@misc{indiciae469e7cc65cee,
  title = {The Gaussian Is Enough: Flow-Matching Priors Do Not Help When Fine-Tuning Large Behavior Models},
  author = {Chen Xu and Rishi Shah and Hadas Kress-Gazit and Haruki Nishimura and Masha Itkina},
  year = {2026},
  url = {https://arxiv.org/abs/2609.27070},
  note = {Source identifier: 2609.27070}
}