AI.info

Ege Erdogan

Explore Ege Erdogan on AI.info.

• Equivariant Sparse Autoencoders: Mechanistic Interpretability of Neural Networks on Symmetric Data