MM-Hallu/PROVE

PROVE Benchmark for evaluating hallucinations in VLM free-form responses using scene-graph representations. 10,606 QA pairs grounded in structured visual property tuples from hyper-detailed DOCCI image captions. Fields Field Description image DOCCI dataset image image_url Original image URL caption Hyper-detailed image caption question Question about the image answer Ground truth answer image_tuples JSON array of scene graph tuples qa_diff_tuples… See the full description on the dataset page: https://huggingface.co/datasets/MM-Hallu/PROVE.

<iframe src="https://silytics.si/embed/item/9721b890-515a-4b0a-959b-91d25a9c1f81" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.