Dehong Xu

large language model (2) reinforcement learning (1) multi-task learning (1) reward modeling (1) offline reinforcement learning (1) variational inference (1) language model alignment (1) reinforcement learning from human feedback (1) model alignment (1) posterior inference (1) proximal policy optimization (1) knowledge-grounded dialogue (1) zero-shot retrieval (1) knowledge selection (1) latent planning (1) latent variable inference (1) faithful generation (1) multi-modal embedding (1) task-adaptive representation (1) token-level supervision (1)