HazardArena: Evaluating Semantic Safety in Vision-Language-Action Models

arXiv:2604.12447 · cs.RO · Submitted 2026-04-14 · Read on arXiv

cs.RO

Submitted: 2026-04-14

Updated: 2026-09-23

Comments: Submitted to conference; 23 pages, 5 figures, including supplementary material; revised version with additional experiments and updated analysis

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Benchmark — 25× in this paper · explained in 2 episodes
• Fine-Tuning — 16× in this paper · explained in 5 episodes
• Actor — 13× in this paper · explained in 3 episodes
• Safety Option Layer — 10× in this paper
• VLA Models — 8× in this paper · explained in 3 episodes
• Generalization — 5× in this paper · explained in 13 episodes
• Alignment — 3× in this paper
• Vision-Language-Action (VLA) Models — 3× in this paper · explained in 4 episodes
• Action Space — 2× in this paper · explained in 1 episode
• Benchmarking — 2× in this paper · explained in 4 episodes
• Decoupling — 2× in this paper · explained in 2 episodes
• Failure Modes — 2× in this paper · explained in 3 episodes
• Fidelity — 2× in this paper · explained in 2 episodes
• Grounding — 2× in this paper · explained in 3 episodes
• Model-Agnostic — 2× in this paper · explained in 2 episodes
• Affordance — 1× in this paper · explained in 1 episode
• Aperture — 1× in this paper
• Consistency — 1× in this paper · explained in 3 episodes
• Control Barrier Functions (CBFs) — 1× in this paper · explained in 6 episodes
• Embodied AI — 1× in this paper · explained in 2 episodes
• Generalizability — 1× in this paper · explained in 6 episodes
• Learning Rate — 1× in this paper
• Objective Mismatch — 1× in this paper · explained in 2 episodes
• Semantic Understanding — 1× in this paper · explained in 2 episodes
• Supervised Fine-Tuning (SFT) — 1× in this paper · explained in 8 episodes
• Task Planning — 1× in this paper · explained in 2 episodes

Sources

• $\pi_0$: A Vision-Language-Action Flow Model for General Robot Control
• VLSA: Vision-Language-Action Models with Plug-and-Play Safety Constraint Layer
• NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks
• Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success
• AGENTSAFE: Benchmarking the Safety of Embodied Agents on Hazardous Instructions
• IS-Bench: Evaluating Interactive Safety of VLM-Driven Embodied Agents in Daily Household Tasks
• VLA-Adapter: An Effective Paradigm for Tiny-Scale Vision-Language-Action Model
• SafeAgentBench: A Benchmark for Safe Task Planning of Embodied LLM Agents

Related papers

• FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
• MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
• RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
• ARC: A Reasoning Recipe for Robot Foundation Models
• HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
• APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies