3 ms·Review of Mechanistic Interpretability for Transformer-Based Language Models2 points by belter 2y ago