{"id":"https://openalex.org/W7157975204","doi":"https://doi.org/10.48550/arxiv.2604.25167","title":"From Insight to Action: A Novel Framework for Interpretability-Guided Data Selection in Large Language Models","display_name":"From Insight to Action: A Novel Framework for Interpretability-Guided Data Selection in Large Language Models","publication_year":2026,"publication_date":"2026-04-28","ids":{"openalex":"https://openalex.org/W7157975204","doi":"https://doi.org/10.48550/arxiv.2604.25167"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.25167","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.25167","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.2604.25167","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134858518","display_name":"Ling Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Shi, Ling","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134871517","display_name":"Xinwei Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Xinwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134839902","display_name":"Xiaohu Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Xiaohu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134821959","display_name":"Hao Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Hao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134855605","display_name":"Heng Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Heng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134853734","display_name":"Yangyang Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yangyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134828856","display_name":"Linlong Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Linlong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134843109","display_name":"Longyue Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Longyue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134840135","display_name":"Deyi Xiong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiong, Deyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134854464","display_name":"Weihua Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Weihua","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5134858518"],"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.7099000215530396,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.7099000215530396,"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":0.04230000078678131,"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/T13910","display_name":"Computational and Text Analysis Methods","score":0.02539999969303608,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/interpretability","display_name":"Interpretability","score":0.9156000018119812},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6639000177383423},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.5986999869346619},{"id":"https://openalex.org/keywords/feature-selection","display_name":"Feature selection","score":0.5055000185966492},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.438400000333786},{"id":"https://openalex.org/keywords/bridge","display_name":"Bridge (graph theory)","score":0.4212999939918518},{"id":"https://openalex.org/keywords/task-analysis","display_name":"Task analysis","score":0.41130000352859497},{"id":"https://openalex.org/keywords/model-selection","display_name":"Model selection","score":0.4036000072956085}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.9156000018119812},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7750999927520752},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.666100025177002},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6639000177383423},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5986999869346619},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5429999828338623},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.5055000185966492},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.438400000333786},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.4212999939918518},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.41659998893737793},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.41130000352859497},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.4036000072956085},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3847000002861023},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.38370001316070557},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3522999882698059},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3508000075817108},{"id":"https://openalex.org/C149364088","wikidata":"https://www.wikidata.org/wiki/Q185917","display_name":"Translation (biology)","level":4,"score":0.32339999079704285},{"id":"https://openalex.org/C24756922","wikidata":"https://www.wikidata.org/wiki/Q1757694","display_name":"Data quality","level":3,"score":0.2870999872684479},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2782999873161316},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.27639999985694885},{"id":"https://openalex.org/C2776145971","wikidata":"https://www.wikidata.org/wiki/Q30673951","display_name":"Labeled data","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2644999921321869},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2612999975681305},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2590000033378601},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.2533000111579895}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.25167","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.25167","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.2604.25167","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.25167","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":{"While":[0],"mechanistic":[1],"interpretability":[2],"tools":[3],"like":[4],"Sparse":[5],"Autoencoders":[6],"(SAEs)":[7],"can":[8],"uncover":[9],"meaningful":[10],"features":[11,47,70,85],"within":[12,98],"Large":[13],"Language":[14],"Models":[15],"(LLMs),":[16],"a":[17,43,49,62,119,143,157],"critical":[18],"gap":[19,34],"remains":[20],"in":[21],"transforming":[22],"these":[23,67],"insights":[24],"into":[25],"practical":[26],"actions":[27],"for":[28,86],"model":[29],"optimization.":[30],"We":[31,88],"bridge":[32],"this":[33],"with":[35],"the":[36,111,129],"hypothesis":[37],"that":[38,64,81],"data":[39,108,137],"selection":[40],"guided":[41],"by":[42,54,118,165],"model's":[44],"internal":[45,168],"task":[46,69,84,151],"is":[48],"effective":[50,160],"training":[51],"strategy.":[52],"Inspired":[53],"this,":[55],"we":[56],"propose":[57],"Interpretability-Guided":[58],"Data":[59],"Selection":[60],"(IGDS),":[61],"framework":[63,161],"first":[65],"identifies":[66],"causal":[68],"through":[71],"frequency":[72],"recall":[73],"and":[74,95,101,131,139,150,159],"interventional":[75],"filtering,":[76],"then":[77],"selects":[78],"``Feature-Resonant":[79],"Data''":[80],"maximally":[82],"activates":[83],"fine-tuning.":[87],"validate":[89],"IGDS":[90,114,154],"on":[91,110,122,136],"mathematical":[92],"reasoning,":[93],"summarization,":[94],"translation":[96],"tasks":[97],"Gemma-2,":[99],"LLaMA-3.1,":[100],"Qwen3":[102],"models.":[103],"Our":[104],"experiments":[105],"demonstrate":[106],"exceptional":[107],"efficiency:":[109],"Math":[112],"task,":[113],"surpasses":[115],"full-dataset":[116],"fine-tuning":[117],"remarkable":[120],"17.4%":[121],"Gemma-2-2B":[123],"while":[124],"using":[125],"only":[126],"50%":[127],"of":[128],"data,":[130],"outperforms":[132],"established":[133],"baselines":[134],"focused":[135],"quality":[138],"diversity.":[140],"Analysis":[141],"confirms":[142],"strong":[144],"positive":[145],"correlation":[146],"between":[147],"feature":[148],"amplification":[149],"performance":[152],"improvement.":[153],"thus":[155],"provides":[156],"direct":[158],"to":[162],"enhance":[163],"LLMs":[164],"leveraging":[166],"their":[167],"mechanisms,":[169],"validating":[170],"our":[171],"core":[172],"hypothesis.":[173]},"counts_by_year":[],"updated_date":"2026-04-30T06:11:10.768123","created_date":"2026-04-30T00:00:00"}