{"id":"https://openalex.org/W4414359498","doi":"https://doi.org/10.24963/ijcai.2025/892","title":"Scalable Multi-Stage Influence Function for Large Language Models via Eigenvalue-Corrected Kronecker-Factored Parameterization","display_name":"Scalable Multi-Stage Influence Function for Large Language Models via Eigenvalue-Corrected Kronecker-Factored Parameterization","publication_year":2025,"publication_date":"2025-09-01","ids":{"openalex":"https://openalex.org/W4414359498","doi":"https://doi.org/10.24963/ijcai.2025/892"},"language":"en","primary_location":{"id":"doi:10.24963/ijcai.2025/892","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/892","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2505.05017","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5110921325","display_name":"Yinghui Bao","orcid":null},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Yuntai Bao","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101722406","display_name":"Xuhong Zhang","orcid":"https://orcid.org/0000-0002-8571-9780"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xuhong Zhang","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102448448","display_name":"Tianyu Du","orcid":null},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tianyu Du","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101765691","display_name":"Xinkui Zhao","orcid":"https://orcid.org/0000-0002-1115-5652"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xinkui Zhao","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113738947","display_name":"Jiang Zong","orcid":null},"institutions":[{"id":"https://openalex.org/I2935922141","display_name":"Universal Technical Institute","ror":"https://ror.org/019njds38","country_code":"US","type":"education","lineage":["https://openalex.org/I2935922141"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jiang Zong","raw_affiliation_strings":["Universal Identification Technology (Hangzhou) Co.,Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Universal Identification Technology (Hangzhou) Co.,Ltd","institution_ids":["https://openalex.org/I2935922141"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048469338","display_name":"Hao Peng","orcid":"https://orcid.org/0000-0003-0586-7132"},"institutions":[{"id":"https://openalex.org/I135237710","display_name":"Zhejiang Normal University","ror":"https://ror.org/01vevwk45","country_code":"CN","type":"education","lineage":["https://openalex.org/I135237710"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao Peng","raw_affiliation_strings":["Zhejiang Normal University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang Normal University","institution_ids":["https://openalex.org/I135237710"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5069353502","display_name":"Jianwei Yin","orcid":"https://orcid.org/0000-0003-4703-7348"},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jianwei Yin","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":7,"corresponding_author_ids":["https://openalex.org/A5110921325"],"corresponding_institution_ids":["https://openalex.org/I76130692"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.12187423,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"8022","last_page":"8030"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9761999845504761,"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":0.9761999845504761,"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/T11550","display_name":"Text and Document Classification Technologies","score":0.9146000146865845,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.7581999897956848},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.6891000270843506},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.5238999724388123},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.40869998931884766},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.3928999900817871},{"id":"https://openalex.org/keywords/empirical-research","display_name":"Empirical research","score":0.3889999985694885},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.3596000075340271}],"concepts":[{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7581999897956848},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6995999813079834},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.6891000270843506},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.5238999724388123},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4214000105857849},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.40869998931884766},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40119999647140503},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.3928999900817871},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.3889999985694885},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.3596000075340271},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.30889999866485596},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.3003999888896942},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.27649998664855957},{"id":"https://openalex.org/C55037315","wikidata":"https://www.wikidata.org/wiki/Q5421151","display_name":"Experimental data","level":2,"score":0.27140000462532043},{"id":"https://openalex.org/C166052673","wikidata":"https://www.wikidata.org/wiki/Q83021","display_name":"Empirical evidence","level":2,"score":0.2687999904155731},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2680000066757202},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2651999890804291},{"id":"https://openalex.org/C133199616","wikidata":"https://www.wikidata.org/wiki/Q25386885","display_name":"Empirical modelling","level":2,"score":0.2554999887943268},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.24963/ijcai.2025/892","is_oa":false,"landing_page_url":"https://doi.org/10.24963/ijcai.2025/892","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2505.05017","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.05017","pdf_url":"https://arxiv.org/pdf/2505.05017","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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"pmh:doi:10.48550/arxiv.2505.05017","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2505.05017","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.05017","pdf_url":"https://arxiv.org/pdf/2505.05017","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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4644492335","display_name":null,"funder_award_id":"62402418","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320323970","display_name":"Ministry of Industry and Information Technology of the People's Republic of China","ror":"https://ror.org/0385nmy68"},{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4414359498.pdf","grobid_xml":"https://content.openalex.org/works/W4414359498.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Pre-trained":[0],"large":[1],"language":[2],"models":[3],"(LLMs)":[4],"are":[5],"commonly":[6],"fine-tuned":[7,26,81],"to":[8,10,28,44,57,64,75,83],"adapt":[9],"downstream":[11,78],"tasks.":[12],"Since":[13],"the":[14,23,77,87,93,114,121],"majority":[15],"of":[16,25,80,97,117,123],"knowledge":[17],"is":[18],"acquired":[19],"during":[20],"pre-training,":[21],"attributing":[22],"predictions":[24,47,79],"LLMs":[27,82],"their":[29],"pre-training":[30,84],"data":[31,85],"may":[32],"provide":[33],"valuable":[34],"insights.":[35],"Influence":[36],"functions":[37,74],"have":[38],"been":[39],"proposed":[40],"as":[41],"a":[42,132],"means":[43],"explain":[45],"model":[46],"based":[48],"on":[49,131],"training":[50],"data.":[51],"However,":[52],"existing":[53],"approaches":[54],"often":[55],"fail":[56],"compute":[58],"\"multi-stage\"":[59],"influence":[60,73,100,126,147],"and":[61,95,120],"lack":[62],"scalability":[63,116],"billion-scale":[65],"LLMs.":[66],"In":[67],"this":[68],"paper,":[69],"we":[70,102],"propose":[71],"multi-stage":[72,99,125,146],"attribute":[76],"under":[86],"full-parameter":[88],"fine-tuning":[89],"paradigm.":[90],"To":[91],"enhance":[92],"efficiency":[94],"practicality":[96],"our":[98,124],"function,":[101],"leverage":[103],"Eigenvalue-corrected":[104],"Kronecker-Factored":[105],"(EK-FAC)":[106],"parameterization":[107],"for":[108],"efficient":[109],"approximation.":[110],"Empirical":[111],"results":[112],"validate":[113],"superior":[115],"EK-FAC":[118],"approximation":[119],"effectiveness":[122],"function.":[127],"Additionally,":[128],"case":[129],"studies":[130],"real-world":[133],"LLM,":[134],"dolly-v2-3b,":[135],"demonstrate":[136],"its":[137],"interpretive":[138],"power,":[139],"with":[140],"exemplars":[141],"illustrating":[142],"insights":[143],"provided":[144],"by":[145],"estimates.":[148]},"counts_by_year":[],"updated_date":"2026-05-23T08:51:43.019350","created_date":"2025-10-10T00:00:00"}