1 paper
Zili Zhang, Yinmin Zhong, Chengxu Yang +5
Agentic Reinforcement Learning (RL) enables LLMs to solve complex tasks by alternating between a data-collection rollout phase and a policy training phase. During rollout, the agen…