Bias and Fairness in Large Language Models: A Survey

arXiv:2309.00770v2 Announce Type: replace-cross
Abstract: Rapid advancements of large language models (LLMs) have enabled the processing, understanding, and generation of human-like text, with increasing integration into systems that touch our social sphere. Despite this success, these models can learn, perpetuate, and amplify harmful social biases. In this paper, we present a comprehensive survey of bias evaluation and mitigation techniques for LLMs. We first consolidate, formalize, and expand notions of social bias and fairness in natural language processing, defining distinct facets …

abstract amplify arxiv bias biases cs.ai cs.cl cs.cy cs.lg fairness human human-like integration language language models large language large language models learn llms paper processing social sphere success survey systems text type understanding