3 minute read
Mechanistic Interpretability: Unraveling the Mysteries of Neural Networks
Neil Kubler discusses mechanistic interpretability of neural networks, including superposition, induction heads, and the implications for AI alignment and safety.
