{"id":"https://openalex.org/W7225824404","doi":"https://doi.org/10.21203/rs.3.rs-10959650/v1","title":"Auditable Release Assurance for Quantized LLM Deployment Configurations: Anytime-Valid Non-Inferiority and Executable Certificates","display_name":"Auditable Release Assurance for Quantized LLM Deployment Configurations: Anytime-Valid Non-Inferiority and Executable Certificates","publication_year":2026,"publication_date":"2026-11-11","ids":{"openalex":"https://openalex.org/W7225824404","doi":"https://doi.org/10.21203/rs.3.rs-10959650/v1"},"language":"en","primary_location":{"id":"doi:10.21203/rs.3.rs-10959650/v1","doi":"https://doi.org/10.21203/rs.3.rs-10959650/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-10959650/v1","pdf_url":null,"source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Research Square","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.21203/rs.3.rs-10959650/v1","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5121243531","display_name":"Myunglae Jo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I31419693","display_name":"Kyungpook National University","ror":"https://ror.org/040c17130","country_code":"KR","type":"education","lineage":["https://openalex.org/I31419693"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Myunglae Jo","raw_affiliation_strings":["Kyungpook National University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Kyungpook National University","institution_ids":["https://openalex.org/I31419693"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5145714652","display_name":"Donghyun Kim","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Donghyun Kim","raw_affiliation_strings":["Palusomni Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Palusomni Co., Ltd","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5161492493","display_name":"Changhan Son","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Changhan Son","raw_affiliation_strings":["Palusomni Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Palusomni Co., Ltd","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5161489903","display_name":"Minseong Park","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Minseong Park","raw_affiliation_strings":["Palusomni Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Palusomni Co., Ltd","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025647721","display_name":"Il-Hyuk Yu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Il-Hyuk Yu","raw_affiliation_strings":["Palusomni Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Palusomni Co., Ltd","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5049298364","display_name":"Hyungkwang Lim","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hyungkwang Lim","raw_affiliation_strings":["Palusomni Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Palusomni Co., Ltd","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"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/T13295","display_name":"Safety Systems Engineering in Autonomy","score":0.2029999941587448,"subfield":{"id":"https://openalex.org/subfields/2213","display_name":"Safety, Risk, Reliability and Quality"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T13295","display_name":"Safety Systems Engineering in Autonomy","score":0.2029999941587448,"subfield":{"id":"https://openalex.org/subfields/2213","display_name":"Safety, Risk, Reliability and Quality"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10743","display_name":"Software Testing and Debugging Techniques","score":0.14149999618530273,"subfield":{"id":"https://openalex.org/subfields/1712","display_name":"Software"},"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/T10430","display_name":"Software Engineering Techniques and Practices","score":0.07100000232458115,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7009999752044678},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.6586999893188477},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.48190000653266907},{"id":"https://openalex.org/C160145156","wikidata":"https://www.wikidata.org/wiki/Q778586","display_name":"Executable","level":2,"score":0.4607999920845032},{"id":"https://openalex.org/C96865113","wikidata":"https://www.wikidata.org/wiki/Q2946816","display_name":"Certificate","level":2,"score":0.45590001344680786},{"id":"https://openalex.org/C46304622","wikidata":"https://www.wikidata.org/wiki/Q374814","display_name":"Certification","level":2,"score":0.436599999666214},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.4214000105857849},{"id":"https://openalex.org/C2780385302","wikidata":"https://www.wikidata.org/wiki/Q367158","display_name":"Protocol (science)","level":3,"score":0.4090999960899353},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.3955000042915344},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3521000146865845},{"id":"https://openalex.org/C153180980","wikidata":"https://www.wikidata.org/wiki/Q19776675","display_name":"Commit","level":2,"score":0.33340001106262207},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.30570000410079956},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.30489999055862427},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.29330000281333923},{"id":"https://openalex.org/C147007100","wikidata":"https://www.wikidata.org/wiki/Q3344190","display_name":"External auditor","level":4,"score":0.2892000079154968},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.2872999906539917},{"id":"https://openalex.org/C2992734406","wikidata":"https://www.wikidata.org/wiki/Q413267","display_name":"One shot","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.28189998865127563},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.2793000042438507},{"id":"https://openalex.org/C2776633867","wikidata":"https://www.wikidata.org/wiki/Q186612","display_name":"Timer","level":3,"score":0.2775000035762787},{"id":"https://openalex.org/C200601418","wikidata":"https://www.wikidata.org/wiki/Q2193887","display_name":"Reliability engineering","level":1,"score":0.27459999918937683},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.27079999446868896},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.25999999046325684},{"id":"https://openalex.org/C75291252","wikidata":"https://www.wikidata.org/wiki/Q1315756","display_name":"TRACE (psycholinguistics)","level":2,"score":0.25999999046325684},{"id":"https://openalex.org/C2779010991","wikidata":"https://www.wikidata.org/wiki/Q2720909","display_name":"Artifact (error)","level":2,"score":0.25850000977516174},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.25440001487731934}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21203/rs.3.rs-10959650/v1","doi":"https://doi.org/10.21203/rs.3.rs-10959650/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-10959650/v1","pdf_url":null,"source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Research Square","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.21203/rs.3.rs-10959650/v1","doi":"https://doi.org/10.21203/rs.3.rs-10959650/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-10959650/v1","pdf_url":null,"source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Research Square","raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Abstract":[0],"Quantization":[1],"changes":[2],"not":[3],"only":[4],"model":[5,37],"size":[6],"but":[7],"the":[8,110,119,140,153,170,173,179,207],"deployed":[9],"software":[10],"configuration":[11,43],",":[12],"yet":[13],"release":[14,218],"decisions":[15],"are":[16],"often":[17],"based":[18],"on":[19,155],"a":[20,33,41,45,69,73,97,113,128,186,225],"single":[21],"benchmark":[22],"gap.":[23],"We":[24,151],"present":[25],"an":[26,135],"auditable":[27],"software-quality":[28],"gate":[29],"for":[30,185,206],"deciding":[31],"whether":[32],"quantized":[34,209],"large":[35],"language":[36],"(LLM)":[38],"may":[39],"replace":[40],"reference":[42,141],"under":[44,88],"declared":[46],"non-inferiority":[47],"margin":[48],"\u03b5":[49],"and":[50,60,72,82,96,104,117,125,148,165,200],"false-certification":[51,176],"budget":[52],"\u03b4.":[53],"The":[54,220],"method":[55],"uses":[56],"paired":[57],"item-level":[58],"outcomes":[59],"supports":[61],"two":[62],"distinct":[63],"targets:":[64],"future":[65],"requests":[66],"sampled":[67],"from":[68],"task":[70],"distribution":[71],"committed":[74],"finite":[75],"audit":[76,191],"frame":[77],"inspected":[78],"without":[79],"replacement.":[80],"Fixed-sample":[81],"anytime-valid":[83],"procedures":[84],"control":[85],"false":[86],"certification":[87],"their":[89],"stated":[90],"assumptions,":[91],"while":[92],"precommitment,":[93],"candidate-selection":[94],"rules,":[95],"lifetime":[98],"risk":[99],"ledger":[100],"address":[101],"optional":[102],"stopping":[103,195],"repeated":[105],"releases.":[106],"Protocol":[107],"3.1":[108],"binds":[109],"decision":[111],"to":[112,118],"versioned":[114],"deployment":[115,167],"profile":[116],"supplied":[120],"task,":[121],"evaluator,":[122],"frame,":[123],"selection,":[124],"ordered-record":[126],"files;":[127],"verifier":[129],"that":[130],"can":[131],"be":[132],"run":[133],"by":[134],"independent":[136],"third":[137],"party":[138],"replays":[139],"implemen-tation\u2019s":[142],"bindings,":[143],"selected":[144],"candidate,":[145],"inspection":[146],"order,":[147],"sequential-evidence":[149],"checks.":[150],"evaluate":[152],"procedure":[154],"twelve":[156],"measured":[157],"model\u2013task\u2013quantization":[158],"populations":[159],"spanning":[160],"NVIDIA":[161],"Jetson":[162],"Orin":[163],"Nano":[164],"Mac":[166],"configurations.":[168],"For":[169],"proposed":[171],"procedures,":[172],"maximum":[174],"observed":[175],"rate":[177],"across":[178],"simulated":[180],"null-region":[181],"cells":[182],"was":[183],"3.8%":[184],"5%":[187],"budget.":[188],"Representative":[189],"500-item":[190],"frames":[192],"reached":[193,211],"retrospective":[194,226],"points":[196],"between":[197],"167":[198],"1":[199],"423":[201],"items.":[202],"Cross-configuration":[203],"parsed-prediction":[204],"disagreement":[205],"same":[208],"artifact":[210],"11.8%,":[212],"supporting":[213],"configuration-specific":[214],"rather":[215],"than":[216],"file-specific":[217],"evidence.":[219],"current":[221],"certificate":[222],"example":[223],"is":[224],"replay;":[227],"prospective":[228],"assurance":[229],"additionally":[230],"requires":[231],"external":[232],"anchoring.":[233]},"counts_by_year":[],"updated_date":"2026-11-12T12:28:57.059296","created_date":"2026-11-12T00:00:00"}
