Unified Multimodal Chain-of-Thought Reward Model through Reinforcement Fine-Tuning | Read Paper on Bytez