HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks

Open in new window