From Tokens to Photons: Test-Time Physical Prompting for Vision-Language Models

This article likely discusses a novel approach to improve the performance of Vision-Language Models (VLMs). The title suggests a method that bridges the gap between abstract token representations and the physical world (photons), potentially by manipulating the input during the testing phase. The use of "physical prompting" implies a focus on real-world characteristics or simulations to enhance model understanding. The source, ArXiv, indicates this is a research paper.