@misc{indiciaeeecf6b056319,
  title = {T3R: Training-Free Two-Stage Token Refinement Towards Efficient and Robust VLA Models},
  author = {Mohammed Hassan and Zhenyang Chen and Zheng Wang and Zhixin Zhu and Tianlong Chen and Yingyan and Lin and Chaojian Li},
  year = {2026},
  url = {https://arxiv.org/abs/2611.08413},
  note = {Source identifier: 2611.08413}
}