Ali Shehper comments on Sparse Autoencoders Work on Attention Layer Outputs

[Ali Shehper]

I see. Thanks for the clarification!