Sign inGet started
Dataset · Hugging Face

IDEAS-Lab-Northwestern/ManiGuard-Bench

ManiGuard-Bench A safety-aware robotic manipulation benchmark. 200 everyday tabletop tasks across 6 families, each rendered as one in-distribution task plus 4 out-of-distribution perturbations, in high-fidelity physics simulation. Unlike a plain success benchmark, every task carries formal safety constraints — a policy is measured not only on whether it finishes the task, but on whether it stays safe while doing it (never knocking objects over, dropping them, spilling, or… See the full description on the dataset page: https://huggingface.co/datasets/IDEAS-Lab-Northwestern/ManiGuard-Bench.

Author / organizationIDEAS-Lab-Northwestern
License (as declared)cc-by-nc-4.0
Downloads910 (as reported by Hugging Face)
Likes2
Created28 June 2026
Last updated at source30 September 2026
robotic-manipulationbenchmarkbehavior-1komnigibsonltlsafetymanipulationfranka
Embed this entry
<iframe src="https://silytics.si/embed/item/d12a21be-d462-45a8-9ab8-96737bea73f5" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

More from IDEAS-Lab-Northwestern

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 Information may be incomplete or out of date.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.