1 paper
Jatin Nainani, Sankaran Vaidyanathan, AJ Yeung +2
Mechanistic interpretability aims to understand the inner workings of large neural networks by identifying circuits, or minimal subgraphs within the model that implement algorithms…