{"id":"https://openalex.org/W4399803256","doi":"https://doi.org/10.1038/s41586-024-07421-0","title":"Detecting hallucinations in large language models using semantic entropy","display_name":"Detecting hallucinations in large language models using semantic entropy","publication_year":2024,"publication_date":"2024-06-19","ids":{"openalex":"https://openalex.org/W4399803256","doi":"https://doi.org/10.1038/s41586-024-07421-0","pmid":"https://pubmed.ncbi.nlm.nih.gov/38898292"},"language":"en","primary_location":{"id":"doi:10.1038/s41586-024-07421-0","is_oa":true,"landing_page_url":"https://doi.org/10.1038/s41586-024-07421-0","pdf_url":"https://www.nature.com/articles/s41586-024-07421-0.pdf","source":{"id":"https://openalex.org/S137773608","display_name":"Nature","issn_l":"0028-0836","issn":["0028-0836","1476-4687"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319908","host_organization_name":"Nature Portfolio","host_organization_lineage":["https://openalex.org/P4310319908","https://openalex.org/P4310319965"],"host_organization_lineage_names":["Nature Portfolio","Springer Nature"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Nature","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","pubmed"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://www.nature.com/articles/s41586-024-07421-0.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5051690066","display_name":"Sebastian Farquhar","orcid":"https://orcid.org/0000-0002-9185-6415"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]},{"id":"https://openalex.org/I4210146410","display_name":"Science Oxford","ror":"https://ror.org/04j8yhy50","country_code":"GB","type":"nonprofit","lineage":["https://openalex.org/I4210146410"]}],"countries":["GB"],"is_corresponding":true,"raw_author_name":"Sebastian Farquhar","raw_affiliation_strings":["OATML, Department of Computer Science, University of Oxford, Oxford, UK. sebfar@gmail.com","OATML, Department of Computer Science, University of Oxford, Oxford, UK"],"raw_orcid":"https://orcid.org/0000-0002-9185-6415","affiliations":[{"raw_affiliation_string":"OATML, Department of Computer Science, University of Oxford, Oxford, UK. sebfar@gmail.com","institution_ids":["https://openalex.org/I4210146410"]},{"raw_affiliation_string":"OATML, Department of Computer Science, University of Oxford, Oxford, UK","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060914941","display_name":"Jannik Kossen","orcid":null},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Jannik Kossen","raw_affiliation_strings":["OATML, Department of Computer Science, University of Oxford, Oxford, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"OATML, Department of Computer Science, University of Oxford, Oxford, UK","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054381323","display_name":"Lorenz Kuhn","orcid":null},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Lorenz Kuhn","raw_affiliation_strings":["OATML, Department of Computer Science, University of Oxford, Oxford, UK"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"OATML, Department of Computer Science, University of Oxford, Oxford, UK","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5029186201","display_name":"Yarin Gal","orcid":"https://orcid.org/0000-0002-2733-2078"},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Yarin Gal","raw_affiliation_strings":["OATML, Department of Computer Science, University of Oxford, Oxford, UK"],"raw_orcid":"https://orcid.org/0000-0002-2733-2078","affiliations":[{"raw_affiliation_string":"OATML, Department of Computer Science, University of Oxford, Oxford, UK","institution_ids":["https://openalex.org/I40120149"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5051690066"],"corresponding_institution_ids":["https://openalex.org/I40120149","https://openalex.org/I4210146410"],"apc_list":{"value":11690,"currency":"USD","value_usd":11690},"apc_paid":{"value":11690,"currency":"USD","value_usd":11690},"fwci":166.8465,"has_fulltext":true,"cited_by_count":697,"citation_normalized_percentile":{"value":0.99987873,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":100},"biblio":{"volume":"630","issue":"8017","first_page":"625","last_page":"630"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9998000264167786,"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":0.9998000264167786,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.991599977016449,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.9857000112533569,"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/computer-science","display_name":"Computer science","score":0.5953117609024048},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5451573133468628},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.45756322145462036},{"id":"https://openalex.org/keywords/meaning","display_name":"Meaning (existential)","score":0.4131714403629303},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.38635456562042236},{"id":"https://openalex.org/keywords/cognitive-psychology","display_name":"Cognitive psychology","score":0.37735870480537415},{"id":"https://openalex.org/keywords/data-science","display_name":"Data