Exploring Autonomous Agents: A Closer Look at Why They Fail When Completing Tasks

Ruofan Lu, Yichen Li, Yintong Huo

Autonomous agent systems powered by Large Language Models (LLMs) have demonstrated promising capabilities in automating complex tasks. However, current evaluations largely rely on…