List your company

Harmful Query on JailbreakB

0.67ASRNSPO
-1.183211.325923.83536.3441Dec 12, 2025

Evaluation results

MethodLinks
NSPOBackbone=Qwen2.5-7B-Instruct2025.120.67
NSPOBackbone=Llama3-8B-Instruct2025.121.33
BFPOBackbone=Qwen2.5-7B-Instruct2025.121.33
Llama3-8B-InstructBackbone=Llama3-8B-Instruct2025.122
MoCANBackbone=Qwen2.5-7B-Instruct2025.122
PeCANBackbone=Llama3-8B-Instruct2025.122.33
DPO-HBackbone=Qwen2.5-7B-Instruct2025.122.33
DPO-SBackbone=Llama3-8B-Instruct2025.122.67
MoCANBackbone=Llama3-8B-Instruct2025.122.67
BFPOBackbone=Llama3-8B-Instruct2025.122.67
DPO-SBackbone=Qwen2.5-7B-Instruct2025.122.67
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct2025.123
W-DOORBackbone=Qwen2.5-7B-Instruct2025.123.33
W-DOORBackbone=Llama3-8B-Instruct2025.124.33
DPO-MixBackbone=Qwen2.5-7B-Instruct2025.124.33
PeCANBackbone=Qwen2.5-7B-Instruct2025.125.67
DPO-MixBackbone=Llama3-8B-Instruct2025.126.67
SafeRLHFBackbone=Llama3-8B-Instruct2025.1216.33
SafeRLHFBackbone=Qwen2.5-7B-Instruct2025.1225
DPO-HBackbone=Llama3-8B-Instruct2025.1247