{"id":"https://openalex.org/W7214734896","doi":"https://doi.org/10.48550/arxiv.2609.35279","title":"Measuring Collapse and Correction in Homogeneous-Panel LLM Debate","display_name":"Measuring Collapse and Correction in Homogeneous-Panel LLM Debate","publication_year":2026,"publication_date":"2026-09-28","ids":{"openalex":"https://openalex.org/W7214734896","doi":"https://doi.org/10.48550/arxiv.2609.35279"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.35279","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35279","pdf_url":"https://arxiv.org/pdf/2609.35279","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","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.35279","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153438183","display_name":"Xin Li","orcid":"https://orcid.org/0000-0002-9175-4849","observed_orcids":["https://orcid.org/0000-0002-9175-4849"]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Li, Xin","raw_affiliation_strings":["Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090268549","display_name":"Mengbing Liu","orcid":"https://orcid.org/0000-0002-4920-5280","observed_orcids":["https://orcid.org/0000-0002-4920-5280","https://orcid.org/0009-0003-8351-0849"]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Liu, Mengbing","raw_affiliation_strings":["Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5140774874","display_name":"Chau Yuen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Yuen, Chau","raw_affiliation_strings":["Nanyang Technological University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University","institution_ids":["https://openalex.org/I172675005"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I172675005"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10028","display_name":"Topic Modeling","score":0.2354000061750412,"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":0.2354000061750412,"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/T10456","display_name":"Multi-Agent Systems and Negotiation","score":0.11990000307559967,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.07329999655485153,"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/collapse","display_name":"collapse","score":0.4560000002384186},{"id":"https://openalex.org/keywords/corrections","display_name":"corrections","score":0.9440000057220459},{"id":"https://openalex.org/keywords/multiple-choice-questions","display_name":"multiple-choice questions","score":0.6140000224113464},{"id":"https://openalex.org/keywords/mmlu-pro","display_name":"MMLU-Pro","score":0.8550000190734863}],"concepts":[{"id":"https://openalex.org/C2780385302","wikidata":"https://www.wikidata.org/wiki/Q367158","display_name":"Protocol (science)","level":3,"score":0.61080002784729},{"id":"https://openalex.org/C27415008","wikidata":"https://www.wikidata.org/wiki/Q7256382","display_name":"Psychological intervention","level":2,"score":0.5141000151634216},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.4643999934196472},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4375},{"id":"https://openalex.org/C2780665704","wikidata":"https://www.wikidata.org/wiki/Q959298","display_name":"Intervention (counseling)","level":2,"score":0.4016000032424927},{"id":"https://openalex.org/C134752490","wikidata":"https://www.wikidata.org/wiki/Q374182","display_name":"Logical consequence","level":2,"score":0.37720000743865967},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.3723999857902527},{"id":"https://openalex.org/C2777120189","wikidata":"https://www.wikidata.org/wiki/Q780067","display_name":"Triage","level":2,"score":0.3596999943256378},{"id":"https://openalex.org/C91302708","wikidata":"https://www.wikidata.org/wiki/Q2732056","display_name":"Ledger","level":2,"score":0.35019999742507935},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3276999890804291},{"id":"https://openalex.org/C61423126","wikidata":"https://www.wikidata.org/wiki/Q187432","display_name":"Scripting language","level":2,"score":0.32510000467300415},{"id":"https://openalex.org/C66882249","wikidata":"https://www.wikidata.org/wiki/Q169336","display_name":"Homogeneous","level":2,"score":0.319599986076355},{"id":"https://openalex.org/C162118730","wikidata":"https://www.wikidata.org/wiki/Q1128453","display_name":"Actuarial science","level":1,"score":0.31940001249313354},{"id":"https://openalex.org/C2778223634","wikidata":"https://www.wikidata.org/wiki/Q224952","display_name":"Suspect","level":2,"score":0.3075000047683716},{"id":"https://openalex.org/C108170787","wikidata":"https://www.wikidata.org/wiki/Q3951828","display_name":"Agency (philosophy)","level":2,"score":0.2953000068664551},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.2833999991416931},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.2775999903678894},{"id":"https://openalex.org/C160798450","wikidata":"https://www.wikidata.org/wiki/Q4230870","display_name":"Concordance","level":2,"score":0.26750001311302185},{"id":"https://openalex.org/C190253527","wikidata":"https://www.wikidata.org/wiki/Q295354","display_name":"Law and