Language selection

Government of Canada / Gouvernement du Canada

Search

This prompt is measuring <mask>: evaluating bias evaluation in language models

From National Research Council Canada

Download
  1. (PDF, 1.3 MiB)
DOIResolve DOI: https://doi.org/10.18653/v1/2023.findings-acl.139
AuthorSearch for: ; Search for: ; Search for: 1; Search for:
Affiliation
  1. National Research Council of Canada. Digital Technologies
FormatText, Article
Conference61st Annual Meeting of the Association for Computational Linguistics, ACL 2023, July 9-14, 2023, Toronto, ON, Canada
Abstract
Bias research in NLP seeks to analyse models for social biases, thus helping NLP practitioners uncover, measure, and mitigate social harms. We analyse the body of work that uses prompts and templates to assess bias in language models. We draw on a measurement modelling framework to create a taxonomy of attributes that capture what a bias test aims to measure and how that measurement is carried out. By applying this taxonomy to 90 bias tests, we illustrate qualitatively and quantitatively that core aspects of bias test conceptualisations and operationalisations are frequently unstated or ambiguous, carry implicit assumptions, or be mismatched. Our analysis illuminates the scope of possible bias types the field is able to measure, and reveals types that are as yet under-researched. We offer guidance to enable the community to explore a wider section of the possible bias space, and to better close the gap between desired outcomes and experimental design, both for bias and for evaluating language models more broadly.
Publication date
PublisherAssociation for Computational Linguistics
Licence
In
Series
  • Proceedings of the Conference.
LanguageEnglish
Peer reviewedYes
Export citationExport as RIS
Report a correctionReport a correction (opens in a new tab)
Record identifier6848018a-b83d-42ca-9f95-0a75ab64e717
Record created2024-12-06
Record modified2024-12-06
Date modified: