| # GenderVL-Bench |
|
|
| **GenderVL-Bench** is a compact vision-language benchmark for evaluating how Vision-Language Models (VLMs) interpret **gender-related representations across different occupations**. |
|
|
| ### Dataset |
|
|
| * **108 images** |
| * **12 occupations** |
| * **9 images per occupation** |
| * **Format:** JPEG / ImageFolder |
| * **Split:** `train` |
|
|
| ### Usage |
|
|
| ```python |
| from datasets import load_dataset |
| |
| dataset = load_dataset("suparnojit/GenderVL-Bench") |
| ``` |
|
|
| ### Citation |
|
|
| ```bibtex |
| @dataset{sarkar2026gendervlbench, |
| author = {Sarkar, Suparnojit}, |
| title = {GenderVL-Bench}, |
| year = {2026}, |
| publisher = {Hugging Face}, |
| url = {https://huggingface.co/datasets/suparnojit/GenderVL-Bench} |
| } |
| ``` |
|
|
