ULMA: Unified Language Model Alignment with Demonstration and Point-wise Human Preference

Open in new window