REVIEW 16 cited by
Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models
read the original abstract
We introduce new jailbreak attacks on vision language models (VLMs), which use aligned LLMs and are resilient to text-only jailbreak attacks. Specifically, we develop cross-modality attacks on alignment where we pair adversarial images going through the vision encoder with textual prompts to break the alignment of the language model. Our attacks employ a novel compositional strategy that combines an image, adversarially targeted towards toxic embeddings, with generic prompts to accomplish the jailbreak. Thus, the LLM draws the context to answer the generic prompt from the adversarial image. The generation of benign-appearing adversarial images leverages a novel embedding-space-based methodology, operating with no access to the LLM model. Instead, the attacks require access only to the vision encoder and utilize one of our four embedding space targeting strategies. By not requiring access to the LLM, the attacks lower the entry barrier for attackers, particularly when vision encoders such as CLIP are embedded in closed-source LLMs. The attacks achieve a high success rate across different VLMs, highlighting the risk of cross-modality alignment vulnerabilities, and the need for new alignment approaches for multi-modal models.
Forward citations
Cited by 16 Pith papers
-
Cross-Modal Backdoors in Multimodal Large Language Models
Poisoning a single connector in MLLMs establishes a reusable latent backdoor pathway that transfers across modalities with over 95% attack success rate under bounded perturbations.
-
Hierarchical Attacks for Multi-Modal Multi-Agent Reasoning
HAM³ achieves up to 78.3% attack success rate on the GQA benchmark by hierarchically attacking perception, communication, and reasoning layers in multi-modal multi-agent systems.
-
LLM-as-Judge Framework for Evaluating Tone-Induced Hallucination in Vision-Language Models
Ghost-100 benchmark shows prompt tone drives hallucination rates and intensities in VLMs, with non-monotonic peaks at intermediate pressure and task-specific differences that aggregate metrics hide.
-
Adversarial Hubness in Multi-Modal Retrieval
Adversarial hubs can be generated to be retrieved as top-1 for over 84% of test queries in text-to-image retrieval, far exceeding natural hubs.
-
One Modality to Forget Them All: Enhancing Cross-Modal Unlearning in Vision-Language Models
Cross-modal unlearning transfer in vision-language models is asymmetric, architecture-dependent, and shallow under typographic attacks; influence-guided block selection reduces the measured gap.
-
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 e...
-
MLingualFC: Evaluating Jailbreak Vulnerabilities in Multilingual Vision-Language Models
MLingualFC benchmark finds flowchart jailbreaks succeed at high rates for Latin-script languages but much lower rates for Punjabi in multilingual VLMs, pointing to language-dependent safety gaps.
-
Exploring Adversarial Robustness and Safety Alignment in Multilingual Multi-Modal Large Language Models
Adversarial images transfer across languages in MLLMs while apparent safety in weaker languages stems from comprehension and visual-grounding failures rather than genuine alignment.
-
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.
-
Hard to Read, Easy to Jailbreak: How Visual Degradation Bypasses MLLM Safety Alignment
Degraded image resolution in MLLMs bypasses safety alignments via cognitive overload, raising jailbreak rates across perturbations.
-
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.
-
Visual Inception: Compromising Long-term Planning in Agentic Recommenders via Multimodal Memory Poisoning
Visual Inception poisons images to hijack long-term memory in agentic recommenders and steer planning, while CognitiveGuard reduces success to about 10% via perceptual sanitization and reasoning verification.
-
StructBreak: Structural Cognitive Overload-Induced Safety Failures in MLLMs
StructBreak framework shows structural cognitive overload induces safety failures in MLLMs with 92% average attack success rate across six models.
-
New Wide-Net-Casting Jailbreak Attacks Risk Large Models
The paper demonstrates that a tailored jailbreak method for querying groups of large models can achieve up to 100% success rate in some experiments on unprotected models, revealing overlooked multi-model safety risks.
-
When the Prompt Becomes Visual: Vision-Centric Jailbreak Attacks for Large Image Editing Models
Picture-only instructions can jailbreak large image editing models with up to 80.9% attack success; a simple appended safety trigger mitigates the threat.
-
Breaking the Illusion: Consensus-Based Generative Mitigation of Adversarial Illusions in Multi-Modal Embeddings
Generative purification with consensus aggregation reduces adversarial illusion attack success rates to near zero on ImageBind while improving alignment on both clean and attacked inputs.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.