Safety Evaluation of VLA Models

VLA: Vision-Language-Action

Overview

This field tests whether robots and driving agents can interpret instructions and act safely in interactive environments. Benchmarks use matched safe and unsafe scenarios, trajectory-level safety rules, counterfactual tests, and simulation-to-physical evaluation to separate safe behavior from task completion. Methods include adversarial red teaming, runtime action screening and abstention, semantic guards, robustness checks, and safety-aware training. Strong hazard recognition or task success often fails to ensure safe execution, leaving open challenges in cross-domain generalization, standardized metrics, and balancing safety with useful performance.

Latest papers

CardsList

41 papers14 papers in the last four weeks, up 250% on the four weeks before. 0.1% of all new papers.

  1. ReactBench-AD: Do Driving VLAs and World-Action Models React to Imminent Hazards in Their Path?

    Nov 15, 2026Nikos Theodoridis, Reenu Mohandas, Sushil Sharma +5Autonomous DrivingLong-Tail Autonomous Driving

  2. When Backdoors Can Be Switched Off: Suppression-Aware Backdoor Attacks on Vision-Language-Action Models

    Nov 15, 2026Nur Afsa Syeda, Mohamed ElmahallawyBackdoor Attacks on VLA ModelsSafety Evaluation of VLA Models

  3. AffectiveVLA: Safe and Intent-Aware VLA Control for Human-Humanoid Social Interaction

    Nov 5, 2026Nikita Kuzmin, Georgii Demianchuk, Mariya Lezina +4Human-Robot InteractionHumanoid Robot Control

  4. VLA Policies Do Not Know When Not to Act: Unintended Task Execution and Start Suppression

    Oct 30, 2026Minsu Jang, Jae-Woo Choi, Youngwoo Yoon +1Vision-Language-Action ModelsSafety Evaluation of VLA Models

  5. TRACE: Target-State Risk-Aligned Conditional Execution for Embodied Tracking

    Oct 30, 2026Mingyang Lyu, Gongce Cao, Yiyang Jia +7Visual Object TrackingRobotic Control

  6. Benchmarking Obstacle-Conditioned Safety in Household Robot Navigation

    Oct 26, 2026Hyunwoong Bae, Jisu Han, Jiyeon Choi +2Safe Robot NavigationRobot Policy Adaptation

  7. When Human Corrections Travel: Safety in Federated VLA Adaptation

    Oct 23, 2026Vigneshwaran Subbaraju, Obaidullah Zaland, Monowar BhuyanVLA Model AdaptationHuman-in-the-Loop AI

  8. Progress Probes P2P^2: Lightweight Runtime Monitoring for Vision-Language-Action Models

    Oct 22, 2026Atiksh Bhardwaj, Edward W. Duan, Junwon Seo +4Failure Detection for VLA ModelsVision-Language-Action Models

  9. Perceives but Doesn't Brake: A Perception-Action Binding Gap in Reasoning Driving VLAs

    Oct 20, 2026Matteo Cercola, Katie Z Luo, Simone Formentin +1Reasoning for VLA ModelsSafety Evaluation of VLA Models

  10. SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving

    Oct 14, 2026Shaoyuan Xie, Zihan Zhang, Jingxuan Wang +6Autonomous DrivingEnd-to-End Autonomous Driving

  11. ReVeal: A Reconstruction-Aware Real-to-Sim Framework for VLA Policy Evaluation

    Sep 20, 2026Xinyi Wang, Heng Hao, Wenjun Hu +53D ReconstructionRobotic Manipulation

  12. Beyond the Leaderboard: Counterfactual Diagnosis of End-to-End and VLA Driving Policies Under Domain Shift

    Sep 18, 2026Ruolin Yang, Zilin Huang, Buoyue Wang +4Autonomous DrivingAutonomous Driving Safety Evaluation

  13. SafeStage: Evaluating Safety Before, During, and After Vision-Language-Conditioned Robot Manipulation

    Sep 18, 2026Jinzhu Luo, Qi Zhang, Wei Wang +1Robot SafetySafety Evaluation of VLA Models

  14. STAGE: Diagnosing Semantic Transfer at Grounded Execution in Embodied Agents

    Sep 11, 2026Baosheng Jin, Yushen Liang, Hua ShenLanguage-Conditioned Robot ManipulationVision-Language-Action Models

  15. MANIGUARD: A Benchmark and Data Suite for Specification-Grounded Safety Evaluation and Improvement of Robotic Manipulation

    Aug 18, 2026Yiyan Peng, Philip Wang, Simon Sinong Zhan +11Robotic ManipulationSafety Evaluation of VLA Models

  16. MAE I Trust Myself? Self-Evaluating VLA Action Generation with Markov Attention Entropy

    Aug 17, 2026Aniri, Chen Yilin, Jinhe Bi +7Uncertainty Estimation for VLA ModelsLLM Uncertainty Estimation

  17. SafeGesture: Evaluating Fine-Grained Hand Gesture Understanding in Vision-Language Models through Scenario-Conditioned Safety Interpretation

    Aug 17, 2026Taegang Kim, Saleh Afroogh, Junfeng JiaoVLM EvaluationReasoning for VLA Models

  18. SSP: An Event-Matched Syn2Sim2Phy Cross-Domain Evaluation Framework for Autonomous Driving VLA Models

    Aug 14, 2026Haojie Feng, Peizhi Zhang, Xinrui Zhang +7Safety Evaluation of VLA ModelsAutonomous Driving Benchmarks

  19. ForesightSafety-VLA: A Unified Diagnostic Safety Benchmark for Vision-Language-Action Models

    Jun 25, 2026Mingyang Lyu, Yinqian Sun, Yiyang Jia +5RoboticsRobot Safety

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

    Jun 22, 2026Rongxu Cui, Zongzheng Zhang, Jingrui Pang +11Robot SafetyVision-Language-Action Models