Where do Large Vision-Language Models Look at when Answering Questions?
Large Vision-Language Models (LVLMs) have shown promising performance in vision-language understanding and reasoning tasks. However, their visual understanding behaviors remain underexplored. A fundamental question arises: to what extent do LVLMs rely on visual input, and which image regions contribute to their responses? It is non-trivial to interpret the free-form generation of LVLMs due to their complicated visual architecture (e.g., multiple encoders and multi-resolution) and variable-length outputs. In this paper, we extend existing heatmap visualization methods (e.g., iGOS++) to support LVLMs for open-ended visual question answering. We propose a method to select visually relevant tokens that reflect the relevance between generated answers and input image. Furthermore, we conduct a comprehensive analysis of state-of-the-art LVLMs on benchmarks designed to require visual information to answer. Our findings offer several insights into LVLM behavior, including the relationship between focus region and answer correctness, differences in visual attention across architectures, and the impact of LLM scale on visual understanding. The code and data are available at https://github.com/bytedance/LVLM_Interpretation.
Code (1)
Tasks
Question AnsweringVisual Question AnsweringMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
When to Think and When to Look: Uncertainty-Guided Lookback
Test-time thinking (that is, generating explicit intermediate reasoning chains) is known to boost performance in large language models and has recently shown strong gains for large vision language models (LVLMs). However…
Visual GroundingVisual ReasoningLookWise: Knowing When and Where to Look for Fine-Grained Visual Reasoning in Multimodal Large Language Models
Multimodal Large Language Models (MLLMs) are shifting towards "Thinking with Images" by actively exploring image details. While effective, large-scale training is computationally expensive, which has spurred growing inte…
Visual ReasoningTeaching Vision-Language-Action Models What to See and Where to Look
Vision-Language-Action (VLA) models have emerged as a promising paradigm for end-to-end autonomous driving. However, existing VLAs' training relies heavily on text-centric visual question answering and chain-of-thought r…
Visual Question AnsweringTrajectory PredictionAutonomous DrivingWhy We Look Where We Look: Emergent Human-like Fixations of a Foveated Visual Language Model Maximizing Scene Understanding
When humans view scenes without a specific task (free-viewing), they initially direct their eye movements toward the scene center and then fixate on people, text, objects being gazed at or grasped, and semantically meani…
Scene UnderstandingGRPO, Dr. GRPO, and DAPO Are Three Operations on One Number: The Group-Standard-Deviation Identity
Three of the most popular methods for training language models to reason look like three different tricks. They are not. All three adjust a single number: standard deviation, reflecting how much a prompt's sampled answer…