@article{howneuralrewardmodelslearnfeaturesforpol,
  title = {How Neural Reward Models Learn Features for Policy Optimization: A Single-Index Analysis},
  author = {Rei Higuchi and Ryotaro Kawata and Akifumi Wachi and Shokichi Takakura and Kohei Miyaguchi and Taiji Suzuki},
  year = {2026},
  eprint = {2605.24749},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.24749},
}