Implements a Delphi of GANs approach to detect bias in pre-trained language models and provides a novel way to assess and compare bias levels in different models across multiple dimensions such as gender, racial, and age bias
-
Updated
Sep 4, 2024 - Jupyter Notebook