1 paper
Ruocan Wei, Shufeng Wang, Ziwei Shi
Large language model (LLM) agents often suffer from high reasoning overhead, excessive token consumption, unstable execution, and inability to reuse past experiences in complex tas…