Hierarchical Reinforcement Learning with Augmented Step-Level Transitions for LLM Agents

ArXi:2604.05808v1 Announce Type: new Large language model (LLM) agents have nstrated strong capabilities in complex interactive decision-making tasks. However, existing LLM agents typically rely on increasingly long interaction histories, resulting in high computational cost and limited scalability. In this paper, we propose STEP-HRL, a hierarchical reinforcement learning (HRL) framework that enables step-level learning by conditioning only on single-step transitions rather than full interaction histories.