1 thinkers, books, and ideas connected to this field.
American AI researcher at Anthropic — who works on mechanistic interpretability — known for research on superposition in neural networks and toy model analysis of how transformers store information.