{"id":"https://openalex.org/W4388691884","doi":"https://doi.org/10.1162/coli_a_00493","title":"Analyzing Semantic Faithfulness of Language Models via Input Intervention on Question Answering","display_name":"Analyzing Semantic Faithfulness of Language Models via Input Intervention on Question Answering","publication_year":2023,"publication_date":"2023-11-15","ids":{"openalex":"https://openalex.org/W4388691884","doi":"https://doi.org/10.1162/coli_a_00493"},"language":"en","primary_location":{"id":"doi:10.1162/coli_a_00493","is_oa":true,"landing_page_url":"https://doi.org/10.1162/coli_a_00493","pdf_url":"https://direct.mit.edu/coli/article-pdf/doi/10.1162/coli_a_00493/2177320/coli_a_00493.pdf","source":{"id":"https://openalex.org/S155526855","display_name":"Computational Linguistics","issn_l":"0891-2017","issn":["0891-2017","1530-9312"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Computational Linguistics","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://direct.mit.edu/coli/article-pdf/doi/10.1162/coli_a_00493/2177320/coli_a_00493.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5019757348","display_name":"Akshay Chaturvedi","orcid":"https://orcid.org/0000-0002-8192-6221"},"institutions":[{"id":"https://openalex.org/I134560555","display_name":"Universit\u00e9 Toulouse III - Paul Sabatier","ror":"https://ror.org/02v6kpv12","country_code":"FR","type":"education","lineage":["https://openalex.org/I134560555"]},{"id":"https://openalex.org/I4210119061","display_name":"Institut de Recherche en Informatique de Toulouse","ror":"https://ror.org/01rx4qw44","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I205747304","https://openalex.org/I4210119061","https://openalex.org/I4210152422","https://openalex.org/I4405258862","https://openalex.org/I4405259414"]}],"countries":["FR"],"is_corresponding":true,"raw_author_name":"Akshay Chaturvedi","raw_affiliation_strings":["IRIT, Universit\u00e9 Paul Sabatier Toulouse, France akshay91.isi@gmail.com"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IRIT, Universit\u00e9 Paul Sabatier Toulouse, France akshay91.isi@gmail.com","institution_ids":["https://openalex.org/I134560555","https://openalex.org/I4210119061"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067158676","display_name":"Swarnadeep Bhar","orcid":null},"institutions":[{"id":"https://openalex.org/I134560555","display_name":"Universit\u00e9 Toulouse III - Paul Sabatier","ror":"https://ror.org/02v6kpv12","country_code":"FR","type":"education","lineage":["https://openalex.org/I134560555"]},{"id":"https://openalex.org/I4210119061","display_name":"Institut de Recherche en Informatique de Toulouse","ror":"https://ror.org/01rx4qw44","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I205747304","https://openalex.org/I4210119061","https://openalex.org/I4210152422","https://openalex.org/I4405258862","https://openalex.org/I4405259414"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Swarnadeep Bhar","raw_affiliation_strings":["IRIT, Universit\u00e9 Paul Sabatier Toulouse, France swarnadeep.bhar@irit.fr"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IRIT, Universit\u00e9 Paul Sabatier Toulouse, France swarnadeep.bhar@irit.fr","institution_ids":["https://openalex.org/I134560555","https://openalex.org/I4210119061"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030585541","display_name":"Soumadeep Saha","orcid":"https://orcid.org/0000-0003-0013-5143"},"institutions":[{"id":"https://openalex.org/I6498739","display_name":"Indian Statistical Institute","ror":"https://ror.org/00q2w1j53","country_code":"IN","type":"education","lineage":["https://openalex.org/I6498739"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Soumadeep Saha","raw_affiliation_strings":["Indian Statistical Institute Kolkata, India soumadeep.saha97@gmail.com","Indian Statistical Institute [Kolkata] (No. 203, Barrackpore Trunk Road, Kolkata, West Bengal 700108 - India)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Indian Statistical Institute Kolkata, India soumadeep.saha97@gmail.com","institution_ids":["https://openalex.org/I6498739"]},{"raw_affiliation_string":"Indian