@article{swann2026sparse,
  title={Sparse Autoencoders Reveal Interpretable and Steerable Features in VLA Models},
  author={Swann, Aiden and McGranahan, Lachlain and Buurmeijer, Hugo and Kennedy III, Monroe and Schwager, Mac},
  journal={arXiv preprint arXiv:2603.19183},
  year={2026}
}