Self-Reflection Makes Large Language Models Safer, Less Biased, and Ideologically Neutral

TL;DR

This paper demonstrates that self-reflection in large language models can significantly improve safety, reduce bias, and neutralize ideological leaning, although its effectiveness on reasoning depends on prompt wording and complexity.

Contribution

The study systematically evaluates self-reflection in LLMs, revealing its potential for safety and bias mitigation and clarifying its impact on reasoning performance with prompt sensitivity.

Findings

Self-reflection reduces toxic responses by 75.8%.

Self-reflection decreases gender bias by 77%.

Self-reflection eliminates partisan bias completely.

Abstract

Previous studies proposed that the reasoning capabilities of large language models (LLMs) can be improved through self-reflection, i.e., letting LLMs reflect on their own output to identify and correct mistakes in the initial responses. However, earlier experiments offer mixed results when it comes to the benefits of self-reflection. Furthermore, prior studies on self-reflection are predominantly concerned with the reasoning capabilities of models, ignoring the potential for self-reflection in safety, bias, and ideological leaning. Here, by conducting a series of experiments testing LLM's self-reflection capability in various tasks using a variety of prompts and different LLMs, we make several contributions to the literature. First, we reconcile conflicting findings regarding the benefit of self-reflection, by demonstrating that the outcome of self-reflection is sensitive to prompt wording -- both the original prompt that are used to elicit an initial answer and the subsequent prompt used to self-reflect. Specifically, although self-reflection may improve the reasoning capability of LLMs when the initial response is simple, the technique cannot improve upon the state-of-the-art chain-of-thought (CoT) prompting. Second, we show that self-reflection can lead to safer (75.8\% reduction in toxic responses while preserving 97.8\% non-toxic ones), less biased (77\% reduction in gender biased responses, while preserving 94.3\% unbiased ones), and more ideologically neutral responses (100\% reduction in partisan leaning response, while preserving 87.7\% non-partisan ones). The paper concludes by discussing the implications of our findings on the deployment of large language models. We release our experiments at https://github.com/Michael98Liu/self-reflection.