Checkpoints for the paper: "One Adapts to Any: Meta Reward Modeling for Personalized LLM Alignment".