{"id":"https://openalex.org/W7225734016","doi":"https://doi.org/10.21203/rs.3.rs-11610550/v1","title":"Auditing the Auditor: Extraction Accuracy, Hallucination, and Abstention in Open-Weight LLMs for ESG Reporting A Longitudinal Empirical Benchmark Across 518 Corporate Disclosures (2010\u20132026)","display_name":"Auditing the Auditor: Extraction Accuracy, Hallucination, and Abstention in Open-Weight LLMs for ESG Reporting A Longitudinal Empirical Benchmark Across 518 Corporate Disclosures (2010\u20132026)","publication_year":2026,"publication_date":"2026-11-10","ids":{"openalex":"https://openalex.org/W7225734016","doi":"https://doi.org/10.21203/rs.3.rs-11610550/v1"},"language":"en","primary_location":{"id":"doi:10.21203/rs.3.rs-11610550/v1","doi":"https://doi.org/10.21203/rs.3.rs-11610550/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11610550/v1","pdf_url":"https://www.researchsquare.com/article/rs-11610550/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.researchsquare.com/article/rs-11610550/latest.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5158513863","display_name":"Redeemer Okekale Salami","orcid":"https://orcid.org/0009-0000-2900-592X","observed_orcids":["https://orcid.org/0009-0000-2900-592X"]},"institutions":[{"id":"https://openalex.org/I138690464","display_name":"University of Ghana","ror":"https://ror.org/01r22mr83","country_code":"GH","type":"education","lineage":["https://openalex.org/I138690464"]}],"countries":["GH"],"is_corresponding":false,"raw_author_name":"Redeemer Salami Okekale","raw_affiliation_strings":["University of Ghana"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Ghana","institution_ids":["https://openalex.org/I138690464"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093260644","display_name":"Kate Markin","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I138690464","display_name":"University of Ghana","ror":"https://ror.org/01r22mr83","country_code":"GH","type":"education","lineage":["https://openalex.org/I138690464"]}],"countries":["GH"],"is_corresponding":false,"raw_author_name":"Kate Markin","raw_affiliation_strings":["University of Ghana"],"raw_orcid":"https://orcid.org/0009-0004-3037-8888","affiliations":[{"raw_affiliation_string":"University of Ghana","institution_ids":["https://openalex.org/I138690464"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160558466","display_name":"Derek Amuna","orcid":"https://orcid.org/0009-0003-1889-7602","observed_orcids":["https://orcid.org/0009-0003-1889-7602"]},"institutions":[{"id":"https://openalex.org/I138690464","display_name":"University of Ghana","ror":"https://ror.org/01r22mr83","country_code":"GH","type":"education","lineage":["https://openalex.org/I138690464"]}],"countries":["GH"],"is_corresponding":false,"raw_author_name":"Derek Amuna","raw_affiliation_strings":["University of Ghana"],"raw_orcid":"https://orcid.org/0009-0003-1889-7602","affiliations":[{"raw_affiliation_string":"University of Ghana","institution_ids":["https://openalex.org/I138690464"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5161413323","display_name":"Felix Coker","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I138690464","display_name":"University of Ghana","ror":"https://ror.org/01r22mr83","country_code":"GH","type":"education","lineage":["https://openalex.org/I138690464"]}],"countries":["GH"],"is_corresponding":false,"raw_author_name":"Felix Coker","raw_affiliation_strings":["University of Ghana"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Ghana","institution_ids":["https://openalex.org/I138690464"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I138690464"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10028","display_name":"Topic Modeling","score":0.8554999828338623,"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":0.8554999828338623,"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/T10115","display_name":"Corporate Social Responsibility Reporting","score":0.05090000107884407,"subfield":{"id":"https://openalex.org/subfields/1408","display_name":"Strategy and Management"},"field":{"id":"https://openalex.org/fields/14","display_name":"Business, Management and Accounting"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T13910","display_name":"Computational and Text Analysis Methods","score":0.03539999946951866,"subfield":{"id":"https://openalex.org/subfields/3300","display_name":"General Social Sciences"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[],"concepts":[{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.7803000211715698},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.7476999759674072},{"id":"https://openalex.org/C39389867","wikidata":"https://www.wikidata.org/wiki/Q380767","display_name":"Corporate governance","level":2,"score":0.6478999853134155},{"id":"https://openalex.org/C121955636","wikidata":"https://www.wikidata.org/wiki/Q4116214","display_name":"Accounting","level":1,"score":0.6171000003814697},{"id":"https://openalex.org/C66204764","wikidata":"https://www.wikidata.org/wiki/Q219416","display_name":"Sustainability","level":2,"score":0.5770000219345093},{"id":"https://openalex.org/C153701036","wikidata":"https://www.wikidata.org/wiki/Q659974","display_name":"Trustworthiness","level":2,"score":0.5720999836921692},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.4408000111579895},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.43470001220703125},{"id":"https://openalex.org/C162118730","wikidata":"https://www.wikidata.org/wiki/Q1128453","display_name":"Actuarial