T$^2$PO: Uncertainty-Guided Exploration Control for Stable Multi-Turn Agentic Reinforcement Learning

TL;DR

T$^2$PO introduces an uncertainty-guided exploration control framework for multi-turn RL, enhancing training stability and efficiency by fine-grained monitoring and resampling of low-information actions.

Contribution

It proposes a novel token- and turn-level policy optimization method that explicitly manages exploration based on uncertainty dynamics in multi-turn RL.

Findings

Significant improvements in training stability across diverse environments.

Enhanced exploration efficiency leading to better performance.

Demonstrated effectiveness of uncertainty-aware control in complex tasks.

Abstract

Recent progress in multi-turn reinforcement learning (RL) has significantly improved reasoning LLMs' performances on complex interactive tasks. Despite advances in stabilization techniques such as fine-grained credit assignment and trajectory filtering, instability remains pervasive and often leads to training collapse. We argue that this instability stems from inefficient exploration in multi-turn settings, where policies continue to generate low-information actions that neither reduce uncertainty nor advance task progress. To address this issue, we propose Token- and Turn-level Policy Optimization (T$^2$PO), an uncertainty-aware framework that explicitly controls exploration at fine-grained levels. At the token level, T$^2$PO monitors uncertainty dynamics and triggers a thinking intervention once the marginal uncertainty change falls below a threshold. At the turn level, T$^2$PO identifies interactions with negligible exploration progress and dynamically resamples such turns to avoid wasted rollouts. We evaluate T$^2$PO in diverse environments, including WebShop, ALFWorld, and Search QA, demonstrating substantial gains in training stability and performance improvements with better exploration efficiency. Code is available at: https://github.com/WillDreamer/T2PO.