ReadingModel-based value expansionNotes on model-based value expansion (MVE)Vision Language Models as Reward ModelsNotes on using VLMs as reward models for RL