Do Sparse Autoencoders Generalize? A Case Study of Answerability

L. Heindrich, P. Torr, F. Barez, V. Thost
July 2025
Type
Conference paper
Publication
ICML 2025 Workshop
Interpretability ICML