Submitted by Chirag Agarwal 9 Towards Understanding the Robustness of Sparse Autoencoders Aikyam Lab 1 2