@misc{indiciae54b6a5e010a4,
  title = {From Demonstrations to Rewards: Test-Time Prompt Optimization for VLM Reward Models},
  author = {Christian Gumbsch and Leonardo Barcellona and Lennard Schünemann and Platon Karageorgis and Andrii Zadaianchuk and Zehao Wang and Sergey Zakharov and Fabien Despinoy and Rahaf Aljundi and Efstratios Gavves},
  year = {2026},
  url = {https://arxiv.org/abs/2606.00083},
  note = {Source identifier: 2606.00083}
}