{"id":"https://openalex.org/W7128536746","doi":"https://doi.org/10.48550/arxiv.2602.07909","title":"SparseEval: Efficient Evaluation of Large Language Models by Sparse Optimization","display_name":"SparseEval: Efficient Evaluation of Large Language Models by Sparse Optimization","publication_year":2026,"publication_date":"2026-02-08","ids":{"openalex":"https://openalex.org/W7128536746","doi":"https://doi.org/10.48550/arxiv.2602.07909"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2602.07909","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.07909","pdf_url":"https://arxiv.org/pdf/2602.07909","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,"listed_in":[],"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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"parent_works":[],"lineage":["https://openalex.org/W7128536746"],"child_works_count":0,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2602.07909","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125541308","display_name":"Taolin Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I180726961","display_name":"Shenzhen University","ror":"https://ror.org/01vy4gh70","country_code":"CN","type":"education","lineage":["https://openalex.org/I180726961"]},{"id":"https://openalex.org/I4405263052","display_name":"Tsinghua Shenzhen International Graduate School","ror":"https://ror.org/01vceef04","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405263052","https://openalex.org/I99065089"]},{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Taolin","raw_affiliation_strings":["College of Computer Science and Software Engineering , Shenzhen University","Tsinghua Shenzhen International Graduate School , Tsinghua University 3 Independent ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Computer Science and Software Engineering , Shenzhen University","institution_ids":["https://openalex.org/I180726961"]},{"raw_affiliation_string":"Tsinghua Shenzhen International Graduate School , Tsinghua University 3 Independent ,","institution_ids":["https://openalex.org/I4405263052","https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5159865225","display_name":"Hang Guo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405263052","display_name":"Tsinghua Shenzhen International Graduate School","ror":"https://ror.org/01vceef04","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405263052","https://openalex.org/I99065089"]},{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Guo, Hang","raw_affiliation_strings":["Tsinghua Shenzhen International Graduate School , Tsinghua University 3 Independent ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tsinghua Shenzhen International Graduate School , Tsinghua University 3 Independent ,","institution_ids":["https://openalex.org/I4405263052","https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125534615","display_name":"Wang Lu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101571196","display_name":"Tao Dai","orcid":"https://orcid.org/0000-0002-1187-5019","observed_orcids":["https://orcid.org/0000-0002-1187-5019"]},"institutions":[{"id":"https://openalex.org/I180726961","display_name":"Shenzhen University","ror":"https://ror.org/01vy4gh70","country_code":"CN","type":"education","lineage":["https://openalex.org/I180726961"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dai, Tao","raw_affiliation_strings":["College of Computer Science and Software Engineering , Shenzhen University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"College of Computer Science and Software Engineering , Shenzhen University","institution_ids":["https://openalex.org/I180726961"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034104790","display_name":"Shu\u2010Tao Xia","orcid":"https://orcid.org/0000-0002-8639-982X","observed_orcids":["https://orcid.org/0000-0002-8639-982X"]},"institutions":[{"id":"https://openalex.org/I4405263052","display_name":"Tsinghua Shenzhen International Graduate School","ror":"https://ror.org/01vceef04","country_code":"CN","type":"education","lineage":["https://openalex.org/I4405263052","https://openalex.org/I99065089"]},{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xia, Shu-Tao","raw_affiliation_strings":["Tsinghua Shenzhen International Graduate School , Tsinghua University 3 Independent ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tsinghua Shenzhen International Graduate School , Tsinghua University 3 Independent ,","institution_ids":["https://openalex.org/I4405263052","https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5125519836","display_name":"Jindong Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Jindong","raw_affiliation_strings":["Microsoft Research Asia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":5,"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":0.9429000020027161,"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":0.9429000020027161,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.016499999910593033,"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"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.010999999940395355,"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/llm-evaluation","display_name":"LLM evaluation","score":0.7860000133514404},{"id":"https://openalex.org/keywords/benchmark-selection","display_name":"benchmark selection","score":0.5830000042915344},{"id":"https://openalex.org/keywords/sparse-optimization","display_name":"sparse optimization","score":0.8790000081062317},{"id":"https://openalex.org/keywords/anchor-selection","display_name":"anchor selection","score":0.7870000004768372},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"gradient descent","score":0.5070000290870667},{"id":"https://openalex.org/keywords/iterative-refinement","display_name":"iterative refinement","score":0.8050000071525574},{"id":"https://openalex.org/keywords/kendall-s-tau","display_name":"Kendall's tau","score":0.5220000147819519},{"id":"https://openalex.org/keywords/evaluation-cost-reduction","display_name":"evaluation cost reduction","score":0.4880000054836273}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.772599995136261},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.6642000079154968},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6172999739646912},{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.5985000133514404},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5701000094413757},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.527899980545044},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.49559998512268066},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44600000977516174},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.43869999051094055},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3995000123977661},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.38519999384880066},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3752000033855438},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.3725000023841858},{"id":"https://openalex.org/C124066611","wikidata":"https://www.wikidata.org/wiki/Q28684319","display_name":"Sparse