AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

LIBERO-Safety: A Comprehensive Benchmark for Physical and Semantic Safety in Vision-Language-Action Models

Rongxu Cui, Zongzheng Zhang, Jingrui Pang, Haohan Chi, Jinbang Guo, Saining Zhang, shaoxuan Xie, Xin Jin, Yao Mu, Jiaolong Yang, Guocai Yao, Xianyuan Zhan and 2 more

ECCV 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models · 15 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train on 500 instead of 50 demonstrations per task (ablation)pi0.5 · FSHOA L2success rate (%): 48.9 -> 51.3Improved
train on 500 instead of 50 demonstrations per task (ablation)pi0.5 · FSHOA L2collision rate (%): 12.7 -> 10.0Improved
train on 500 instead of 50 demonstrations per task (ablation)pi0.5 · FSHOA L2LDLJ: -17.78 -> -17.47Improved
train on 500 instead of 50 demonstrations per task (ablation)pi0.5 · FSHOA L2execution time (s): 362.5 -> 355.0Improved
perturb object layout (ablation)pi0.5 · TSA L0collision rate (%): 2.7 -> 6.3Hurt
perturb object layout (ablation)pi0.5 · TSA L0success rate (%): 58.0 -> 56.3Hurt
introduce unseen objects (ablation)pi0.5 · TSA L0collision rate (%): 2.7 -> 3.3No clear effect
perturb image noise, init state, view, or scene (ablation)pi0.5 · TSA L0success rate (%): 58.0 -> 56.3-60.7No clear effect
use reasoning-based RynnBrain-CoP instead of RoboBrain2.0-7Bembodied foundation model · SSR L1refusal rate (%): 40.0 -> 72.0Improved
use reasoning-based RynnBrain-CoP instead of RoboBrain2.0-7Bembodied foundation model · SSR L0refusal rate (%): 80.0 -> 56.0Hurt
use keypose-driven pipeline instead of human teleoperationdata generation · data collectionhuman effort (min/task): 7.4 -> 1.8Improved
evaluate on held-out L2 OOD tasks instead of L0OpenVLA-OFT · AAGsuccess rate (%): 50.0 -> 1.3Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy