Causal-Guided Active Learning for Debiasing Large Language Models

Zhouhao Sun, Li Du, Xiao Ding, Yixuan Ma, Yang Zhao, Kaitao Qiu, Ting Liu, Bing Qin


Abstract
Although achieving promising performance, recent analyses show that current generative large language models (LLMs) may still capture dataset biases and utilize them for generation, leading to poor generalizability and harmfulness of LLMs. However, due to the diversity of dataset biases and the over-optimization problem, previous prior-knowledge-based debiasing methods and fine-tuning-based debiasing methods may not be suitable for current LLMs.To address this issue, we explore combining active learning with the causal mechanisms and propose a casual-guided active learning (CAL) framework, which utilizes LLMs itself to automatically and autonomously identify informative biased samples and induce the bias patterns. Then a cost-effective and efficient in-context learning based method is employed to prevent LLMs from utilizing dataset biases during generation.Experimental results show that CAL can effectively recognize typical biased instances and induce various bias patterns for debiasing LLMs.
Anthology ID:
2024.acl-long.778
Volume:
Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
14455–14469
Language:
URL:
https://aclanthology.org/2024.acl-long.778
DOI:
10.18653/v1/2024.acl-long.778
Bibkey:
Cite (ACL):
Zhouhao Sun, Li Du, Xiao Ding, Yixuan Ma, Yang Zhao, Kaitao Qiu, Ting Liu, and Bing Qin. 2024. Causal-Guided Active Learning for Debiasing Large Language Models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 14455–14469, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
Causal-Guided Active Learning for Debiasing Large Language Models (Sun et al., ACL 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.acl-long.778.pdf