# Localize and Neutralize: Gradient-guided Token Suppression against Visual Prompt Injection Attack

arXiv: https://arxiv.org/abs/2605.25194 · HTML: https://mlsanity.com/paper/2605.25194/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Vision-language and text-to-image safety, one of 12 directions in LLM safety and security.
- Builds on 41 of its 47 references that are in the map.
- First posted on arXiv 24 May 2026.

## Placed in

- LLM safety and security / [Vision-language and text-to-image safety](https://mlsanity.com/safety/vision-language-and-text-to-image-safety/)

## Builds on (most cited first, 30 of 41)

- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal](https://mlsanity.com/paper/2402.04249/) (2024)
- [MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities](https://mlsanity.com/paper/2308.02490/) (2023)
- [MiniGPT-v2: large language model as a unified interface for vision-language multi-task learning](https://mlsanity.com/paper/2310.09478/) (2023)
- [Diffusion Models for Adversarial Purification](https://mlsanity.com/paper/2205.07460/) (2022)
- [On Evaluating Adversarial Robustness](https://mlsanity.com/paper/1902.06705/) (2019)
- [FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts](https://mlsanity.com/paper/2311.05608/) (2023)
- [Visual Adversarial Examples Jailbreak Aligned Large Language Models](https://mlsanity.com/paper/2306.13213/) (2023)
- [Generic Attention-model Explainability for Interpreting Bi-Modal and Encoder-Decoder Transformers](https://mlsanity.com/paper/2103.15679/) (2021)
- [MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models](https://mlsanity.com/paper/2311.17600/) (2023)
- [Image Hijacks: Adversarial Images can Control Generative Models at Runtime](https://mlsanity.com/paper/2309.00236/) (2023)
- [Divide, Conquer and Combine: A Training-Free Framework for High-Resolution Image Perception in Multimodal Large Language Models](https://mlsanity.com/paper/2408.15556/) (2024)
- [Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models](https://mlsanity.com/paper/2403.09792/) (2024)
- [AdaShield: Safeguarding Multimodal Large Language Models from Structure-based Attack via Adaptive Shield Prompting](https://mlsanity.com/paper/2403.09513/) (2024)
- [Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs](https://mlsanity.com/paper/2307.10490/) (2023)
- [Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation](https://mlsanity.com/paper/2403.09572/) (2024)
- [Jailbreak Vision Language Models via Bi-Modal Adversarial Prompt](https://mlsanity.com/paper/2406.04031/) (2024)
- [PromptArmor: Simple yet Effective Prompt Injection Defenses](https://mlsanity.com/paper/2507.15219/) (2025)
- [A Survey of Attacks on Large Vision-Language Models: Resources, Advances, and Future Trends](https://mlsanity.com/paper/2407.07403/) (2024)
- [VLATTACK: Multimodal Adversarial Attacks on Vision-Language Tasks via Pre-trained Models](https://mlsanity.com/paper/2310.04655/) (2023)
- [Visual-RolePlay: Universal Jailbreak Attack on MultiModal Large Language Models via Role-playing Image Character](https://mlsanity.com/paper/2405.20773/) (2024)
- [White-box Multimodal Jailbreaks Against Large Vision-Language Models](https://mlsanity.com/paper/2405.17894/) (2024)
- [Mind the box: l1-APGD for sparse adversarial attacks on image classifiers](https://mlsanity.com/paper/2103.01208/) (2021)
- [Steering Away from Harm: An Adaptive Approach to Defending Vision Language Model Against Jailbreaks](https://mlsanity.com/paper/2411.16721/) (2024)
- [VPI-Bench: Visual Prompt Injection Attacks for Computer-Use Agents](https://mlsanity.com/paper/2506.02456/) (2025)

## Abstract (arXiv)

Adversarial images pose a severe security threat to multimodal large language models through prompt injection. Existing defenses largely lack a principled understanding of the underlying mechanisms and struggle to balance efficiency and defense utility. In this work, we show that successful adversarial attacks do not rely on the entire image uniformly but instead depend on a small subset of critical image tokens. Based on this insight, we propose Gradient Token Masking (GTM), which localizes these tokens via gradient analysis and neutralizes them through masking. We find that attribution based on the first generated token's output probability fails when attacks preserve the predicted token. To overcome this, GTM utilizes the Hidden-State Gradient Norm score for generation-influence attribution under adversarial inputs. We prove that its ranking is consistent with that of the full adversarial loss gradient, providing a theoretical guarantee for accurate localization. Our method requires only a single forward-backward pass to identify and zero out a small number of high-scoring tokens, effectively disrupting the adversarial attack path. Extensive experiments on prompt injection and multimodal jailbreak attacks demonstrate that our approach reduces attack success rates (ASR) to near zero while preserving model utility with negligible computational overhead.