Improving Task-Specific Multimodal Sentiment Analysis with General MLLMs via Prompting

Open in new window