Colluding LoRA: A Composite Attack on LLM Safety Alignment
Colluding LoRA: A Composite Attack on LLM Safety Alignment explores Colluding LoRA presents a novel attack on LLM safety alignment through composition-triggered vulnerabilities.. Commercial viability score: 2/10 in LLM Safety.
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 $10K - $14K 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
- Higher ViabilityWeight space Detection of Backdoors in LoRA AdaptersViability: 6.0
- Higher ViabilityWhen Safe Models Merge into Danger: Exploiting Latent Vulnerabilities in LLM FusionViability: 7.0
- Higher ViabilityMoLoRA: Composable Specialization via Per-Token Adapter RoutingViability: 8.0
- Higher ViabilityCascade: Composing Software-Hardware Attack Gadgets for Adversarial Threat Amplification in Compound AI SystemsViability: 4.0
- Higher ViabilityLAAF: Logic-layer Automated Attack Framework A Systematic Red-Teaming Methodology for LPCI Vulnerabilities in Agentic Large Language Model Systems