1 paper · 1 filter
Chenxi Wang, Zhuoyun Yu, Xin Xie +8
Learning from experience is critical for building capable large language model (LLM) agents, yet prevailing self-evolving paradigms remain inefficient: agents learn in isolation, r…