ChatGPT Based Data Augmentation for Improved Parameter-Efficient Debiasing of LLMs¶
Arxiv Link - 2024-02-19 01:28:48
Abstract¶
Large Language models (LLMs), while powerful, exhibit harmful social biases. Debiasing is often challenging due to computational costs, data constraints, and potential degradation of multi-task language capabilities. This work introduces a novel approach utilizing ChatGPT to generate synthetic training data, aiming to enhance the debiasing of LLMs. We propose two strategies: Targeted Prompting, which provides effective debiasing for known biases but necessitates prior specification of bias in question; and General Prompting, which, while slightly less effective, offers debiasing across various categories. We leverage resource-efficient LLM debiasing using adapter tuning and compare the effectiveness of our synthetic data to existing debiasing datasets. Our results reveal that: (1) ChatGPT can efficiently produce high-quality training data for debiasing other LLMs; (2) data produced via our approach surpasses existing datasets in debiasing performance while also preserving internal knowledge of a pre-trained LLM; and (3) synthetic data exhibits generalizability across categories, effectively mitigating various biases, including intersectional ones. These findings underscore the potential of synthetic data in advancing the fairness of LLMs with minimal retraining cost.
Socials¶
| X | |
|---|---|
| 🚀 Exciting new research alert! 🚀 Debiasing Large Language models (LLMs) just got a major boost with a novel approach using ChatGPT to generate synthetic training data. This innovative method aims to enhance debiasing while maintaining multi-task language capabilities. Learn more about this cutting-edge research on leveraging synthetic data for LLM debiasing and its impressive results here: http://arxiv.org/abs/2402.11764v1 #AI #NLP #LLMs #Debiasing #Research #TechInnovation |
🚀 Exciting new research on debiasing Large Language Models (LLMs) using ChatGPT synthetic data! Learn how this innovative approach enhances debiasing while preserving multi-task capabilities. Find out more at: http://arxiv.org/abs/2402.11764v1 #AI #NLP #LLMs #Debiasing #ChatGPT |