Statistical Institute [Kolkata] (No. 203, Barrackpore Trunk Road, Kolkata, West Bengal 700108 - India)","institution_ids":["https://openalex.org/I6498739"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024170958","display_name":"Utpal Garain","orcid":"https://orcid.org/0000-0001-7207-5018"},"institutions":[{"id":"https://openalex.org/I6498739","display_name":"Indian Statistical Institute","ror":"https://ror.org/00q2w1j53","country_code":"IN","type":"education","lineage":["https://openalex.org/I6498739"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Utpal Garain","raw_affiliation_strings":["Indian Statistical Institute Kolkata, India utpal@isical.ac.in","Indian Statistical Institute [Kolkata] (No. 203, Barrackpore Trunk Road, Kolkata, West Bengal 700108 - India)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Indian Statistical Institute Kolkata, India utpal@isical.ac.in","institution_ids":["https://openalex.org/I6498739"]},{"raw_affiliation_string":"Indian Statistical Institute [Kolkata] (No. 203, Barrackpore Trunk Road, Kolkata, West Bengal 700108 - India)","institution_ids":["https://openalex.org/I6498739"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5019863253","display_name":"Nicholas Asher","orcid":"https://orcid.org/0000-0002-7689-8246"},"institutions":[{"id":"https://openalex.org/I134560555","display_name":"Universit\u00e9 Toulouse III - Paul Sabatier","ror":"https://ror.org/02v6kpv12","country_code":"FR","type":"education","lineage":["https://openalex.org/I134560555"]},{"id":"https://openalex.org/I4210119061","display_name":"Institut de Recherche en Informatique de Toulouse","ror":"https://ror.org/01rx4qw44","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I205747304","https://openalex.org/I4210119061","https://openalex.org/I4210152422","https://openalex.org/I4405258862","https://openalex.org/I4405259414"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Nicholas Asher","raw_affiliation_strings":["IRIT, Universit\u00e9 Paul Sabatier Toulouse, France nicholas.asher@irit.fr"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"IRIT, Universit\u00e9 Paul Sabatier Toulouse, France nicholas.asher@irit.fr","institution_ids":["https://openalex.org/I134560555","https://openalex.org/I4210119061"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5019757348"],"corresponding_institution_ids":["https://openalex.org/I134560555","https://openalex.org/I4210119061"],"apc_list":null,"apc_paid":null,"fwci":0.3954,"has_fulltext":true,"cited_by_count":3,"citation_normalized_percentile":{"value":0.70428641,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":97},"biblio":{"volume":"50","issue":"1","first_page":"119","last_page":"155"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9998999834060669,"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":0.9998999834060669,"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":0.9984999895095825,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9807000160217285,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.6918975710868835},{"id":"https://openalex.org/keywords/negation","display_name":"Negation","score":0.6844673156738281},{"id":"https://openalex.org/keywords/predicate","display_name":"Predicate (mathematical logic)","score":0.6539375185966492},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.6505517959594727},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.6234028339385986},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.5355206727981567},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5115953087806702},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.511557400226593},{"id":"https://openalex.org/keywords/argument","display_name":"Argument (complex analysis)","score":0.4737711250782013},{"id":"https://openalex.org/keywords/semantic-role-labeling","display_name":"Semantic role labeling","score":0.46033698320388794},{"id":"https://openalex.org/keywords/psychological-intervention","display_name":"Psychological