@misc{indiciae5a04c563d22f,
  title = {Coherent Off-Policy Improvement of Large Behavior Models with Learned Rewards},
  author = {Christian Scherer and Joe Watson and Theo Gruner and Daniel Palenicek and Ingmar Posner and Jan Peters},
  year = {2026},
  url = {https://arxiv.org/abs/2606.02194},
  note = {Source identifier: 2606.02194}
}