# See What You Are Told: Visual Attention Sink in Large Multimodal Models

arXiv: https://arxiv.org/abs/2503.03321 · HTML: https://mlsanity.com/paper/2503.03321/

## Key facts

- Cited by 132 papers in the map (data through 14 Oct 2026).
- Placed in Visual token pruning and efficiency, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 59 of its 69 references that are in the map.
- First posted on arXiv 5 Mar 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Visual token pruning and efficiency](https://mlsanity.com/mllm/visual-token-pruning-and-efficiency/)

## Builds on (most cited first, 30 of 59)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [MMBench: Is Your Multi-modal Model an All-around Player?](https://mlsanity.com/paper/2307.06281/) (2023)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering](https://mlsanity.com/paper/2209.09513/) (2022)
- [Evaluating Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2305.10355/) (2023)
- [Towards VQA Models That Can Read](https://mlsanity.com/paper/1904.08920/) (2019)
- [MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities](https://mlsanity.com/paper/2308.02490/) (2023)
- [Kosmos-2: Grounding Multimodal Large Language Models to the World](https://mlsanity.com/paper/2306.14824/) (2023)
- [SEED-Bench: Benchmarking Multimodal LLMs with Generative Comprehension](https://mlsanity.com/paper/2307.16125/) (2023)
- [A Survey on Multimodal Large Language Models](https://mlsanity.com/paper/2306.13549/) (2023)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [VizWiz Grand Challenge: Answering Visual Questions from Blind People](https://mlsanity.com/paper/1802.08218/) (2018)
- [Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs](https://mlsanity.com/paper/2406.16860/) (2024)
- [Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMs](https://mlsanity.com/paper/2401.06209/) (2024)
- [Aligning Large Multimodal Models with Factually Augmented RLHF](https://mlsanity.com/paper/2309.14525/) (2023)
- [A-OKVQA: A Benchmark for Visual Question Answering using World Knowledge](https://mlsanity.com/paper/2206.01718/) (2022)
- [VILA: On Pre-training for Visual Language Models](https://mlsanity.com/paper/2312.07533/) (2023)

## Built on by (most cited first, 30 of 132)

- [Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Grounding](https://mlsanity.com/paper/2503.06287/) (2025)
- [LiveVLM: Efficient Online Video Understanding via Streaming-Oriented KV Cache and Retrieval](https://mlsanity.com/paper/2505.15269/) (2025)
- [Multimodal Chain of Continuous Thought for Latent-Space Reasoning in Vision-Language Models](https://mlsanity.com/paper/2508.12587/) (2025)
- [A Survey of Multimodal Hallucination Evaluation and Detection](https://mlsanity.com/paper/2507.19024/) (2025)
- [To Sink or Not to Sink: Visual Information Pathways in Large Vision-Language Models](https://mlsanity.com/paper/2510.08510/) (2025)
- [HiDe: Rethinking The Zoom-IN method in High Resolution MLLMs via Hierarchical Decoupling](https://mlsanity.com/paper/2510.00054/) (2025)
- [See&Trek: Training-Free Spatial Prompting for Multimodal Large Language Model](https://mlsanity.com/paper/2509.16087/) (2025)
- [SparseVILA: Decoupling Visual Sparsity for Efficient VLM Inference](https://mlsanity.com/paper/2510.17777/) (2025)
- [MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMs](https://mlsanity.com/paper/2511.14159/) (2025)
- [When Semantics Mislead Vision: Mitigating Large Multimodal Models Hallucinations in Scene Text Spotting and Understanding](https://mlsanity.com/paper/2506.05551/) (2025)
- [Restoring Linguistic Grounding in VLA Models via Train-Free Attention Recalibration](https://mlsanity.com/paper/2603.06001/) (2026)
- [GLSim: Detecting Object Hallucinations in LVLMs via Global-Local Similarity](https://mlsanity.com/paper/2508.19972/) (2025)
- [Unleashing the Potential of Multimodal LLMs for Zero-Shot Spatio-Temporal Video Grounding](https://mlsanity.com/paper/2509.15178/) (2025)
- [Catching the Details: Self-Distilled RoI Predictors for Fine-Grained MLLM Perception](https://mlsanity.com/paper/2509.16944/) (2025)
- [Tell Model Where to Look: Mitigating Hallucinations in MLLMs by Vision-Guided Attention](https://mlsanity.com/paper/2511.20032/) (2025)
- [Hallucination Begins Where Saliency Drops](https://mlsanity.com/paper/2601.20279/) (2026)
- [Value-State Gated Attention for Mitigating Extreme-Token Phenomena in Transformers](https://mlsanity.com/paper/2510.09017/) (2025)
- [VLA Knows Its Limits: Adaptive Execution Horizons for Robot Policies](https://mlsanity.com/paper/2602.21445/) (2026)
- [Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients](https://mlsanity.com/paper/2603.17809/) (2026)
- [Don't Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs](https://mlsanity.com/paper/2504.09456/) (2025)
- [SHIELD: Suppressing Hallucinations In LVLM Encoders via Bias and Vulnerability Defense](https://mlsanity.com/paper/2510.16596/) (2025)
- [Direct Visual Grounding by Directing Attention of Visual Tokens](https://mlsanity.com/paper/2511.12738/) (2025)
- [Revealing Perception and Generation Dynamics in LVLMs: Mitigating Hallucinations via Validated Dominance Correction](https://mlsanity.com/paper/2512.18813/) (2025)
- [What Do Visual Tokens Really Encode? Uncovering Sparsity and Redundancy in Multimodal Large Language Models](https://mlsanity.com/paper/2603.00510/) (2026)
- [Rethinking Explainability in the Era of Multimodal AI](https://mlsanity.com/paper/2506.13060/) (2025)
- [Can Vision-Language Models Count? A Synthetic Benchmark and Analysis of Attention-Based Interventions](https://mlsanity.com/paper/2511.17722/) (2025)
- [On the Nature of Attention Sink that Shapes Decoding Strategy in Omni-LLMs](https://mlsanity.com/paper/2603.14337/) (2026)
- [Q-Zoom: Query-Aware Adaptive Perception for Efficient Multimodal Large Language Models](https://mlsanity.com/paper/2604.06912/) (2026)
- [The Point, the Vision and the Text: Does Point Cloud Boost Spatial Reasoning of Large Language Models? A Bias-Controlled Study](https://mlsanity.com/paper/2504.04540/) (2025)
- [Segmentation as A Plug-and-Play Capability for Frozen Multimodal LLMs](https://mlsanity.com/paper/2510.16785/) (2025)

## Abstract (arXiv)

Large multimodal models (LMMs) "see" images by leveraging the attention mechanism between text and visual tokens in the transformer decoder. Ideally, these models should focus on key visual information relevant to the text token. However, recent findings indicate that LMMs have an extraordinary tendency to consistently allocate high attention weights to specific visual tokens, even when these tokens are irrelevant to the corresponding text. In this study, we investigate the property behind the appearance of these irrelevant visual tokens and examine their characteristics. Our findings show that this behavior arises due to the massive activation of certain hidden state dimensions, which resembles the attention sink found in language models. Hence, we refer to this phenomenon as the visual attention sink. In particular, our analysis reveals that removing the irrelevant visual sink tokens does not impact model performance, despite receiving high attention weights. Consequently, we recycle the attention to these tokens as surplus resources, redistributing the attention budget to enhance focus on the image. To achieve this, we introduce Visual Attention Redistribution (VAR), a method that redistributes attention in image-centric heads, which we identify as innately focusing on visual information. VAR can be seamlessly applied across different LMMs to improve performance on a wide range of tasks, including general vision-language tasks, visual hallucination tasks, and vision-centric tasks, all without the need for additional training, models, or inference steps. Experimental results demonstrate that VAR enables LMMs to process visual information more effectively by adjusting their internal attention mechanisms, offering a new direction to enhancing the multimodal capabilities of LMMs.