@misc{indiciae793003d2c0c9,
  title = {Focusing on What Matters: Object-Agent-centric Tokenization for Vision Language Action models},
  author = {Rokas Bendikas and Daniel Dijkman and Markus Peschl and Sanjay Haresh and Pietro Mazzaglia},
  year = {2025},
  url = {https://arxiv.org/abs/2509.23655},
  note = {Source identifier: 2509.23655}
}