Measuring bias in Instruction-Following models with P-AT

Dario Onorati, Elena Sofia Ruzzetti, Davide Venditti, Leonardo Ranaldi, Fabio Massimo Zanzotto

Abstract
Instruction-Following Language Models (IFLMs) are promising and versatile tools for solving many downstream, information-seeking tasks. Given their success, there is an urgent need to have a shared resource to determine whether existing and new IFLMs are prone to produce biased language interactions. In this paper, we propose Prompt Association Test (P-AT): a new resource for testing the presence of social biases in IFLMs. P-AT stems from WEAT (Caliskan et al., 2017) and generalizes the notion of measuring social biases to IFLMs. Basically, we cast WEAT word tests in promptized classification tasks, and we associate a metric - the bias score. Our resource consists of 2310 prompts. We then experimented with several families of IFLMs discovering gender and race biases in all the analyzed models. We expect P-AT to be an important tool for quantifying bias across different dimensions and, therefore, for encouraging the creation of fairer IFLMs before their distortions have consequences in the real world.
Anthology ID:
2023.findings-emnlp.539
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2023
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
8006–8034
Language:
URL:
https://aclanthology.org/2023.findings-emnlp.539
DOI:
10.18653/v1/2023.findings-emnlp.539
Bibkey:
Cite (ACL):
Dario Onorati, Elena Sofia Ruzzetti, Davide Venditti, Leonardo Ranaldi, and Fabio Massimo Zanzotto. 2023. Measuring bias in Instruction-Following models with P-AT. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 8006–8034, Singapore. Association for Computational Linguistics.
Cite (Informal):
Measuring bias in Instruction-Following models with P-AT (Onorati et al., Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/dois-2013-emnlp/2023.findings-emnlp.539.pdf
PDF Search