intervention","score":0.4585217833518982},{"id":"https://openalex.org/keywords/linguistics","display_name":"Linguistics","score":0.329107403755188},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.19427835941314697},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.08040454983711243},{"id":"https://openalex.org/keywords/medicine","display_name":"Medicine","score":0.0785650908946991}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6918975710868835},{"id":"https://openalex.org/C2185349","wikidata":"https://www.wikidata.org/wiki/Q190558","display_name":"Negation","level":2,"score":0.6844673156738281},{"id":"https://openalex.org/C140146324","wikidata":"https://www.wikidata.org/wiki/Q1144319","display_name":"Predicate (mathematical logic)","level":2,"score":0.6539375185966492},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.6505517959594727},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.6234028339385986},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5355206727981567},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5115953087806702},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.511557400226593},{"id":"https://openalex.org/C98184364","wikidata":"https://www.wikidata.org/wiki/Q1780131","display_name":"Argument (complex analysis)","level":2,"score":0.4737711250782013},{"id":"https://openalex.org/C67277372","wikidata":"https://www.wikidata.org/wiki/Q7449085","display_name":"Semantic role labeling","level":3,"score":0.46033698320388794},{"id":"https://openalex.org/C27415008","wikidata":"https://www.wikidata.org/wiki/Q7256382","display_name":"Psychological intervention","level":2,"score":0.4585217833518982},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.329107403755188},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.19427835941314697},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.08040454983711243},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.0785650908946991},{"id":"https://openalex.org/C2777530160","wikidata":"https://www.wikidata.org/wiki/Q41796","display_name":"Sentence","level":2,"score":0.0},{"id":"https://openalex.org/C118552586","wikidata":"https://www.wikidata.org/wiki/Q7867","display_name":"Psychiatry","level":1,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C126322002","wikidata":"https://www.wikidata.org/wiki/Q11180","display_name":"Internal medicine","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.1162/coli_a_00493","is_oa":true,"landing_page_url":"https://doi.org/10.1162/coli_a_00493","pdf_url":"https://direct.mit.edu/coli/article-pdf/doi/10.1162/coli_a_00493/2177320/coli_a_00493.pdf","source":{"id":"https://openalex.org/S155526855","display_name":"Computational Linguistics","issn_l":"0891-2017","issn":["0891-2017","1530-9312"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Computational Linguistics","raw_type":"journal-article"},{"id":"pmh:oai:HAL:hal-04829090v1","is_oa":true,"landing_page_url":"https://hal.science/hal-04829090","pdf_url":"https://hal.science/hal-04829090v1/document","source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ISSN: 0891-2017","raw_type":"info:eu-repo/semantics/article"},{"id":"pmh:oai:HAL:hal-04344736v1","is_oa":true,"landing_page_url":"https://hal.science/hal-04344736","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"https://direct.mit.edu/coli/article/doi/10.1162/coli_a_00493/118135/Analyzing-Semantic-Faithfulness-of-Language-Models","raw_type":"info:eu-repo/semantics/article"},{"id":"pmh:oai:doaj.org/article:ac32178958144801b323c7fca98f5499","is_oa":true,"landing_page_url":"https://doaj.org/article/ac32178958144801b323c7fca98f5499","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Computational Linguistics, Vol 50, Iss 1 (2024)","raw_type":"article"}],"best_oa_location":{"id":"doi:10.1162/coli_a_00493","is_oa":true,"landing_page_url":"https://doi.org/10.1162/coli_a_00493","pdf_url":"https://direct.mit.edu/coli/article-pdf/doi/10.1162/coli_a_00493/2177320/coli_a_00493.pdf","source":{"id":"https://openalex.org/S155526855","display_name":"Computational Linguistics","issn_l":"0891-2017","issn":["0891-2017","1530-9312"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by-nc-nd","license_id":"https://openalex.org/licenses/cc-by-nc-nd","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Computational