@misc{indiciae94e444a6ad3a,
  title = {On the Power of Pre-training for Generalization in RL: Provable Benefits and Hardness},
  author = {Haotian Ye and Xiaoyu Chen and Liwei Wang and Simon S. Du},
  year = {2023},
  url = {https://arxiv.org/abs/2210.10464},
  note = {Source identifier: 2210.10464}
}