Improving Task-Specific Multimodal Sentiment Analysis with General MLLMs via Prompting | Read Paper on Bytez