urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64:0 SND doi-10-23695-v3wp-6c64 0 urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.ResourcePackage:2.0 urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.OtherMaterialScheme:2.0 urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.OrganizationScheme-0:2.0 urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.Individual-0:2.0 Södahl Bladsjö, Tom urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.Individual-0:2.0 Muñoz Sánchez, Ricardo urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.StudyUnit:2.0 doi-10-23695-v3wp-6c64 MARB MARB urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.Individual-0:2.0 Individual Göteborgs universitet University of Gothenburg Göteborgs universitet University of Gothenburg 2024-05-27 10.23695/V3WP-6C64 DOI Reporting bias (the human tendency to not mention obvious or redundant information) and social bias (societal attitudes toward specific demographic groups) have both been shown to propagate from human text data to language models trained on such data. However, the two phenomena have not previously been studied in combination. The MARB dataset was developed to begin to fill this gap by studying the interaction between social biases and reporting bias in language models. Unlike many existing benchmark datasets, MARB does not rely on artificially constructed templates or crowdworkers to create contrasting examples. Instead, the templates used in MARB are based on naturally occurring written language from the 2021 version of the enTenTen corpus (Jakubíček et al., 2013). Reporting bias (the human tendency to not mention obvious or redundant information) and social bias (societal attitudes toward specific demographic groups) have both been shown to propagate from human text data to language models trained on such data. However, the two phenomena have not previously been studied in combination. The MARB dataset was developed to begin to fill this gap by studying the interaction between social biases and reporting bias in language models. Unlike many existing benchmark datasets, MARB does not rely on artificially constructed templates or crowdworkers to create contrasting examples. Instead, the templates used in MARB are based on naturally occurring written language from the 2021 version of the enTenTen corpus (Jakubíček et al., 2013). urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.TopicalCoverage:2.0 Gender Studies Genusstudier Natural Language Processing Språkbehandling och datorlingvistik urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.Archive:2.0 urn:ddi:se.researchdata:doi-10-23695-v3wp-6c64.Archive-ArchiveSpecificType-AccessType:2.0 0