XianjingHan/CultureVidBench
CultureVidBench Dataset Dataset Description CultureVidBench is a benchmark for evaluating cultural understanding in text-to-video (T2V) generation. The benchmark contains 1,000 prompts covering: 12 countries 6 continents 8 cultural regions 3 cultural categories 14 cultural aspects Project Page: https://hanxjing.github.io/CultureVidBench/ Cultural Categories Category Aspects Number of Prompts Material Culture Food, Clothing, Architecture… See the full description on the dataset page: https://huggingface.co/datasets/XianjingHan/CultureVidBench.
CultureVidBench Dataset
Dataset Description
CultureVidBench is a benchmark for evaluating cultural understanding in text-to-video (T2V) generation.
The benchmark contains 1,000 prompts covering:
- 12 countries
- 6 continents
- 8 cultural regions
- 3 cultural categories
- 14 cultural aspects
Project Page: https://hanxjing.github.io/CultureVidBench/
Cultural Categories
Prompt Distribution by Country
Dataset Format
countryaspectculture_elementprompt
Dataset Sources
- Cultural elements were collected from CulturalAtlas and Wikipedia.
Citation
If you find this dataset helpful, please consider citing our work:
@inproceedings{han2026culturevidbench,
title={CultureVidBench: Benchmarking Cultural Understanding in Text-to-Video Generation},
author={Han, Xianjing and Su, Yuhan and Deng, Yang and Ma, Dong and Tay, Wee Peng and Zhu, Bin},
booktitle={Proceedings of the Conference on Empirical Methods in Natural Language Processing},
year={2026}
}