This repository contains the official implementation of $\mathbf{T^3}$ as described in the paper Reducing Belief Deviation in Reinforcement Learning for Active Reasoning of LLM Agents by Deyu Zou, Yongqiang Chen, Jianxiang Wang, Garry YANG, Mufei Li, Qing Da, James Cheng, Pan Li, Yu Gong, which has been selected as ICLR 2026 Oral Presentation.

The codebase is currently under preparation. We will make the full implementation publicly available by March 22, 2026.

Thanks for your patience and attention!