oking0197/DiCoBench

DiCoBench Dataset This dataset card documents the DiCoBench Hugging Face release. Dataset Summary DiCoBench is a high-resolution pairwise visual multiple-choice benchmark for multi-image fine-grained perception. Each example provides two images, one question, five answer choices, and one ground-truth answer letter. The benchmark evaluates whether multimodal models can discover implicit cross-image visual cues without explicit textual target prompts. The dataset… See the full description on the dataset page: https://huggingface.co/datasets/oking0197/DiCoBench.

<iframe src="https://silytics.si/embed/item/5a4e4074-d6af-44d3-878d-76a03ed3b161" 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 Information may be incomplete or out of date.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 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.