{"id":"https://openalex.org/W7165676248","doi":"https://doi.org/10.48550/arxiv.2606.20820","title":"CELEUS: Certifiable and Efficient LLM Evaluation via E-Processes","display_name":"CELEUS: Certifiable and Efficient LLM Evaluation via E-Processes","publication_year":2026,"publication_date":"2026-06-18","ids":{"openalex":"https://openalex.org/W7165676248","doi":"https://doi.org/10.48550/arxiv.2606.20820"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.20820","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.20820","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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.20820","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139208511","display_name":"Zhijian Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Zhijian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5021782599","display_name":"Z Zhenyu Ye","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ye, Zesheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139167492","display_name":"Zhaorun Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Zhaorun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139152078","display_name":"Bo Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Bo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139149221","display_name":"Feng Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Feng","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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.5047000050544739,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.5047000050544739,"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/T11719","display_name":"Data Quality and Management","score":0.11020000278949738,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.06689999997615814,"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/oracle","display_name":"Oracle","score":0.6240000128746033},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.5015000104904175},{"id":"https://openalex.org/keywords/estimator","display_name":"Estimator","score":0.444599986076355},{"id":"https://openalex.org/keywords/logarithm","display_name":"Logarithm","score":0.43059998750686646},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.4251999855041504},{"id":"https://openalex.org/keywords/cover","display_name":"Cover (algebra)","score":0.3456999957561493},{"id":"https://openalex.org/keywords/measure","display_name":"Measure (data warehouse)","score":0.34389999508857727}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6639999747276306},{"id":"https://openalex.org/C55166926","wikidata":"https://www.wikidata.org/wiki/Q2892946","display_name":"Oracle","level":2,"score":0.6240000128746033},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.5015000104904175},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.4593000113964081},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.444599986076355},{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.43059998750686646},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4251999855041504},{"id":"https://openalex.org/C2780428219","wikidata":"https://www.wikidata.org/wiki/Q16952335","display_name":"Cover (algebra)","level":2,"score":0.3456999957561493},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.34389999508857727},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.34119999408721924},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3375000059604645},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.33239999413490295},{"id":"https://openalex.org/C3018395757","wikidata":"https://www.wikidata.org/wiki/Q1379672","display_name":"Evaluation methods","level":2,"score":0.3131999969482422},{"id":"https://openalex.org/C3018076075","wikidata":"https://www.wikidata.org/wiki/Q1826427","display_name":"Variance components","level":2,"score":0.27250000834465027},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.2662999927997589},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2628999948501587},{"id":"https://openalex.org/C44492722","wikidata":"https://www.wikidata.org/wiki/Q327069","display_name":"Conditional probability","level":2,"score":0.26170000433921814},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.26080000400543213},{"id":"https://openalex.org/C2776292839","wikidata":"https://www.wikidata.org/wiki/Q5179217","display_name":"Coverage probability","level":3,"score":0.26030001044273376},{"id":"https://openalex.org/C62644790","wikidata":"https://www.wikidata.org/wiki/Q3454689","display_name":"Variance reduction","level":3,"score":0.25459998846054077}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.20820","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.20820","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.20820","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.20820","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":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Can":[0],"we":[1,117],"trust":[2],"evaluation":[3,13,29,149],"scores":[4],"to":[5,81,84,112,127,194],"capture":[6],"an":[7],"LLM's":[8],"true":[9,40],"real-world":[10],"performance?":[11],"Certifiable":[12,103],"answers":[14],"this":[15,97],"question":[16],"by":[17,99,185],"providing":[18],"guarantee":[19],"for":[20,105,131,137,147],"LLM":[21,107],"evaluation.":[22],"In":[23],"particular,":[24],"existing":[25,61],"methods":[26,62],"sequentially":[27],"curate":[28],"samples":[30,130,222],"and":[31,79,92,133,157,169,197],"keep":[32],"updating":[33],"confidence":[34],"intervals":[35],"(CIs)":[36],"that":[37,120,142,182,204,212],"cover":[38],"the":[39,53,67,148,153,163,172,199,206,215],"performance":[41],"with":[42,175],"high":[43],"probability":[44],"(e.g.,":[45,70],"95%)":[46,71],"until":[47],"some":[48],"conditions":[49],"are":[50,63,76],"satisfied,":[51],"e.g.,":[52],"CI":[54],"width":[55],"reaches":[56,214],"a":[57,87,102,190],"target":[58,173,216],"precision.":[59],"However,":[60],"not":[64],"generally":[65],"anytime-valid:":[66],"claimed":[68],"coverage":[69],"may":[72],"fail":[73],"when":[74,83],"CIs":[75,183],"repeatedly":[77],"updated":[78],"used":[80],"decide":[82],"stop,":[85],"leaving":[86],"gap":[88,98],"between":[89],"theoretical":[90],"rigor":[91],"practice.":[93],"This":[94],"paper":[95],"bridges":[96],"proposing":[100],"Celeus,":[101],"framework":[104],"Efficient":[106],"evaluation,":[108,132],"which":[109],"leverages":[110],"E-processes":[111],"build":[113],"anytime-valid":[114,158,227],"CIs.":[115,160],"Concretely,":[116],"propose":[118],"signals":[119,144],"combine":[121],"two":[122,164],"ingredients:":[123],"(i)":[124],"Uncertainty-guided":[125],"sampling":[126,202],"select":[128],"informative":[129],"(ii)":[134],"Surrogate-assisted":[135],"approximations":[136],"unevaluated":[138],"samples.":[139,178],"We":[140,179],"prove":[141,181],"such":[143],"remain":[145],"unbiased":[146],"score":[150],"conditional":[151],"on":[152],"past,":[154],"enabling":[155],"statistically-grounded":[156],"$e$-process":[159],"More":[161],"importantly,":[162],"ingredients":[165],"reduce":[166],"estimation":[167],"variance":[168],"help":[170],"reach":[171],"precision":[174,217],"fewer":[176,220],"evaluated":[177,221],"also":[180],"obtained":[184],"Celeus":[186,213],"can":[187],"shrink":[188],"at":[189],"near-parametric":[191],"rate":[192],"up":[193],"logarithmic":[195],"factors":[196],"analyze":[198],"oracle":[200],"variance-optimal":[201],"rule":[203],"motivates":[205],"empirical":[207],"uncertainty-guided":[208],"one.":[209],"Experiments":[210],"show":[211],"using":[218],"54-62%":[219],"than":[223],"baselines,":[224],"while":[225],"preserving":[226],"coverage.":[228]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-24T00:00:00"}