metadata
language:
- en
license: apache-2.0
size_categories:
- 10K<n<100K
task_categories:
- image-text-to-text
tags:
- image-generation
- text_prompt
- safety
- bias
- fairness
- toxicity
- privacy
T2ISafety: Benchmark for Assessing Fairness, Toxicity, and Privacy in Image Generation
This repository contains the dataset for the paper T2ISafety: Benchmark for Assessing Fairness, Toxicity, and Privacy in Image Generation. The dataset includes 68K manually annotated images and is designed to evaluate Text-to-Image (T2I) models across toxicity, fairness, and privacy.
Code: https://github.com/adwardlee/t2i_safety
The dataset is organized according to a detailed hierarchy of 12 tasks and 44 categories based on the three domains of fairness, toxicity, and privacy.