List your company

Malicious Prompt Detection on JailbreakV_28K Image-based (test)

0.19FNRVLMShield
-3.00418.555540.11561.6745Apr 7, 2026

Evaluation results

MethodLinks
VLMShieldTarget VLM Architecture=LLaVA-1.5-13B2026.040.19
VLMShieldTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.040.19
ASTRATarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.042.14
ASTRATarget VLM Architecture=LLaVA-1.5-13B2026.045.21
VLMGuardTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.0411.82
JailGuardTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.0414
VLMGuardTarget VLM Architecture=LLaVA-1.5-13B2026.0416.37
MirrorCheckTarget VLM Architecture=LLaVA-1.5-13B2026.0417.19
MirrorCheckTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.0417.19
JailGuardTarget VLM Architecture=LLaVA-1.5-13B2026.0422.05
SelfReminderTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.0434.8
CIDERTarget VLM Architecture=LLaVA-1.5-13B2026.0437.2
CIDERTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.0437.2
ECSOTarget VLM Architecture=LLaVA-1.5-13B2026.0439.68
ECSOTarget VLM Architecture=Qwen2.5-VL-7B-Instruct2026.0443.06
SelfReminderTarget VLM Architecture=LLaVA-1.5-13B2026.0480.04