Abstract page for arXiv paper 2610.28801: JUICE: Unified Multimodal Modeling in Pretrained Representation Spaces