{"id":"https://openalex.org/W4415016237","doi":"https://doi.org/10.48550/arxiv.2504.14150","title":"Walk the Talk? Measuring the Faithfulness of Large Language Model Explanations","display_name":"Walk the Talk? Measuring the Faithfulness of Large Language Model Explanations","publication_year":2025,"publication_date":"2025-04-19","ids":{"openalex":"https://openalex.org/W4415016237","doi":"https://doi.org/10.48550/arxiv.2504.14150"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2504.14150","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.14150","pdf_url":"https://arxiv.org/pdf/2504.14150","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/2504.14150","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5099111284","display_name":"Katie Matton","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I63966007","display_name":"Massachusetts Institute of Technology","ror":"https://ror.org/042nb2s44","country_code":"US","type":"education","lineage":["https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Matton, Katie","raw_affiliation_strings":["MIT"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT","institution_ids":["https://openalex.org/I63966007"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5008771487","display_name":"Robert Osazuwa Ness","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ness, Robert Osazuwa","raw_affiliation_strings":["Microsoft Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007282049","display_name":"John V. Guttag","orcid":"https://orcid.org/0000-0003-0992-0906","observed_orcids":["https://orcid.org/0000-0003-0992-0906"]},"institutions":[{"id":"https://openalex.org/I63966007","display_name":"Massachusetts Institute of Technology","ror":"https://ror.org/042nb2s44","country_code":"US","type":"education","lineage":["https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Guttag, John","raw_affiliation_strings":["MIT"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT","institution_ids":["https://openalex.org/I63966007"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5079458476","display_name":"Emre K\u0131c\u0131man","orcid":"https://orcid.org/0000-0001-5429-468X","observed_orcids":["https://orcid.org/0000-0001-5429-468X"]},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"K\u0131c\u0131man, Emre","raw_affiliation_strings":["Microsoft Research"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research","institution_ids":["https://openalex.org/I1290206253"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":30,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":1.0,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":1.0,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.0,"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":0.0,"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/large-language-models","display_name":"large language models","score":0.7289999723434448},{"id":"https://openalex.org/keywords/explanation-faithfulness","display_name":"explanation fidelity","score":0.9340000152587891},{"id":"https://openalex.org/keywords/evaluation-of-explanations","display_name":"explanation evaluation","score":0.3970000147819519},{"id":"https://openalex.org/keywords/counterfactual-thinking","display_name":"counterfactual reasoning","score":0.7319999933242798},{"id":"https://openalex.org/keywords/bayesian-hierarchical-model","display_name":"Bayesian hierarchical model","score":0.8539999723434448},{"id":"https://openalex.org/keywords/social-bias","display_name":"social bias","score":0.6869999766349792},{"id":"https://openalex.org/keywords/medical-question-answering","display_name":"medical question answering","score":0.9559999704360962},{"id":"https://openalex.org/keywords/interpretability","display_name":"interpretability","score":0.5730000138282776}],"concepts":[{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.6424999833106995},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6122000217437744},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4787999987602234},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4652000069618225},{"id":"https://openalex.org/C11671645","wikidata":"https://www.wikidata.org/wiki/Q5054567","display_name":"Causal model","level":2,"score":0.44830000400543213},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.43959999084472656},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.3573000133037567},{"id":"https://openalex.org/C64357122","wikidata":"https://www.wikidata.org/wiki/Q1149766","display_name":"Causality (physics)","level":2,"score":0.34549999237060547},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.34459999203681946},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3192000091075897},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.31709998846054077},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.3147999942302704},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.30090001225471497},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2775999903678894},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2524000108242035}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2504.14150","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.14150","pdf_url":"https://arxiv.org/pdf/2504.14150","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.2504.14150","doi":"https://doi.org/10.48550/arxiv.2504.14150","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2504.14150","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:2504.14150","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2504.14150","pdf_url":"https://arxiv.org/pdf/2504.14150","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":[{"id":"https://openalex.org/G9000874036","display_name":"Clinical Research Center for the Improved Prevention, Diagnosis, and Treatment of Vocal Hyperfunction","funder_award_id":"P50DC015446","funder_id":"https://openalex.org/F4320337352","funder_display_name":"National Institute on Deafness and Other Communication Disorders"}],"funders":[{"id":"https://openalex.org/F4320332161","display_name":"National Institutes of Health","ror":"https://ror.org/01cwqze88"},{"id":"https://openalex.org/F4320337352","display_name":"National Institute on Deafness and Other Communication Disorders","ror":"https://ror.org/04mhx6838"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4415016237.pdf","grobid_xml":"https://content.openalex.org/works/W4415016237.grobid-xml"},"referenced_works_count":16,"referenced_works":["https://openalex.org/W561663288","https://openalex.org/W2143891888","https://openalex.org/W2473286717","https://openalex.org/W2934842096","https://openalex.org/W2948140294","https://openalex.org/W2988346595","https://openalex.org/W3162922479","https://openalex.org/W3207316473","https://openalex.org/W4385291106","https://openalex.org/W4385327568","https://openalex.org/W4385571986","https://openalex.org/W4387799763","https://openalex.org/W4392637287","https://openalex.org/W4393946427","https://openalex.org/W4402671681","https://openalex.org/W4404781621"],"related_works":["https://openalex.org/W4391673068","https://openalex.org/W7102412901","https://openalex.org/W4402683840","https://openalex.org/W4402683961","https://openalex.org/W7211919242","https://openalex.org/W7154685440","https://openalex.org/W4416365771","https://openalex.org/W4416036619","https://openalex.org/W7166888719","https://openalex.org/W4403752369"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2],"(LLMs)":[3],"are":[4,100],"capable":[5],"of":[6,10,52,61,88,94,130,151,174,190,211],"generating":[7],"plausible":[8],"explanations":[9,22,65,98,186,204],"how":[11],"they":[12,30,69],"arrived":[13],"at":[14,153],"an":[15,123],"answer":[16],"to":[17,39,126,135,146,168],"a":[18,45,58,111,142,177,194],"question.":[19],"However,":[20],"these":[21],"can":[23,31,37,165],"misrepresent":[24],"the":[25,50,75,81,89,92,103,128,148,155,188,214],"model's":[26,215],"\"reasoning\"":[27],"process,":[28],"i.e.,":[29],"be":[32,166],"unfaithful.":[33],"This,":[34],"in":[35,74,86],"turn,":[36],"lead":[38],"over-trust":[40],"and":[41,102,139,157,170],"misuse.":[42],"We":[43,83],"introduce":[44],"new":[46],"approach":[47],"for":[48,114],"measuring":[49],"faithfulness":[51,85,116],"LLM":[53,64,97,125,185,203],"explanations.":[54],"First,":[55],"we":[56,109,181,199],"provide":[57,205],"rigorous":[59],"definition":[60],"faithfulness.":[62],"Since":[63],"mimic":[66],"human":[67],"explanations,":[68],"often":[70],"reference":[71],"high-level":[72],"concepts":[73,95,131,152],"input":[76],"question":[77,196],"that":[78,96,105,117,162],"purportedly":[79],"influenced":[80,213],"model.":[82],"define":[84],"terms":[87],"difference":[90],"between":[91],"set":[93,104],"imply":[99],"influential":[101],"truly":[106],"are.":[107],"Second,":[108],"present":[110],"novel":[112],"method":[113,164],"estimating":[115],"is":[118],"based":[119],"on:":[120],"(1)":[121],"using":[122,141],"auxiliary":[124],"modify":[127],"values":[129],"within":[132],"model":[133,145],"inputs":[134],"create":[136],"realistic":[137],"counterfactuals,":[138],"(2)":[140],"Bayesian":[143],"hierarchical":[144],"quantify":[147,169],"causal":[149],"effects":[150],"both":[154],"example-":[156],"dataset-level.":[158],"Our":[159],"experiments":[160],"show":[161],"our":[163],"used":[167],"discover":[171],"interpretable":[172],"patterns":[173],"unfaithfulness.":[175],"On":[176,193],"social":[178,191],"bias":[179],"task,":[180,198],"uncover":[182,200],"cases":[183,201],"where":[184,202],"hide":[187],"influence":[189],"bias.":[192],"medical":[195],"answering":[197],"misleading":[206],"claims":[207],"about":[208],"which":[209],"pieces":[210],"evidence":[212],"decisions.":[216]},"counts_by_year":[{"year":2026,"cited_by_count":24},{"year":2025,"cited_by_count":6}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-10T00:00:00"}