Donghai Hong

large language model (4) reinforcement learning from human feedback (2) preference learning (1) knowledge distillation (1) question answering (1) model safety (1) responsible ai (1) safety alignment (1) monte carlo tree search (1) human feedback (1) reward model (1) hallucination reduction (1) alignment method (1) model-agnostic approach (1) residual correction (1) human preference (1) policy model (1) preference datum (1) test-time scaling (1) leaderboard evaluation (1)