@misc{indiciae123378dae82d,
  title = {OPERA: Aligning Open-Ended Reasoning via Objective Perplexity-based Reinforcement Learning},
  author = {Wenxuan Jiang and Zining Fan and Zijian Zhang and Xuecheng Wu and Hongming Tan and Haoyang Dai and Xiaoyu Li and Xuezhi Cao and Ninghao Liu},
  year = {2026},
  url = {https://arxiv.org/abs/2606.25757},
  note = {Source identifier: 2606.25757}
}