baulab/openai-comic-strips

OpenAI Comic Strips 500 six-panel comic strips (3,000 images) generated with OpenAI's gpt-image-1, each paired with structured metadata: an art style, a recurring protagonist, and a one-sentence caption for every panel. The dataset was built to study spatial grounding in vision-language models: specifically, how a VLM's attention tracks which panel of a multi-panel image it is currently describing. Because each strip is laid out as six discrete panels with known per-panel… See the full description on the dataset page: https://huggingface.co/datasets/baulab/openai-comic-strips.

<iframe src="https://silytics.si/embed/item/006a9583-2426-42bc-bb52-38cb0bb025e0" 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 Information may be incomplete or out of date.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 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.