TAPS-TRM-Maze-CoDesign

A hard-maze TRM checkpoint trained with progress--stability co-design for TAPS adaptive recurrent inference. The model checkpoint is provided as step_651, together with its configuration and model definition.

Scheduling Recursive Reasoning in Looped Transformers

Paper: arXiv:2609.36653

Citation

@misc{wang2026schedulingrecursivereasoninglooped,
  title={Scheduling Recursive Reasoning in Looped Transformers},
  author={Boyuan Wang and Chengyao Yu and Jiaxi Ren and Hongxin Wei and Bingyi Jing and Yuxin Tao},
  year={2026},
  eprint={2609.36653},
  archivePrefix={arXiv},
  primaryClass={cs.LG},
  url={https://arxiv.org/abs/2609.36653}
}