Chenkun Tan

large language model (4) direct preference optimization (2) reinforcement learning from human feedback (2) inference-time alignment (2) in-context learning (1) multimodal learning (1) vision-language alignment (1) model safety (1) safety alignment (1) model training (1) vision-language model (1) context length (1) length extrapolation (1) multimodal understanding (1) steering vector (1) content safety (1) language prior (1) long-context modeling (1) attack success rate (1) safety training (1)