Tzu-Han Lin

model merging (2) reward model (2) knowledge editing (1) factual knowledge (1) language model alignment (1) model editing (1) reinforcement learning from human feedback (1) parameter efficient (1) domain knowledge (1) vision-language model (1) preference datum (1) llm alignment (1) large language model (1) knowledge distortion (1) reward modeling (1) model deterioration (1) preference learning (1)