JonaRuthardt/CORE

CORE: COnditional REtrieval Benchmark Paper | Project Page | GitHub | SteerViT Model CORE is a benchmark for evaluating whether visual representations can be steered by natural-language prompts. The task is conditional image retrieval: given a query image from a scene and a text prompt naming an object of interest, retrieve other images from the same scene that contain the same object. CORE is designed to test whether global image features can shift away from dominant… See the full description on the dataset page: https://huggingface.co/datasets/JonaRuthardt/CORE.

<iframe src="https://silytics.si/embed/item/49d2486c-69a4-42f1-83a4-08cbc1c00b14" 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 Information may be incomplete or out of date.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.