approximation","level":2,"score":0.3666999936103821},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.3522000014781952},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.3407999873161316},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.3239000141620636},{"id":"https://openalex.org/C77637269","wikidata":"https://www.wikidata.org/wiki/Q7002051","display_name":"Neural coding","level":2,"score":0.3190999925136566},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.30250000953674316},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.2973000109195709},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.29010000824928284},{"id":"https://openalex.org/C159694833","wikidata":"https://www.wikidata.org/wiki/Q2321565","display_name":"Iterative method","level":2,"score":0.2842000126838684},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.2833999991416931},{"id":"https://openalex.org/C2987595161","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Optimization algorithm","level":2,"score":0.27639999985694885},{"id":"https://openalex.org/C83546350","wikidata":"https://www.wikidata.org/wiki/Q1139051","display_name":"Regression","level":2,"score":0.272599995136261},{"id":"https://openalex.org/C134261354","wikidata":"https://www.wikidata.org/wiki/Q938438","display_name":"Statistical inference","level":2,"score":0.2662000060081482},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.2547999918460846},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25380000472068787},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.25110000371932983}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2602.07909","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.07909","pdf_url":"https://arxiv.org/pdf/2602.07909","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,"listed_in":[],"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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2602.07909","doi":"https://doi.org/10.48550/arxiv.2602.07909","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.07909","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,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2602.07909","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2602.07909","pdf_url":"https://arxiv.org/pdf/2602.07909","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,"listed_in":[],"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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G3025693001","display_name":null,"funder_award_id":"62302309","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G824329526","display_name":null,"funder_award_id":"62571298","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"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7128536746.pdf","grobid_xml":"https://content.openalex.org/works/W7128536746.grobid-xml"},"referenced_works_count":17,"referenced_works":["https://openalex.org/W2786672974","https://openalex.org/W2794325560","https://openalex.org/W2950700477","https://openalex.org/W3030163527","https://openalex.org/W3083410900","https://openalex.org/W3210904070","https://openalex.org/W4327810158","https://openalex.org/W4387225871","https://openalex.org/W4401042935","https://openalex.org/W4401660192","https://openalex.org/W4403556531","https://openalex.org/W4407231762","https://openalex.org/W4407719847","https://openalex.org/W4411630215","https://openalex.org/W4412945299","https://openalex.org/W4414281281","https://openalex.org/W6948070318"],"related_works":["https://openalex.org/W4415152267","https://openalex.org/W6948070318","https://openalex.org/W7160840613","https://openalex.org/W7221755302","https://openalex.org/W4400601001","https://openalex.org/W4400488479","https://openalex.org/W4401042935","https://openalex.org/W4417134733","https://openalex.org/W7147165704","https://openalex.org/W6947952751"],"abstract_inverted_index":{"As":[0],"large":[1,31],"language":[2],"models":[3],"(LLMs)":[4],"continue":[5],"to":[6,95,115,129],"scale":[7],"up,":[8],"their":[9,20],"performance":[10,47],"on":[11,29,79],"various":[12],"downstream":[13],"tasks":[14],"has":[15,22],"significantly":[16],"improved.":[17],"However,":[18],"evaluating":[19],"capabilities":[21],"become":[23],"increasingly":[24],"expensive,":[25],"as":[26,61,73],"performing":[27],"inference":[28],"a":[30,74,85,153],"number":[32],"of":[33,67,113,133,149,155],"benchmark":[34],"samples":[35],"incurs":[36],"high":[37,147],"computational":[38],"costs.":[39],"In":[40],"this":[41],"paper,":[42],"we":[43,82],"revisit":[44],"the":[45,65,89,110,121,131,142],"model-item":[46],"matrix":[48],"and":[49,63,99,119,125,146,161],"show":[50],"that":[51,55,64],"it":[52],"exhibits":[53],"sparsity,":[54],"representative":[56],"items":[57],"can":[58,70],"be":[59,71],"selected":[60],"anchors,":[62],"task":[66],"efficient":[68],"benchmarking":[69],"formulated":[72],"sparse":[75,117],"optimization":[76,118],"problem.":[77],"Based":[78],"these":[80],"insights,":[81],"propose":[83,120],"SparseEval,":[84],"method":[86,151],"that,":[87],"for":[88,105,136],"first":[90],"time,":[91],"adopts":[92],"gradient":[93],"descent":[94],"optimize":[96],"anchor":[97,106],"weights":[98],"employs":[100],"an":[101],"iterative":[102],"refinement":[103],"strategy":[104],"selection.":[107],"We":[108],"utilize":[109],"representation":[111],"capacity":[112],"MLP":[114],"handle":[116],"Anchor":[122],"Importance":[123,127],"Score":[124,128],"Candidate":[126],"evaluate":[130],"value":[132],"each":[134],"item":[135],"task-aware":[137],"refinement.":[138],"Extensive":[139],"experiments":[140],"demonstrate":[141],"low":[143],"estimation":[144],"error":[145],"Kendall's~$\u03c4$":[148],"our":[150],"across":[152],"variety":[154],"benchmarks,":[156],"showcasing":[157],"its":[158],"superior":[159],"robustness":[160],"practicality":[162],"in":[163],"real-world":[164],"scenarios.":[165],"Code":[166],"is":[167],"available":[168],"at":[169],"{https://github.com/taolinzhang/SparseEval}.":[170]},"counts_by_year":[{"year":2026,"cited_by_count":5}],"updated_date":"2026-11-14T11:04:58.165337","created_date":"2026-02-11T00:00:00"}