Zhiyuan Sun

large language model (3) multi-agent system (2) reinforcement learning from human feedback (2) generative reward model (2) prompt engineering (1) instruction following (1) preference modeling (1) game ai (1) embodied agent (1) information gathering (1) ai agent (1) role-playing game (1) positional bia (1) embodied instruction following (1) world dynamics (1) embodied learning (1) detective game (1) reinforcement learning (1) consistency-aware reward (1) visual perception (1)