science","score":0.35610896348953247},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.3233878016471863},{"id":"https://openalex.org/keywords/psychotherapist","display_name":"Psychotherapist","score":0.11443063616752625}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5953117609024048},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5451573133468628},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45756322145462036},{"id":"https://openalex.org/C2780876879","wikidata":"https://www.wikidata.org/wiki/Q3054749","display_name":"Meaning (existential)","level":2,"score":0.4131714403629303},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.38635456562042236},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.37735870480537415},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.35610896348953247},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3233878016471863},{"id":"https://openalex.org/C542102704","wikidata":"https://www.wikidata.org/wiki/Q183257","display_name":"Psychotherapist","level":1,"score":0.11443063616752625},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"score":0.0},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0}],"mesh":[{"descriptor_ui":"D006212","descriptor_name":"Hallucinations","qualifier_ui":"Q000175","qualifier_name":"diagnosis","is_major_topic":true},{"descriptor_ui":"D006212","descriptor_name":"Hallucinations","qualifier_ui":"Q000523","qualifier_name":"psychology","is_major_topic":true},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D007802","descriptor_name":"Language","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D012660","descriptor_name":"Semantics","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D019277","descriptor_name":"Entropy","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D035501","descriptor_name":"Uncertainty","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false}],"locations_count":5,"locations":[{"id":"doi:10.1038/s41586-024-07421-0","is_oa":true,"landing_page_url":"https://doi.org/10.1038/s41586-024-07421-0","pdf_url":"https://www.nature.com/articles/s41586-024-07421-0.pdf","source":{"id":"https://openalex.org/S137773608","display_name":"Nature","issn_l":"0028-0836","issn":["0028-0836","1476-4687"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319908","host_organization_name":"Nature Portfolio","host_organization_lineage":["https://openalex.org/P4310319908","https://openalex.org/P4310319965"],"host_organization_lineage_names":["Nature Portfolio","Springer Nature"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Nature","raw_type":"journal-article"},{"id":"pmid:38898292","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/38898292","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Nature","raw_type":null},{"id":"pmh:oai:pubmedcentral.nih.gov:11186750","is_oa":true,"landing_page_url":"https://www.ncbi.nlm.nih.gov/pmc/articles/11186750","pdf_url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC11186750/pdf/41586_2024_Article_7421.pdf","source":{"id":"https://openalex.org/S2764455111","display_name":"PubMed Central","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"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":"Nature","raw_type":"Text"},{"id":"pmh:oai:RePEc:nat:nature:v:630:y:2024:i:8017:d:10.1038_s41586-024-07421-0","is_oa":false,"landing_page_url":"https://www.nature.com/articles/s41586-024-07421-0","pdf_url":null,"source":{"id":"https://openalex.org/S4306401271","display_name":"RePEc: Research Papers in Economics","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I77793887","host_organization_name":"Federal Reserve Bank of St. Louis","host_organization_lineage":["https://openalex.org/I77793887"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},{"id":"pmh:oai:ora.ox.ac.uk:uuid:0653d09e-9368-4eb1-98bb-50d9dda7d3e5","is_oa":true,"landing_page_url":"https://ora.ox.ac.uk/objects/uuid:0653d09e-9368-4eb1-98bb-50d9dda7d3e5","pdf_url":null,"source":{"id":"https://openalex.org/S4306402636","display_name":"Oxford University Research Archive (ORA) (University of Oxford)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I40120149","host_organization_name":"University of Oxford","host_organization_lineage":["https://openalex.org/I40120149"],"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":"Jisc Publications Router","raw_type":"Journal article"}],"best_oa_location":{"id":"doi:10.1038/s41586-024-07421-0","is_oa":true,"landing_page_url":"https://doi.org/10.1038/s41586-024-07421-0","pdf_url":"https://www.nature.com/articles/s41586-024-07421-0.pdf","source":{"id":"https://openalex.org/S137773608","display_name":"Nature","issn_l":"0028-0836","issn":["0028-0836","1476-4687"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319908","host_organization_name":"Nature Portfolio","host_organization_lineage":["https://openalex.org/P4310319908","https://openalex.org/P4310319965"],"host_organization_lineage_names":["Nature Portfolio","Springer Nature"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Nature","raw_type":"journal-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.6000000238418579}],"awards":[],"funders":[{"id":"https://openalex.org/F4320314707","display_name":"Government of the United Kingdom","ror":"https://ror.org/05wnh3t63"},{"id":"https://openalex.org/F4320314731","display_name":"UK Research and