Debiasing should be Good and Bad: Measuring the Consistency of Debiasing Techniques in Language Models - ACL Anthology

Debiasing should be Good and Bad: Measuring the Consistency of Debiasing Techniques in Language Models

Robert Morabito, Jad Kabbara, Ali Emami


Abstract
Debiasing methods that seek to mitigate the tendency of Language Models (LMs) to occasionally output toxic or inappropriate text have recently gained traction. In this paper, we propose a standardized protocol which distinguishes methods that yield not only desirable results, but are also consistent with their mechanisms and specifications. For example, we ask, given a debiasing method that is developed to reduce toxicity in LMs, if the definition of toxicity used by the debiasing method is reversed, would the debiasing results also be reversed? We used such considerations to devise three criteria for our new protocol: Specification Polarity, Specification Importance, and Domain Transferability. As a case study, we apply our protocol to a popular debiasing method, Self-Debiasing, and compare it to one we propose, called Instructive Debiasing, and demonstrate that consistency is as important an aspect to debiasing viability as is simply a desirable result. We show that our protocol provides essential insights into the generalizability and interpretability of debiasing methods that may otherwise go overlooked.
Anthology ID:
2023.findings-acl.280
Volume:
Findings of the Association for Computational Linguistics: ACL 2023
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Anna Rogers, Jordan Boyd-Graber, Naoaki Okazaki
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4581–4597
Language:
URL:
https://aclanthology.org/2023.findings-acl.280
DOI:
10.18653/v1/2023.findings-acl.280
Bibkey:
Cite (ACL):
Robert Morabito, Jad Kabbara, and Ali Emami. 2023. Debiasing should be Good and Bad: Measuring the Consistency of Debiasing Techniques in Language Models. In Findings of the Association for Computational Linguistics: ACL 2023, pages 4581–4597, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
Debiasing should be Good and Bad: Measuring the Consistency of Debiasing Techniques in Language Models (Morabito et al., Findings 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.findings-acl.280.pdf
Video:
 https://aclanthology.org/2023.findings-acl.280.mp4