Systemic Risk Detection on AdvWeb
0Prompt Injection (PI)Claude-3.5
Evaluation results
| Method | Links | |
|---|---|---|
| Claude-3.5Framework Category=Model-based, Shots=0-Shot2025.02 | 0 | |
| AgentMonitorFramework Category=Guardrail-based, Backbone=GPT-4o2025.02 | 0 | |
| AGrailFramework Category=Guardrail-based, Backbone=Claude-3.52025.02 | 0 | |
| GPT-4oFramework Category=Model-based, Shots=0-Shot2025.02 | 500 | |
| AGrailFramework Category=Guardrail-based, Backbone=GPT-4o2025.02 | 880 | |
| LLaMA-Guard 3Framework Category=Guardrail-based2025.02 | 10,000 |