Detoxifying Language Models with a Toxic Corpus

Yoona Park, Frank Rudzicz


Abstract
Existing studies have investigated the tendency of autoregressive language models to generate contexts that exhibit undesired biases and toxicity. Various debiasing approaches have been proposed, which are primarily categorized into data-based and decoding-based. In our study, we investigate the ensemble of the two debiasing paradigms, proposing to use toxic corpus as an additional resource to reduce the toxicity. Our result shows that toxic corpus can indeed help to reduce the toxicity of the language generation process substantially, complementing the existing debiasing methods.
Anthology ID:
2022.ltedi-1.6
Volume:
Proceedings of the Second Workshop on Language Technology for Equality, Diversity and Inclusion
Month:
May
Year:
2022
Address:
Dublin, Ireland
Editors:
Bharathi Raja Chakravarthi, B Bharathi, John P McCrae, Manel Zarrouk, Kalika Bali, Paul Buitelaar
Venue:
LTEDI
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
41–46
Language:
URL:
https://aclanthology.org/2022.ltedi-1.6
DOI:
10.18653/v1/2022.ltedi-1.6
Bibkey:
Cite (ACL):
Yoona Park and Frank Rudzicz. 2022. Detoxifying Language Models with a Toxic Corpus. In Proceedings of the Second Workshop on Language Technology for Equality, Diversity and Inclusion, pages 41–46, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
Detoxifying Language Models with a Toxic Corpus (Park & Rudzicz, LTEDI 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.ltedi-1.6.pdf
Data
WebText