readme created

README.md

@@ -0,0 +1,31 @@
+# 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}
+}
+```