With all renewed discussion about "Sparse AutoEncoders (#SAE)" as a way of doing #MechanisticInterpretability of #LLMs, I am resharing a part of my PhD where we proved years ago about how sparsity automatically emerges in autoencoding.
arxiv.org/abs/1708.03735
arxiv.org
Sparse Coding and Autoencoders
In "Dictionary Learning" one tries to recover incoherent matrices $A^* \in \mathbb{R}^{n \times h}$ (typically overcomplete and whose columns are assumed to be normalized) and sparse vectors $x^* \in ...
