140],"fine-tune":[142],"larger":[143,187],"with":[145,191],"7B,":[146],"8B,":[147],"13B":[149],"parameters.":[150],"Experimental":[151],"results":[152,199],"Alpaca-GPT4":[155],"dataset,":[156],"comprises":[158],"52K":[159],"instruction-response":[160],"pairs,":[161],"show":[162],"set,":[166],"sized":[167],"at":[168],"15\\%":[169],"original":[172],"dataset":[173],"Llama-3.2-3B-Instruct,":[177],"achieves":[178],"average":[180],"improvement":[181],"2.5\\%":[183],"four":[186],"base":[188],"compared":[190],"full":[195],"dataset.":[196],"The":[197],"experimental":[198],"demonstrate":[200],"enhances":[204],"model":[205],"performance":[206],"multiple":[208],"downstream":[209],"tasks":[210],"while":[211],"reducing":[212],"requirements.":[214]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-06-02T00:00:00"}