Uncertainty-aware Reward Model: Teaching Reward Models to Know What is Unknown