Logan Riggs comments on Sparse Autoencoders: Future Work