Exploring Multimodal Perception in Large Language Models Through Perceptual Strength Ratings