Towards Understanding the Robustness of Sparse Autoencoders

arXiv:2605.04072v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) have been applied to large language models and protein language models, but not systematically to

arXiv:2605.04209v1 Announce Type: cross Abstract: We present Sparse Backdoor, a supply-chain attack that plants a \emph{provably undetectable} backdoor in pre-trained image

arXiv:2606.27784v1 Announce Type: cross Abstract: The existence of adversarial attacks is often attributed to the presence of non-robust features in neural networks. While prior

arXiv:2603.22300v1 Announce Type: cross Abstract: Scaling Transformers to ultra-long contexts is bottlenecked by the $O(n^2 d)$ cost of self-attention. Existing methods reduce

arXiv:2606.27698v1 Announce Type: cross Abstract: Automatic Speech Recognition systems are notoriously both sensitive to adversarial and benign perturbations. While this has been