{"id":"https://openalex.org/W7161053567","doi":"https://doi.org/10.48550/arxiv.2605.12384","title":"Scalable Token-Level Hallucination Detection in Large Language Models","display_name":"Scalable Token-Level Hallucination Detection in Large Language Models","publication_year":2026,"publication_date":"2026-05-12","ids":{"openalex":"https://openalex.org/W7161053567","doi":"https://doi.org/10.48550/arxiv.2605.12384"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.12384","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.12384","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.12384","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136016884","display_name":"Rui Min","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Min, Rui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136034441","display_name":"Tianyu Pang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pang, Tianyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136083914","display_name":"Chao Du","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Du, Chao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000534132","display_name":"Minhao Cheng","orcid":"https://orcid.org/0000-0003-3965-4215"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Minhao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5124957280","display_name":"Yi R. Fung","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fung, Yi R.","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":0.1793999969959259,"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":0.1793999969959259,"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/T12488","display_name":"Mental Health via Writing","score":0.14139999449253082,"subfield":{"id":"https://openalex.org/subfields/3207","display_name":"Social Psychology"},"field":{"id":"https://openalex.org/fields/32","display_name":"Psychology"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.11429999768733978,"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":0.7110999822616577},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.6995000243186951},{"id":"https://openalex.org/keywords/detector","display_name":"Detector","score":0.6568999886512756},{"id":"https://openalex.org/keywords/granularity","display_name":"Granularity","score":0.6413999795913696},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6317999958992004},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.6001999974250793},{"id":"https://openalex.org/keywords/segmentation","display_name":"Segmentation","score":0.5758000016212463},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.4440000057220459}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7670999765396118},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7110999822616577},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6995000243186951},{"id":"https://openalex.org/C94915269","wikidata":"https://www.wikidata.org/wiki/Q1834857","display_name":"Detector","level":2,"score":0.6568999886512756},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.6413999795913696},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6395000219345093},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6317999958992004},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6001999974250793},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.5758000016212463},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5120999813079834},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4691999852657318},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4440000057220459},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.36739999055862427},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.34450000524520874},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.34220001101493835},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3147999942302704},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.31220000982284546},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3118000030517578},{"id":"https://openalex.org/C2776145971","wikidata":"https://www.wikidata.org/wiki/Q30673951","display_name":"Labeled data","level":2,"score":0.3095000088214874},{"id":"https://openalex.org/C2776151529","wikidata":"https://www.wikidata.org/wiki/Q3045304","display_name":"Object detection","level":3,"score":0.3089999854564667},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.3043000102043152},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2971999943256378},{"id":"https://openalex.org/C155092808","wikidata":"https://www.wikidata.org/wiki/Q182557","display_name":"Computational linguistics","level":2,"score":0.26460000872612},{"id":"https://openalex.org/C124504099","wikidata":"https://www.wikidata.org/wiki/Q56933","display_name":"Image segmentation","level":3,"score":0.2551000118255615},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.25130000710487366}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.12384","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.12384","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.12384","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.12384","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,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,161],"(LLMs)":[3],"have":[4],"demonstrated":[5],"remarkable":[6],"capabilities,":[7],"but":[8,28],"they":[9],"still":[10],"frequently":[11],"produce":[12],"hallucinations.":[13],"These":[14],"hallucinations":[15],"are":[16],"difficult":[17],"to":[18,44,57,128,174,192],"detect":[19,45],"in":[20],"reasoning-intensive":[21],"tasks,":[22],"where":[23],"the":[24,107,132],"content":[25],"appears":[26],"coherent":[27],"contains":[29],"errors":[30],"like":[31],"logical":[32],"flaws":[33],"and":[34,53,164,189],"unreliable":[35],"intermediate":[36],"results.":[37],"While":[38],"step-level":[39],"analysis":[40],"is":[41],"commonly":[42],"used":[43],"internal":[46],"hallucinations,":[47,130],"it":[48],"suffers":[49],"from":[50,172],"limited":[51],"granularity":[52],"poor":[54],"scalability":[55],"due":[56],"its":[58,195],"reliance":[59],"on":[60,125],"step":[61,136],"segmentation.":[62],"To":[63,104],"address":[64],"these":[65],"limitations,":[66],"we":[67,110,177],"propose":[68],"TokenHD,":[69,120],"a":[70,82,93,113,147],"holistic":[71],"pipeline":[72],"for":[73,86,100,134],"training":[74,94,118],"token-level":[75],"hallucination":[76,89],"detectors.":[77],"Specifically,":[78],"TokenHD":[79],"consists":[80],"of":[81],"scalable":[83],"data":[84],"engine":[85],"synthesizing":[87],"large-scale":[88],"annotations":[90],"along":[91],"with":[92,169],"recipe":[95],"featuring":[96],"an":[97],"importance-weighted":[98],"strategy":[99],"robust":[101],"model":[102,170],"training.":[103],"systematically":[105],"assess":[106],"detection":[108,165],"performance,":[109],"also":[111],"provide":[112],"rigorous":[114],"evaluation":[115],"protocol.":[116],"Through":[117],"within":[119],"our":[121,180],"detector":[122,149,181],"operates":[123],"directly":[124],"free-form":[126],"text":[127,140],"identify":[129],"eliminating":[131],"need":[133],"predefined":[135],"segmentation":[137],"or":[138],"additional":[139],"reformatting.":[141],"Our":[142],"experiments":[143],"show":[144,178],"that":[145,179],"even":[146],"small":[148],"(0.6B)":[150],"achieves":[151],"substantial":[152],"performance":[153,166],"gains":[154],"after":[155],"training,":[156],"surpassing":[157],"much":[158],"larger":[159],"reasoning":[160],"(e.g.,":[162],"QwQ-32B),":[163],"scales":[167],"consistently":[168],"size":[171],"0.6B":[173],"8B.":[175],"Finally,":[176],"can":[182],"generalize":[183],"well":[184],"across":[185],"diverse":[186],"practical":[187],"scenarios":[188],"explore":[190],"strategies":[191],"further":[193],"enhance":[194],"cross-domain":[196],"generalization":[197],"capability.":[198]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-14T00:00:00"}