Factual Self-Awareness in Language Models: Representation, Robustness, and Scaling - View it on GitHub