IAAILoopOPD
Recurrent Self-Improvement: Dynamic Cross-Loop On-Policy Distillation for Looped Language Models
Uses a looped model's own terminal loop as a compute-privileged teacher for an intermediate-loop student on student-generated rollouts. The dynamic variant keeps refreshing the teacher as shared weights update, giving recurrent self-improvement without an external teacher.
Inside the method
Simplified conceptual schematic. Consult the paper for the complete architecture.
- Recurrence family
- Full-stack recurrence
- Depth control
- Supervision flows from deeper to shallower loops
- KV / state strategy
- See paper
Reading note
On Ouro-Thinking models, LoopOPD improves mathematical reasoning and the dynamic variant adds further gains; training only on math also helps general reasoning and code. Code and checkpoints are promised upon acceptance.
Sources checked 2026-10-28. This catalog does not imply independent reproduction.
Related works
Cite this work
@misc{loopopd2026,
title = {Recurrent Self-Improvement: Dynamic Cross-Loop On-Policy Distillation for Looped Language Models},
author = {Yi Wang and Rui Qian and Yu Li and Haoyang Yao and Wenjie Wang},
year = {2026},
eprint = {2610.10623},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.10623}
}