Do Language Models Have Beliefs? Methods for Detecting, Updating, and Visualizing Model Beliefs