Towards Global-level Mechanistic Interpretability: A Perspective of Modular Circuits of Large Language Models

Devs

Towards Global-level Mechanistic Interpretability: A Perspective of Modular Circuits of Large Language Models | Read Paper on Bytez