Models as Lego Builders: Assembling Malice from Benign Blocks via Semantic Blueprints
Models as Lego Builders: Assembling Malice from Benign Blocks via Semantic Blueprints explores StructAttack is a jailbreak framework that exploits LVLMs' reasoning to assemble benign-looking visual prompts into coherent harmful semantics, demonstrating a novel vulnerability via semantic slot filling.. Commercial viability score: 7/10 in Vision-Language Models Security.
BUILDER'S SANDBOX
Build This Paper
Use an AI coding agent to implement this research.
OpenAI CodexAI Agent
Lightweight coding agent in your terminal.
Claude CodeAI Agent
Agentic coding tool for terminal workflows.
AntiGravity IDEScaffolding
AI agent mindset installer and workflow scaffolder.
CursorIDE
AI-first code editor built on VS Code.
VS CodeIDE
Free, open-source editor by Microsoft.
Recommended Stack
Startup Essentials
Estimated $11K - $15K over 6-10 weeks.
See exactly what it costs to build this -- with 3 comparable funded startups.
7-day free trial. Cancel anytime.
Discover the researchers behind this paper and find similar experts.
7-day free trial. Cancel anytime.
Related Papers
- Builds On ThisRobustness of Vision Language Models Against Split-Image Harmful Input AttacksViability: 5.0
- Builds On ThisRisk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising UtilityViability: 6.0
- Builds On ThisJailbreaks on Vision Language Model via Multimodal ReasoningViability: 3.0
- Builds On ThisStructured Visual Narratives Undermine Safety Alignment in Multimodal Large Language ModelsViability: 4.0
- Builds On ThisStructured Semantic Cloaking for Jailbreak Attacks on Large Language Models