Author name: Seonglae Cho, Zekun Wu, Adriano Koshiyama

arXiv:2508.12535v3 Announce Type: replace
Abstract: Sparse Autoencoders (SAEs) can extract interpretable features from large language models (LLMs) without supervision. However, their effectiveness in downstream steering tasks is limited by the requir…

arXiv:2602.10437v3 Announce Type: replace-cross
Abstract: Sparse autoencoders (SAEs) decompose language model activations into interpretable features, but existing methods reveal only which features activate, not which change model outputs when amplif…