Alignment of vision-language models with human V1-V3 early visual cortex negatively predicts resistance to sycophantic gaslighting attacks.
Query-relevant images jailbreak large multi-modal models
9 Pith papers cite this work, alongside 7 external citations. Polarity classification is still indexing.
9
Pith papers citing it
7
external citations · Pith
abstract
The security concerns surrounding Large Language Models (LLMs) have been extensively explored, yet the safety of Multimodal Large Language Models (MLLMs) remains understudied. In this paper, we observe that Multimodal Large Language Models (MLLMs) can be easily compromised by query-relevant images, as if the text query itself were malicious. To address this, we introduce MM-SafetyBench, a comprehensive framework designed for conducting safety-critical evaluations of MLLMs against such image-based manipulations. We have compiled a dataset comprising 13 scenarios, resulting in a total of 5,040 text-image pairs. Our analysis across 12 state-of-the-art models reveals that MLLMs are susceptible to breaches instigated by our approach, even when the equipped LLMs have been safety-aligned. In response, we propose a straightforward yet effective prompting strategy to enhance the resilience of MLLMs against these types of attacks. Our work underscores the need for a concerted effort to strengthen and enhance the safety measures of open-source MLLMs against potential malicious exploits. The resource is available at https://github.com/isXinLiu/MM-SafetyBench
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
unclear 1representative citing papers
PolicyShiftGuard, trained with randomized policy SFT plus matched pass/block boundary pairs, reaches 76.9 Avg. F1 and 72.1 Avg. PSS on a new policy-shift image-guardrail benchmark where existing VLMs stay brittle.
A narrative survey that catalogs fifty papers on diffusion-based adversarial techniques across text, vision, and vision-language models, proposes a six-class taxonomy of diffusion roles plus a unified five-dimension evaluation framework, and releases a companion catalog.
Introduces MM-Privacy dataset and evaluations showing MLLMs leak sensitive data from images in various tasks, highlighting task inconsistency effects.
SafeSteer improves safety in multimodal large language models by up to 33.4% via a decoding probe and modal alignment vector without any fine-tuning.
Universal adversarial attacks cause output perturbation 90 times more often than precise target injection in VLMs, with only 2 verbatim successes out of 6615 tests.
AICA-Bench evaluates 23 VLMs on affective image analysis, identifies weak intensity calibration and shallow descriptions as limitations, and proposes training-free Grounded Affective Tree Prompting to improve performance.
SafeDIG applies position-aware sparse feature transfer via SAEs in DiT models to reduce unsafe generations in target risk domains on FLUX.1 Dev and SD 3.5 while keeping source safety and quality.
citing papers explorer
-
Gaslight, Gatekeep, V1-V3: Early Visual Cortex Alignment Shields Vision-Language Models from Sycophantic Manipulation
Alignment of vision-language models with human V1-V3 early visual cortex negatively predicts resistance to sycophantic gaslighting attacks.
-
PolicyShiftGuard: Benchmarking and Improving Policy-Adaptive Image Guardrails
PolicyShiftGuard, trained with randomized policy SFT plus matched pass/block boundary pairs, reaches 76.9 Avg. F1 and 72.1 Avg. PSS on a new policy-shift image-guardrail benchmark where existing VLMs stay brittle.
-
Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models
A narrative survey that catalogs fifty papers on diffusion-based adversarial techniques across text, vision, and vision-language models, proposes a six-class taxonomy of diffusion roles plus a unified five-dimension evaluation framework, and releases a companion catalog.
-
Unveiling Privacy Risks in Multi-modal Large Language Models: Task-specific Vulnerabilities and Mitigation Challenges
Introduces MM-Privacy dataset and evaluations showing MLLMs leak sensitive data from images in various tasks, highlighting task inconsistency effects.
-
SafeSteer: A Decoding-level Defense Mechanism for Multimodal Large Language Models
SafeSteer improves safety in multimodal large language models by up to 33.4% via a decoding probe and modal alignment vector without any fine-tuning.
-
VisInject: Disruption != Injection -- A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language Models
Universal adversarial attacks cause output perturbation 90 times more often than precise target injection in VLMs, with only 2 verbatim successes out of 6615 tests.
-
AICA-Bench: Holistically Examining the Capabilities of VLMs in Affective Image Content Analysis
AICA-Bench evaluates 23 VLMs on affective image analysis, identifies weak intensity calibration and shallow descriptions as limitations, and proposes training-free Grounded Affective Tree Prompting to improve performance.
-
Robust and Generalizable Safety Steering for Text-to-Image Diffusion Transformers
SafeDIG applies position-aware sparse feature transfer via SAEs in DiT models to reduce unsafe generations in target risk domains on FLUX.1 Dev and SD 3.5 while keeping source safety and quality.
- RedDiffuser: Auditing Multimodal Safety Failures in Vision-Language Models via Reinforced Diffusion