{"id":"https://openalex.org/W4281748205","doi":"https://doi.org/10.48550/arxiv.2205.14334","title":"Teaching Models to Express Their Uncertainty in Words","display_name":"Teaching Models to Express Their Uncertainty in Words","publication_year":2022,"publication_date":"2022-05-28","ids":{"openalex":"https://openalex.org/W4281748205","doi":"https://doi.org/10.48550/arxiv.2205.14334"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2205.14334","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2205.14334","pdf_url":"https://arxiv.org/pdf/2205.14334","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":null,"license_id":null,"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/2205.14334","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5007914769","display_name":"Stephanie Lin","orcid":"https://orcid.org/0000-0003-0334-6595","observed_orcids":["https://orcid.org/0000-0003-0334-6595"]},"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":"Lin, Stephanie","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007861777","display_name":"Jacob K. Hilton","orcid":"https://orcid.org/0000-0003-1931-9516","observed_orcids":["https://orcid.org/0000-0003-1931-9516"]},"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":"Hilton, Jacob","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5089789418","display_name":"Owain Evans","orcid":null,"observed_orcids":[]},"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":"Evans, Owain","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I40120149"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":511,"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":0.63919997215271,"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":0.63919997215271,"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/T10028","display_name":"Topic Modeling","score":0.25049999356269836,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.020600000396370888,"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/epistemic-uncertainty","display_name":"epistemic uncertainty","score":0.7730000019073486},{"id":"https://openalex.org/keywords/uncertainty-calibration","display_name":"uncertainty calibration","score":0.6650000214576721},{"id":"https://openalex.org/keywords/gpt-3","display_name":"GPT-3","score":0.9340000152587891},{"id":"https://openalex.org/keywords/distribution-shift","display_name":"distribution shift","score":0.7919999957084656},{"id":"https://openalex.org/keywords/logit","display_name":"logit","score":0.47600001096725464},{"id":"https://openalex.org/keywords/latent-representations","display_name":"latent representations","score":0.6729999780654907},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.5920000076293945}],"concepts":[{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.8084547519683838},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6270586252212524},{"id":"https://openalex.org/C2776608160","wikidata":"https://www.wikidata.org/wiki/Q4785462","display_name":"Natural (archaeology)","level":2,"score":0.5399406552314758},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4881339967250824},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.47656893730163574},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.43963316082954407},{"id":"https://openalex.org/C32230216","wikidata":"https://www.wikidata.org/wiki/Q7882499","display_name":"Uncertainty quantification","level":2,"score":0.42431509494781494},{"id":"https://openalex.org/C137209882","wikidata":"https://www.wikidata.org/wiki/Q1403517","display_name":"Measurement uncertainty","level":2,"score":0.4156966209411621},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4000963568687439},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.391432523727417},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.2998993396759033},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.24887686967849731},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.07056844234466553},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0},{"id":"https://openalex.org/C166957645","wikidata":"https://www.wikidata.org/wiki/Q23498","display_name":"Archaeology","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2205.14334","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2205.14334","pdf_url":"https://arxiv.org/pdf/2205.14334","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2205.14334","doi":"https://doi.org/10.48550/arxiv.2205.14334","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2205.14334","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2205.14334","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2205.14334","pdf_url":"https://arxiv.org/pdf/2205.14334","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4281748205.pdf","grobid_xml":"https://content.openalex.org/works/W4281748205.grobid-xml"},"referenced_works_count":20,"referenced_works":["https://openalex.org/W1933798900","https://openalex.org/W2756073160","https://openalex.org/W2904981516","https://openalex.org/W2969338701","https://openalex.org/W3012209028","https://openalex.org/W3021338988","https://openalex.org/W3035441651","https://openalex.org/W3083410900","https://openalex.org/W3099142828","https://openalex.org/W3130319171","https://openalex.org/W3151462836","https://openalex.org/W3188772518","https://openalex.org/W3189951784","https://openalex.org/W3199958362","https://openalex.org/W3200980294","https://openalex.org/W3201174429","https://openalex.org/W3211582110","https://openalex.org/W3215698779","https://openalex.org/W4224308101","https://openalex.org/W7133225605"],"related_works":["https://openalex.org/W7134281628","https://openalex.org/W6929465362","https://openalex.org/W4403963829","https://openalex.org/W7169862453","https://openalex.org/W7164409435","https://openalex.org/W4416033841","https://openalex.org/W4399554616","https://openalex.org/W4414902044","https://openalex.org/W7036906873","https://openalex.org/W7214989192"],"abstract_inverted_index":{"We":[0,111,140],"show":[1],"that":[2,50,144,155],"a":[3,26,35,85],"GPT-3":[4],"model":[5,22,29,55,86,126],"can":[6],"learn":[7],"to":[8,48,66,90,122,147],"express":[9,91],"uncertainty":[10,67,93,116,123,131,159],"about":[11,94],"its":[12,69,95,161],"own":[13,70,96],"answers":[14,97],"in":[15,68,98,118],"natural":[16,99],"language":[17],"--":[18],"without":[19],"use":[20],"of":[21,37,109,115,130,134],"logits.":[23,127],"When":[24],"given":[25],"question,":[27],"the":[28,82,106,113],"generates":[30],"both":[31],"an":[32],"answer":[33],"and":[34,63],"level":[36],"confidence":[38],"(e.g.":[39],"\"90%":[40],"confidence\"":[41],"or":[42],"\"high":[43],"confidence\").":[44],"These":[45],"levels":[46],"map":[47],"probabilities":[49],"are":[51,132],"well":[52],"calibrated.":[53],"The":[54],"also":[56,141],"remains":[57],"moderately":[58],"calibrated":[59,92],"under":[60,137],"distribution":[61,138],"shift,":[62],"is":[64,81],"sensitive":[65],"answers,":[71],"rather":[72],"than":[73],"imitating":[74],"human":[75],"examples.":[76],"To":[77],"our":[78],"knowledge,":[79],"this":[80],"first":[83],"time":[84],"has":[87],"been":[88],"shown":[89],"language.":[100],"For":[101],"testing":[102],"calibration,":[103],"we":[104],"introduce":[105],"CalibratedMath":[107],"suite":[108],"tasks.":[110],"compare":[112],"calibration":[114,136,149],"expressed":[117],"words":[119],"(\"verbalized":[120],"probability\")":[121],"extracted":[124],"from":[125],"Both":[128],"kinds":[129],"capable":[133],"generalizing":[135],"shift.":[139],"provide":[142],"evidence":[143],"GPT-3's":[145],"ability":[146],"generalize":[148],"depends":[150],"on":[151],"pre-trained":[152],"latent":[153],"representations":[154],"correlate":[156],"with":[157],"epistemic":[158],"over":[160],"answers.":[162]},"counts_by_year":[{"year":2026,"cited_by_count":227},{"year":2025,"cited_by_count":143},{"year":2024,"cited_by_count":84},{"year":2023,"cited_by_count":51},{"year":2022,"cited_by_count":6}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2022-06-13T00:00:00"}