How Gender Debiasing Affects Internal Model Representations, and Why It Matters

Hadas Orgad, Seraphina Goldfarb-Tarrant, Yonatan Belinkov


Abstract
Common studies of gender bias in NLP focus either on extrinsic bias measured by model performance on a downstream task or on intrinsic bias found in models’ internal representations. However, the relationship between extrinsic and intrinsic bias is relatively unknown. In this work, we illuminate this relationship by measuring both quantities together: we debias a model during downstream fine-tuning, which reduces extrinsic bias, and measure the effect on intrinsic bias, which is operationalized as bias extractability with information-theoretic probing. Through experiments on two tasks and multiple bias metrics, we show that our intrinsic bias metric is a better indicator of debiasing than (a contextual adaptation of) the standard WEAT metric, and can also expose cases of superficial debiasing. Our framework provides a comprehensive perspective on bias in NLP models, which can be applied to deploy NLP systems in a more informed manner. Our code and model checkpoints are publicly available.
Anthology ID:
2022.naacl-main.188
Volume:
Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Month:
July
Year:
2022
Address:
Seattle, United States
Editors:
Marine Carpuat, Marie-Catherine de Marneffe, Ivan Vladimir Meza Ruiz
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2602–2628
Language:
URL:
https://aclanthology.org/2022.naacl-main.188
DOI:
10.18653/v1/2022.naacl-main.188
Bibkey:
Cite (ACL):
Hadas Orgad, Seraphina Goldfarb-Tarrant, and Yonatan Belinkov. 2022. How Gender Debiasing Affects Internal Model Representations, and Why It Matters. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2602–2628, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
How Gender Debiasing Affects Internal Model Representations, and Why It Matters (Orgad et al., NAACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.naacl-main.188.pdf
Video:
 https://aclanthology.org/2022.naacl-main.188.mp4
Code
 technion-cs-nlp/gender_internal +  additional community code
Data
GAP Coreference DatasetWinoBias