Linguistics","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1938070154","display_name":"Artificial and Natural Intelligence Toulouse Institute","funder_award_id":"ANR-19-P3IA-0004","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"},{"id":"https://openalex.org/G3010527862","display_name":null,"funder_award_id":"ANR\u201019","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"},{"id":"https://openalex.org/G5430966381","display_name":"Natural Language Programming for Conversational Cobots","funder_award_id":"ANR-21-FAI2-0005","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"},{"id":"https://openalex.org/G6176599885","display_name":"Structured, Consolidated and Unified Interactive Dialogue for Robot Task Performance","funder_award_id":"ANR-21-ASIA-0005","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"},{"id":"https://openalex.org/G6769792828","display_name":null,"funder_award_id":"6702-2","funder_id":"https://openalex.org/F4320321027","funder_display_name":"Indo-French Centre for the Promotion of Advanced Research"},{"id":"https://openalex.org/G8840324868","display_name":null,"funder_award_id":"SPR/2020/000495","funder_id":"https://openalex.org/F4320320719","funder_display_name":"Department of Science and Technology, Ministry of Science and Technology, India"}],"funders":[{"id":"https://openalex.org/F4320320719","display_name":"Department of Science and Technology, Ministry of Science and Technology, India","ror":"https://ror.org/0101xrq71"},{"id":"https://openalex.org/F4320320883","display_name":"Agence Nationale de la Recherche","ror":"https://ror.org/00rbzpz17"},{"id":"https://openalex.org/F4320321027","display_name":"Indo-French Centre for the Promotion of Advanced Research","ror":"https://ror.org/017aem598"},{"id":"https://openalex.org/F4320328356","display_name":"R\u00e9gion Occitanie Pyr\u00e9n\u00e9es-M\u00e9diterran\u00e9e","ror":null},{"id":"https://openalex.org/F4320334771","display_name":"Science and Engineering Research Board","ror":"https://ror.org/03ffdsr55"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4388691884.pdf","grobid_xml":"https://content.openalex.org/works/W4388691884.grobid-xml"},"referenced_works_count":74,"referenced_works":["https://openalex.org/W114519366","https://openalex.org/W128386071","https://openalex.org/W651567163","https://openalex.org/W1495022714","https://openalex.org/W2085574295","https://openalex.org/W2125960523","https://openalex.org/W2137831583","https://openalex.org/W2473186245","https://openalex.org/W2612690371","https://openalex.org/W2889787757","https://openalex.org/W2946359678","https://openalex.org/W2946417913","https://openalex.org/W2963969878","https://openalex.org/W2964223283","https://openalex.org/W2964303116","https://openalex.org/W2965373594","https://openalex.org/W2970742161","https://openalex.org/W2970862333","https://openalex.org/W2975291987","https://openalex.org/W2979826702","https://openalex.org/W2990138404","https://openalex.org/W2995920266","https://openalex.org/W3011279327","https://openalex.org/W3015001695","https://openalex.org/W3034995113","https://openalex.org/W3035137491","https://openalex.org/W3035305735","https://openalex.org/W3037636427","https://openalex.org/W3044438666","https://openalex.org/W3099843385","https://openalex.org/W3102659883","https://openalex.org/W3105928338","https://openalex.org/W3106229690","https://openalex.org/W3111372685","https://openalex.org/W3113484723","https://openalex.org/W3118485687","https://openalex.org/W3156012351","https://openalex.org/W3159900299","https://openalex.org/W3161374759","https://openalex.org/W3169306793","https://openalex.org/W3175987672","https://openalex.org/W3177813494","https://openalex.org/W3185341429","https://openalex.org/W3193162699","https://openalex.org/W3202712981","https://openalex.org/W4205927329","https://openalex.org/W4211142623","https://openalex.org/W4226226396","https://openalex.org/W4236828747","https://openalex.org/W4248023279","https://openalex.org