Tag
This paper introduces PRISM, a framework that integrates Vision-Language Models and Large Language Models through a dynamic question-answering pipeline to improve sequential decision-making in embodied AI tasks.