Limt: Language-Informed Multi-Task Visual World ModelsICRA, 2025Share on Twitter Facebook LinkedIn Previous Next