The paper titled "ShortV: Efficient Multimodal Large Language Models by Freezing Visual Tokens in Ineffective Layers" addresses key challenges faced by Multimodal Large Language Models (MLLMs), particularly their high computational costs and the management of visual tokens. To tackle these issues, the authors introduce a novel metric called Layer Contribution (LC), designed to evaluate the effectiveness of different layers in processing both visual and textual tokens. This metric aims to optimize model performance by identifying and freezing visual tokens in layers deemed ineffective, thereby reducing computational overhead without compromising accuracy. The positive stance on the Layer Contribution metric suggests it is a promising tool for enhancing MLLM efficiency. This approach aligns with ongoing research efforts to balance model complexity and resource demands in multimodal AI systems. The study contributes to the broader discourse on improving large language models by offering a targeted strategy for layer-wise optimization.