economics","level":1,"score":0.2655999958515167},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.259799987077713},{"id":"https://openalex.org/C2779010991","wikidata":"https://www.wikidata.org/wiki/Q2720909","display_name":"Artifact (error)","level":2,"score":0.2515000104904175}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.35279","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35279","pdf_url":"https://arxiv.org/pdf/2609.35279","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","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.35279","doi":"https://doi.org/10.48550/arxiv.2609.35279","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.35279","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":"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.35279","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35279","pdf_url":"https://arxiv.org/pdf/2609.35279","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","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214734896.pdf","grobid_xml":"https://content.openalex.org/works/W7214734896.grobid-xml"},"referenced_works_count":13,"referenced_works":["https://openalex.org/W2794325560","https://openalex.org/W4385714610","https://openalex.org/W4415058938","https://openalex.org/W4416050235","https://openalex.org/W4416118450","https://openalex.org/W4416189552","https://openalex.org/W7104027795","https://openalex.org/W7104049147","https://openalex.org/W7128479515","https://openalex.org/W7134052852","https://openalex.org/W7150948673","https://openalex.org/W7155169533","https://openalex.org/W7166837933"],"related_works":["https://openalex.org/W7163272606","https://openalex.org/W7117655689","https://openalex.org/W7160924702","https://openalex.org/W7160437446","https://openalex.org/W7162667636","https://openalex.org/W7161866828","https://openalex.org/W7166681978","https://openalex.org/W7160324419","https://openalex.org/W7166837933","https://openalex.org/W4389217455"],"abstract_inverted_index":{"Multi-agent":[0],"large":[1],"language":[2],"model":[3,221,246],"(LLM)":[4],"debate":[5,53,108,206],"is":[6,18,125,146,157,172],"often":[7],"evaluated":[8],"by":[9],"whether":[10],"final":[11],"answers":[12],"improve,":[13],"but":[14,112,169],"a":[15,21,63,120,147,173],"changed":[16],"answer":[17],"not":[19,190],"necessarily":[20],"better":[22],"one:":[23],"the":[24,78,85,91,103,138,204,252],"same":[25,253],"discussion":[26],"can":[27,136,211,248],"correct":[28,37],"an":[29,35,48],"initially":[30,36],"wrong":[31,139],"majority":[32],"or":[33,195],"collapse":[34,133],"one.":[38],"Standard":[39],"final-accuracy":[40],"evaluations":[41],"conflate":[42],"these":[43],"opposing":[44],"transitions.":[45],"We":[46,232],"introduce":[47],"auditable":[49],"protocol":[50,79],"for":[51,127],"homogeneous":[52],"on":[54],"multiple-choice":[55],"questions":[56],"(MCQs)":[57],"that":[58,87,124],"records":[59,80],"each":[60],"run":[61],"as":[62,193],"transition":[64],"ledger":[65],"over":[66],"collapse,":[67],"correction,":[68],"onset,":[69],"and":[70,90,119,215,230,240,255],"signed":[71,256],"intervention":[72],"utility.":[73],"On":[74],"6,925":[75],"MMLU-Pro":[76],"debates,":[77],"253":[81],"collapses":[82,111,202,214,229],"together":[83],"with":[84,154],"corrections":[86,115],"offset":[88],"them,":[89],"correction":[92],"side":[93],"changes":[94],"how":[95],"interventions":[96],"should":[97],"be":[98,249],"judged.":[99],"Replay":[100],"experiments":[101],"reveal":[102],"central":[104],"tradeoff:":[105],"freezing":[106],"every":[107],"prevents":[109],"251":[110],"loses":[113],"804":[114],"under":[116,251],"equal":[117],"weights,":[118],"leave-one-model-out":[121],"probe-gated":[122],"freeze":[123],"tuned":[126],"net":[128],"utility":[129,257],"never":[130],"fires,":[131],"so":[132,187,244],"prevention":[134],"alone":[135],"recommend":[137],"policy.":[140],"A":[141],"compact":[142],"pre-debate":[143],"8-probe":[144],"screen":[145],"triage":[148],"signal:":[149],"its":[150],"unadjusted":[151],"family-level":[152],"association":[153],"conditional-collapse":[155],"risk":[156],"high":[158],"(G=7,":[159],"Spearman":[160],"$\u03c1$":[161,181],"=":[162,167,177,182,185],"0.893,":[163],"exact":[164],"two-sided":[165],"p":[166,184],"0.0123),":[168],"initial-majority":[170],"accuracy":[171],"close":[174],"comparator":[175],"($\u03c1$":[176],"-0.821;":[178],"family":[179],"partial":[180],"0.767,":[183],"0.0877),":[186],"we":[188],"do":[189],"treat":[191],"it":[192],"calibrated":[194],"capability-adjusted":[196],"prediction.":[197],"Round-level":[198],"traces":[199],"localize":[200],"many":[201],"to":[203],"first":[205],"round,":[207],"where":[208],"early":[209],"disagreement":[210],"precede":[212],"both":[213,228],"corrections.":[216,231],"Reasoning-mode":[217],"models,":[218],"including":[219],"one":[220],"at":[222],"two":[223],"reasoning-effort":[224],"levels,":[225],"also":[226],"show":[227],"release":[233],"replayable":[234],"schemas,":[235],"coders,":[236],"audits,":[237],"cost":[238],"cards,":[239],"zero-API":[241],"rebuild":[242],"scripts":[243],"future":[245],"rows":[247],"compared":[250],"denominators":[254],"ledger.":[258]},"counts_by_year":[],"updated_date":"2026-11-11T12:16:26.678958","created_date":"2026-09-30T00:00:00"}