@misc{indiciaef8b58be8172f,
  title = {RolloutPipe: Overlapping Pipelined Rollout and Training in Disaggregated On-Policy LLM Reinforcement Learning},
  author = {Rongjian Chen and Jianmin Hu and Kejiang Ye and Minxian Xu},
  year = {2026},
  url = {https://arxiv.org/abs/2606.26997},
  note = {Source identifier: 2606.26997}
}