science","level":1,"score":0.41350001096725464},{"id":"https://openalex.org/C166052673","wikidata":"https://www.wikidata.org/wiki/Q83021","display_name":"Empirical evidence","level":2,"score":0.39329999685287476},{"id":"https://openalex.org/C141261163","wikidata":"https://www.wikidata.org/wiki/Q218600","display_name":"Corporate social responsibility","level":2,"score":0.3278999924659729},{"id":"https://openalex.org/C2778253041","wikidata":"https://www.wikidata.org/wiki/Q699735","display_name":"Annual report","level":2,"score":0.32199999690055847},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.32109999656677246},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3188000023365021},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.3165999948978424},{"id":"https://openalex.org/C2781460075","wikidata":"https://www.wikidata.org/wiki/Q1399332","display_name":"Compliance (psychology)","level":2,"score":0.29919999837875366},{"id":"https://openalex.org/C41282006","wikidata":"https://www.wikidata.org/wiki/Q830689","display_name":"Auditor's report","level":3,"score":0.28360000252723694},{"id":"https://openalex.org/C2778012447","wikidata":"https://www.wikidata.org/wiki/Q1034415","display_name":"Scope (computer science)","level":2,"score":0.2800999879837036},{"id":"https://openalex.org/C199033989","wikidata":"https://www.wikidata.org/wiki/Q1318295","display_name":"Narrative","level":2,"score":0.2736000120639801},{"id":"https://openalex.org/C2780473208","wikidata":"https://www.wikidata.org/wiki/Q1668727","display_name":"Sustainability reporting","level":3,"score":0.2581999897956848}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21203/rs.3.rs-11610550/v1","doi":"https://doi.org/10.21203/rs.3.rs-11610550/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11610550/v1","pdf_url":"https://www.researchsquare.com/article/rs-11610550/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.21203/rs.3.rs-11610550/v1","doi":"https://doi.org/10.21203/rs.3.rs-11610550/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11610550/v1","pdf_url":"https://www.researchsquare.com/article/rs-11610550/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7225734016.pdf","grobid_xml":"https://content.openalex.org/works/W7225734016.grobid-xml"},"referenced_works_count":8,"referenced_works":["https://openalex.org/W4387635896","https://openalex.org/W4389984066","https://openalex.org/W4415538812","https://openalex.org/W4416036405","https://openalex.org/W4416602526","https://openalex.org/W7120088840","https://openalex.org/W7155416622","https://openalex.org/W7213276223"],"related_works":[],"abstract_inverted_index":{"Abstract":[0],"Corporate":[1],"sustainability":[2,67,116],"auditing":[3],"increasingly":[4],"relies":[5],"on":[6],"Large":[7],"Language":[8],"Models":[9],"(LLMs)":[10],"to":[11,27,45,131,195,202,231],"extract":[12,132],"verified":[13],"quantitative":[14],"metrics":[15],"from":[16,39,155,185],"lengthy":[17],"Environmental,":[18],"Social,":[19],"and":[20,62,72,178,215],"Governance":[21],"(ESG)":[22],"reports.":[23],"For":[24],"an":[25,97,164],"LLM":[26],"serve":[28],"as":[29,137],"a":[30,48,78,156,182],"trustworthy":[31],"compliance":[32],"tool,":[33],"it":[34],"must":[35],"pull":[36],"exact":[37,133,152],"numbers":[38,199],"dense":[40,229],"ta-":[41],"bles":[42],"while":[43,222],"refusing":[44],"guess":[46],"when":[47,149,180],"disclosure":[49],"is":[50],"missing.":[51],"We":[52],"evaluate":[53],"five":[54,85],"leading":[55],"open-weight":[56],"LLMs":[57],"(Qwen3-4B,":[58],"DeepSeek-R1-7B,":[59],"GPT-OSS-20B,":[60],"Qwen3.5-27B,":[61],"Gemma-4-31B)":[63],"across":[64,228],"518":[65],"corporate":[66],"reports":[68],"published":[69],"between":[70],"2010":[71],"2026":[73],"(279,862":[74],"text":[75],"passages).":[76],"Using":[77],"stratified":[79],"120-question":[80],"audit":[81,224],"benchmark":[82],"tested":[83],"under":[84],"controlled":[86],"evidence":[87],"conditions":[88],"(3,000":[89],"total":[90],"responses),":[91],"we":[92],"grade":[93],"every":[94],"output":[95],"using":[96],"independent":[98],"referee":[99],"model":[100,105,160],"(Phi-4-14B)":[101],"so":[102],"that":[103,170],"no":[104],"grades":[106],"its":[107],"own":[108],"family\u2019s":[109],"work.":[110],"Three":[111],"findings":[112],"stand":[113],"out":[114],"for":[115],"practi-":[117],"tioners:":[118],"(1)":[119],"models":[120,194,230],"summarize":[121],"narrative":[122],"policies":[123],"reliably":[124],"(F1":[125,144],"accuracy":[126,165],"\u2248":[127,145],"0.729)":[128],"but":[129],"struggle":[130],"numerical":[134],"indicators":[135],"such":[136],"Scope":[138],"1":[139],"emissions":[140],"or":[141,189,197],"water":[142],"withdrawal":[143],"0.335);":[146],"(2)":[147],"even":[148],"handed":[150],"the":[151,158,186,213,219],"gold-standard":[153],"page":[154],"report,":[157],"strongest":[159],"(GPT-":[161],"OSS-20B)":[162],"reaches":[163],"ceiling":[166],"of":[167,204,218],"0.731,":[168],"showing":[169],"table-reading":[171],"errors":[172],"matter":[173],"more":[174],"than":[175],"search":[176],"errors;":[177],"(3)":[179],"given":[181],"\u201cnear-miss\u201d":[183],"passage":[184],"wrong":[187],"year":[188],"company,":[190],"standard":[191],"prompts":[192],"cause":[193],"invent":[196],"substitute":[198],"in":[200,210],"up":[201],"46.7%":[203],"cases.":[205],"By":[206],"contrast,":[207],"step-by-step":[208],"reasoning":[209],"DeepSeek-R1-7B":[211],"catches":[212],"mismatch":[214],"abstains":[216],"77.5%":[217],"time":[220],"unprompted,":[221],"strict":[223],"instructions":[225],"raise":[226],"abstention":[227],"74.2%\u201380.0%.":[232]},"counts_by_year":[],"updated_date":"2026-11-11T12:16:26.678958","created_date":"2026-11-11T00:00:00"}
