@misc{indiciae751ee7f8da88,
  title = {Retrieving from Your Own Mistakes: On-Policy Training of Embedding Models for Multi-Hop Retrieval-Augmented Generation},
  author = {Chenjie Qiu and Zixuan Zeng and Shaobin Shi and Bingfeng Chen and Dongsheng Luo and Boyan Xu},
  year = {2026},
  url = {https://arxiv.org/abs/2610.40409},
  note = {Source identifier: 2610.40409}
}