List your company

Direct Malicious on VLSafe OOD

90.67ASRSelfReminder
-1.94222.101546.14570.1885Apr 7, 2026

Evaluation results

MethodLinks
SelfReminderBase Model=LLaVA-1.5-13B2026.0490.67
JailGuardBase Model=LLaVA-1.5-13B2026.0472.43
CIDERBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.0450
CIDERBase Model=LLaVA-1.5-13B2026.0450
JailGuardBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.0431.35
MirrorCheckBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.0426.33
MirrorCheckBase Model=LLaVA-1.5-13B2026.0426.33
ECSOBase Model=LLaVA-1.5-13B2026.0424
ECSOBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.0419.63
VLMGuardBase Model=LLaVA-1.5-13B2026.0415.27
VLMGuardBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=Internal Defense2026.0410.32
SelfReminderBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.048.55
ASTRABase Model=LLaVA-1.5-13B2026.048.03
ASTRABase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=Internal Defense2026.043.7
VLMShieldBase VLM=Qwen2.5-VL-7B-Instruct, Defense Category=External Defense2026.041.62
VLMShieldBase Model=LLaVA-1.5-13B2026.041.62