/W4250738323","https://openalex.org/W4252757565","https://openalex.org/W4254941661","https://openalex.org/W4283705032","https://openalex.org/W4285192810","https://openalex.org/W4285225959","https://openalex.org/W4288019212","https://openalex.org/W4311824882","https://openalex.org/W4365799947","https://openalex.org/W4378189609","https://openalex.org/W6675453446","https://openalex.org/W6744110554","https://openalex.org/W6746141323","https://openalex.org/W6751744669","https://openalex.org/W6755207826","https://openalex.org/W6757635932","https://openalex.org/W6763701032","https://openalex.org/W6766673545","https://openalex.org/W6768106268","https://openalex.org/W6768299147","https://openalex.org/W6774862694","https://openalex.org/W6798182279","https://openalex.org/W6810738896","https://openalex.org/W6849590751"],"related_works":["https://openalex.org/W2168154661","https://openalex.org/W4288267738","https://openalex.org/W2964413124","https://openalex.org/W4388937922","https://openalex.org/W3113264705","https://openalex.org/W2121717955","https://openalex.org/W1990677253","https://openalex.org/W2348729074","https://openalex.org/W2567185627","https://openalex.org/W2386406349"],"abstract_inverted_index":{"Abstract":[0],"Transformer-based":[1],"language":[2],"models":[3,101,176,241],"have":[4],"been":[5],"shown":[6],"to":[7,42,115,167,177,207,213,229,235,253,256],"be":[8,116],"highly":[9],"effective":[10],"for":[11,124,132,141,157,185,226],"several":[12],"NLP":[13],"tasks.":[14],"In":[15],"this":[16,76,192],"article,":[17],"we":[18,110,120,189],"consider":[19],"three":[20],"transformer":[21,100],"models,":[22],"BERT,":[23],"RoBERTa,":[24],"and":[25,30,33,96,135,234,259],"XLNet,":[26],"in":[27,55,66,70,139],"both":[28],"small":[29],"large":[31],"versions,":[32],"investigate":[34],"how":[35],"faithful":[36,118],"their":[37],"representations":[38],"are":[39],"with":[40,209],"respect":[41],"the":[43,57,154,172,175,180,204,215,231],"semantic":[44,53,58,93,200],"content":[45,59],"of":[46,52,60,128,174,182,199,218],"texts.":[47,219],"We":[48,73,144,170,220],"formalize":[49],"a":[50,61,67,80,87,125,161],"notion":[51,77],"faithfulness,":[54],"which":[56],"text":[62],"should":[63],"causally":[64],"figure":[65],"model\u2019s":[68,81],"inferences":[69],"question":[71,107],"answering.":[72],"then":[74,145],"test":[75,222],"by":[78,160],"observing":[79],"behavior":[82],"on":[83,105,247],"answering":[84,108],"questions":[85],"about":[86],"story":[88],"after":[89],"performing":[90],"two":[91,232],"novel":[92],"interventions\u2014deletion":[94],"intervention":[95,159,211],"negation":[97,142,210,260],"intervention.":[98,187],"While":[99,239],"achieve":[102,243],"high":[103,245],"performance":[104,246],"standard":[106],"tasks,":[109],"show":[111,190],"that":[112,151,191],"they":[113,251],"fail":[114,252],"semantically":[117],"once":[119],"perform":[121],"these":[122],"interventions":[123,233],"significant":[126,162],"number":[127],"cases":[129],"(\u223c":[130],"50%":[131,166],"deletion":[133,158,186,258],"intervention,":[134],"\u223c":[136,165,168],"20%":[137],"drop":[138],"accuracy":[140],"intervention).":[143],"propose":[146],"an":[147],"intervention-based":[148,183],"training":[149,184,193],"regime":[150],"can":[152],"mitigate":[153],"undesirable":[155],"effects":[156],"margin":[163],"(from":[164],"6%).":[169],"analyze":[171],"inner-workings":[173],"better":[178],"understand":[179],"effectiveness":[181],"But":[188],"does":[194],"not":[195],"attenuate":[196],"other":[197],"aspects":[198],"unfaithfulness":[201],"such":[202],"as":[203],"models\u2019":[205],"inability":[206],"deal":[208],"or":[212],"capture":[214,236],"predicate\u2013argument":[216,237,248],"structure":[217,249],"also":[221],"InstructGPT,":[223],"via":[224],"prompting,":[225],"its":[227],"ability":[228],"handle":[230],"structure.":[238],"InstructGPT":[240],"do":[242],"very":[244],"task,":[250],"respond":[254],"adequately":[255],"our":[257],"interventions.":[261]},"counts_by_year":[{"year":2025,"cited_by_count":3}],"updated_date":"2026-08-26T07:47:46.906454","created_date":"2023-11-16T00:00:00"}