@misc{indiciae41779cc146a5,
  title = {GTR: Guided Thought Reinforcement Prevents Thought Collapse in RL-based VLM Agent Training},
  author = {Tong Wei and Yijun Yang and Junliang Xing and Yuanchun Shi and Zongqing Lu and Deheng Ye},
  year = {2025},
  url = {https://arxiv.org/abs/2503.08525},
  note = {Source identifier: 2503.08525}
}