Image-based Jailbreak on FigStep OOD (ASR%)
0ASRVLMShield
Evaluation Results
| Method | Links | |
|---|---|---|
| VLMShieldBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.04 | 0 | |
| VLMShieldBase Model=LLaVA-1.5-13B2026.04 | 0 | |
| ASTRABase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=Internal Defense2026.04 | 1.8 | |
| ASTRABase Model=LLaVA-1.5-13B2026.04 | 7.33 | |
| VLMGuardBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=Internal Defense2026.04 | 9.46 | |
| JailGuardBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.04 | 13.82 | |
| VLMGuardBase Model=LLaVA-1.5-13B2026.04 | 13.83 | |
| MirrorCheckBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.04 | 15.36 | |
| MirrorCheckBase Model=LLaVA-1.5-13B2026.04 | 15.36 | |
| JailGuardBase Model=LLaVA-1.5-13B2026.04 | 20.3 | |
| ECSOBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.04 | 26.72 | |
| SelfReminderBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.04 | 27.8 | |
| ECSOBase Model=LLaVA-1.5-13B2026.04 | 29.05 | |
| CIDERBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.04 | 40.03 | |
| CIDERBase Model=LLaVA-1.5-13B2026.04 | 40.03 | |
| SelfReminderBase Model=LLaVA-1.5-13B2026.04 | 58 |