PRISM: A Methodology for Auditing Biases in Large Language Models

Leif Azzopardi, Yashar Moshfeghi

Auditing Large Language Models (LLMs) to discover their biases and preferences is an emerging challenge in creating Responsible Artificial Intelligence (AI). While various methods…