Deep Dive into Anthropic's Sparse Autoencoders by Hand ✍️ | Towards Data Science

Explore the concepts behind the interpretability quest for LLMs