Walking in Others' Shoes: How Perspective-Taking Guides Large Language Models in Reducing Toxicity and Bias

TL;DR

This paper introduces perspective-taking prompting (PeT), a novel method inspired by social psychology, that effectively reduces toxicity and bias in large language models without requiring model access or extensive retraining.

Contribution

The paper presents PeT, a new prompting strategy that enables LLMs to self-regulate responses by integrating diverse perspectives, significantly decreasing harmful content.

Findings

PeT reduces toxicity by up to 89%.

PeT decreases bias by up to 73%.

PeT outperforms five baseline methods across multiple LLMs.

Abstract

The common toxicity and societal bias in contents generated by large language models (LLMs) necessitate strategies to reduce harm. Present solutions often demand white-box access to the model or substantial training, which is impractical for cutting-edge commercial LLMs. Moreover, prevailing prompting methods depend on external tool feedback and fail to simultaneously lessen toxicity and bias. Motivated by social psychology principles, we propose a novel strategy named \textbf{perspective-taking prompting (\textsc{PeT})} that inspires LLMs to integrate diverse human perspectives and self-regulate their responses. This self-correction mechanism can significantly diminish toxicity (up to $89\%$) and bias (up to $73\%$) in LLMs' responses. Rigorous evaluations and ablation studies are conducted on two commercial LLMs (ChatGPT and GLM) and three open-source LLMs, revealing \textsc{PeT}'s superiority in producing less harmful responses, outperforming five strong baselines.