Beyond Scalar Reward Model: Learning Generative Judge from Preference Data

Re-assign community