Benchmarking Jailbreak Guardrails for Embodied Agents
cs.AI
Submitted: 2026-10-05
Updated: 2026-10-05
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- VLMGuard: Bootstrapping Malicious Prompt Detectors from Unlabeled Vision-Language Prompts in the Wild
- Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations
- Safety Guardrails for LLM-Enabled Robots
- AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents
- Advancing Embodied Agent Security: From Safety Benchmarks to Input Moderation
- GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning
- OMNIGUARD: An Efficient Approach for AI Safety Moderation Across Languages and Modalities
- Kelp: A Streaming Safeguard for Large Models via Latent Dynamics-Guided Risk Detection
- MADRA: Multi-Agent Debate for Risk-Aware Embodied Planning
- Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring
- Optimized Hybrid Feature Engineering for Resource-Efficient Arrhythmia Detection in ECG Signals: An Optimization Framework
- HomeGuard: VLM-based Embodied Safeguard for Identifying Contextual Risk in Household Task
- EMBGuard: Constructing Hazard-Aware Guardrails for Safe Planning in Embodied Agents
- SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning
- When Are Reasoning-Based Guardrails Not Efficient? ResponseGuard: A Fast Vision-Language Guard for Real-Time Moderation
- GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
- DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
- Gemini Robotics: Bringing AI into the Physical World
- AI2-THOR: An Interactive 3D Environment for Visual AI
- CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulation
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems