MLLMs Fail to Refuse when Using Tools Agentically
cs.AI
Submitted: 2026-10-02
Updated: 2026-10-02
Sources
- AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
- Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond
- Qwen3-VL Technical Report
- Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Memory Injection Attacks on LLM Agents via Query-Only Interaction
- Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation
- GLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents
- Kimi K2.5: Visual Agentic Intelligence
- Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents
- MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
- Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models
- LLaVA-Plus: Learning to Use Tools for Creating Multimodal Agents
- Safety of Multimodal Large Language Models on Images and Texts
- Identifying the Risks of LM Agents with an LM-Emulated Sandbox
- General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model
- Visual ChatGPT: Talking, Drawing and Editing with Visual Foundation Models
- RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents
- MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action
- A Survey on Agentic Multimodal Large Language Models
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