Are Vision-Language Models Safe in the Wild? A Meme-Based Benchmark Study

Rapid deployment of vision-language models VLMs magnifies safety risks, yet most evaluations rely on artificial images. This study asks: How safe are current VLMs when confronted with meme images that ordinary users share? To investigate this question, we introduce MemeSafetyBench, a...

FragFake: a Dataset for Fine-Grained Detection of Edited Images with Vision Language Models

Fine-grained edited image detection of localized edits in images is crucial for assessing content authenticity, especially given that modern diffusion models and image editing methods can produce highly realistic manipulations. However, this domain faces three challenges: 1 Binary classifiers yie...

Diverging Towards Hallucination: Detection of Failures in Vision-Language Models Via Multi-Token Aggregation

Vision-language models VLMs now rival human performance on many multimodal tasks, yet they still hallucinate objects or generate unsafe text. Current hallucination detectors, e.g., single-token linear probing SLP and PTrue, typically analyze only the logit of the first generated token or just its...

Revisiting Data Auditing in Large Vision-Language Models

With the surge of large language models LLMs, Large Vision-Language Models VLMs--which integrate vision encoders with LLMs for accurate visual grounding--have shown great potential in tasks like generalist agents and robotic control. However, VLMs are typically trained on massive web-scraped...

R-TPT: Improving Adversarial Robustness of Vision-Language Models through Test-Time Prompt Tuning

Vision-language models VLMs, such as CLIP, have gained significant popularity as foundation models, with numerous fine-tuning methods developed to enhance performance on downstream tasks. However, due to their inherent vulnerability and the common practice of selecting from a limited set of...

MAI-2024-0012

Large Vision-Language Models VLMs are susceptible to a sophisticated black-box jailbreak attack known as IDEATOR. This attack utilizes a separate VLM to craft malicious image-text pairs designed to circumvent the target VLM's safety protocols. By iteratively refining prompts based on the target...

MAI-2024-0035

This vulnerability affects various Large Vision-Language Models LVLMs, wherein ostensibly safe images, when combined with additional safe images and specific prompts, can be exploited to produce unsafe and harmful content. The attack methodology, known as the Safety Snowball Agent, leverages the...

MAI-2024-0055

Large Vision Language Models LVLMs are susceptible to a sophisticated attack known as the bi-modal adversarial prompt BAP. This attack exploits the integration of textual and visual prompts to circumvent established safety protocols, thereby eliciting harmful outputs from models engineered to...

MAI-2024-0065

The ImgTrojan attack represents a sophisticated data poisoning strategy targeting Vision-Language Models VLMs. This method enables adversaries to circumvent built-in safety protocols by introducing a minimal number of maliciously designed image-caption pairs into the training dataset. These...

MAI-2023-0006

Large Vision-Language Models VLMs are susceptible to exploitation through typographic visual prompts, a method known as jailbreaking. This vulnerability arises from the VLM's capability to process and interpret text embedded within images, circumventing safety protocols that are typically effecti...