Fetching the paper…

TACO: Enhancing Multimodal In-context Learning via Task Mapping-Guided Sequence Configuration · Around