Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks

摘要

Large language models (LLMs) inherit biases from their training data and alignment processes, influencing their responses in subtle ways. While many studies have examined these biases, little work has explored their robustness during interactions. In this paper, we introduce a novel approach where two instances of an LLM engage in self-debate, arguing opposing viewpoints to persuade a neutral version of the model. Through this, we evaluate how firmly biases hold and whether models are susceptible to reinforcing misinformation or shifting to harmful viewpoints. Our experiments span multiple LLMs of varying sizes, origins, and languages, providing deeper insights into bias persistence and flexibility across linguistic and cultural contexts.