Multimodal Pragmatic Jailbreak on Text-to-image Models

Project page | Paper | Code

The Multimodal Pragmatic Unsafe Prompts (MPUP) is a dataset designed to assess the multimodal pragmatic safety in Text-to-Image (T2I) models. It comprises two key sections: image_prompt, and text_prompt.

Dataset Usage

Downloading the Data

To download the dataset, install Huggingface Datasets and then use the following command:

from datasets import load_dataset
dataset = load_dataset("tongliuphysics/multimodalpragmatic")

Data Format

The MPUP datasets contain the following fields:

• id: prompt id.
• class: Category of the prompt.
• subclass: Subclasses of the prompt.
• label: 1 (unsafe) or 0 (safe).
• text_prompt: Visual text prompt that generates the visual text.
• image_prompt: Image-generation prompt that generates the image.