Innovation","ror":"https://ror.org/001aqnf71"}],"has_content":{"pdf":true,"grobid_xml":false},"content_urls":{"pdf":"https://content.openalex.org/works/W4399803256.pdf"},"referenced_works_count":49,"referenced_works":["https://openalex.org/W1981208470","https://openalex.org/W2006336194","https://openalex.org/W2040103855","https://openalex.org/W2076580309","https://openalex.org/W2111959010","https://openalex.org/W2115305054","https://openalex.org/W2145815109","https://openalex.org/W2295477204","https://openalex.org/W2912924812","https://openalex.org/W2951211142","https://openalex.org/W2951434086","https://openalex.org/W2962735233","https://openalex.org/W2963091658","https://openalex.org/W2963096510","https://openalex.org/W2963339397","https://openalex.org/W2963382396","https://openalex.org/W2963748441","https://openalex.org/W2963846996","https://openalex.org/W3034383590","https://openalex.org/W3103178756","https://openalex.org/W3104939451","https://openalex.org/W3153712677","https://openalex.org/W3170403598","https://openalex.org/W3174418826","https://openalex.org/W3176948526","https://openalex.org/W3199958362","https://openalex.org/W3205287819","https://openalex.org/W3213990450","https://openalex.org/W4206775337","https://openalex.org/W4229019162","https://openalex.org/W4285240908","https://openalex.org/W4285298706","https://openalex.org/W4309674289","https://openalex.org/W4318069287","https://openalex.org/W4376959396","https://openalex.org/W4379468930","https://openalex.org/W4384918448","https://openalex.org/W4385573914","https://openalex.org/W4386065840","https://openalex.org/W4389520749","https://openalex.org/W4390041933","https://openalex.org/W4391215636","https://openalex.org/W6718836005","https://openalex.org/W6778883912","https://openalex.org/W6810242208","https://openalex.org/W6839548382","https://openalex.org/W6852907298","https://openalex.org/W6858023062","https://openalex.org/W6949250862"],"related_works":["https://openalex.org/W2088836968","https://openalex.org/W2369480852","https://openalex.org/W2017063170","https://openalex.org/W2360473631","https://openalex.org/W2897855253","https://openalex.org/W2809651922","https://openalex.org/W2313927466","https://openalex.org/W2352901769","https://openalex.org/W2041613042","https://openalex.org/W4211075255"],"abstract_inverted_index":{"Abstract":[0],"Large":[1],"language":[2],"model":[3],"(LLM)":[4],"systems,":[5],"such":[6,69],"as":[7,70],"ChatGPT":[8],"1":[9],"or":[10,33,51,78],"Gemini":[11],"2":[12],",":[13],"can":[14,146],"show":[15],"impressive":[16],"reasoning":[17],"and":[18,26,58,102,135,171,184,217],"question-answering":[19],"capabilities":[20],"but":[21],"often":[22],"\u2018hallucinate\u2019":[23],"false":[24],"outputs":[25],"unsubstantiated":[27],"answers":[28],"3,4":[29],".":[30,73,86],"Answering":[31],"unreliably":[32],"without":[34,173],"the":[35,111,141,156,178],"necessary":[36],"information":[37],"prevents":[38],"adoption":[39],"in":[40,54,66,95,119,149],"diverse":[41],"fields,":[42],"with":[43,100,215],"problems":[44],"including":[45],"fabrication":[46],"of":[47,131,158,164,177],"legal":[48],"precedents":[49],"5":[50],"untrue":[52],"facts":[53],"news":[55],"articles":[56],"6":[57],"even":[59,99],"posing":[60],"a":[61,89,129,174,196,202],"risk":[62],"to":[63,105,127,187,200],"human":[64],"life":[65],"medical":[67],"domains":[68],"radiology":[71],"7":[72],"Encouraging":[74],"truthfulness":[75],"through":[76],"supervision":[77],"reinforcement":[79],"has":[80],"been":[81],"only":[82],"partially":[83],"successful":[84],"8":[85],"Researchers":[87],"need":[88],"general":[90],"method":[91,139,167,205],"for":[92,125,222],"detecting":[93,194],"hallucinations":[94],"LLMs":[96,126,216,224],"that":[97,143,225],"works":[98,168],"new":[101,116,188,220],"unseen":[103],"questions":[104],"which":[106],"humans":[107],"might":[108],"not":[109,190],"know":[110],"answer.":[112],"Here":[113],"we":[114],"develop":[115],"methods":[117],"grounded":[118],"statistics,":[120],"proposing":[121],"entropy-based":[122],"uncertainty":[123,154],"estimators":[124],"detect":[128],"subset":[130],"hallucinations\u2014confabulations\u2014which":[132],"are":[133,226],"arbitrary":[134],"incorrect":[136],"generations.":[137],"Our":[138,166],"addresses":[140],"fact":[142],"one":[144],"idea":[145],"be":[147],"expressed":[148],"many":[150],"ways":[151],"by":[152,229],"computing":[153],"at":[155],"level":[157],"meaning":[159],"rather":[160],"than":[161],"specific":[162],"sequences":[163],"words.":[165],"across":[169],"datasets":[170],"tasks":[172,189],"priori":[175],"knowledge":[176],"task,":[179],"requires":[180],"no":[181],"task-specific":[182],"data":[183],"robustly":[185],"generalizes":[186],"seen":[191],"before.":[192],"By":[193],"when":[195,209],"prompt":[197],"is":[198],"likely":[199],"produce":[201],"confabulation,":[203],"our":[204],"helps":[206],"users":[207],"understand":[208],"they":[210],"must":[211],"take":[212],"extra":[213],"care":[214],"opens":[218],"up":[219],"possibilities":[221],"using":[223],"otherwise":[227],"prevented":[228],"their":[230],"unreliability.":[231]},"counts_by_year":[{"year":2026,"cited_by_count":291},{"year":2025,"cited_by_count":354},{"year":2024,"cited_by_count":51},{"year":2021,"cited_by_count":1}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2025-10-10T00:00:00"}