Shenzhi Wang

large language model (3) agent system (2) policy constraint (2) offline reinforcement learning (2) adversarial learning (1) reward modeling (1) contrastive learning (1) preference learning (1) benchmark evaluation (1) direct preference optimization (1) preference alignment (1) data annotation (1) efficient inference (1) deep learning (1) reinforcement learning from human feedback (1) feature extraction (1) ai safety (1) model alignment (1) early exit (1) efficient computing (1)