1 paper · 1 filter
Weixiang Zhao, Yingshuo Wang, Yichen Zhang +5
Self-evolving large language model (LLM) agents continually improve by accumulating and reusing past experience, yet it remains unclear whether they faithfully rely on that experie…