List your company

Jailbreak Attack on AdvBench-50 + Malicious Instruct

100ASRStegoAttack
-4235077May 22, 2025

Evaluation results

MethodLinks
StegoAttackTarget Model=Gemini-32025.051004.974.71
StegoAttackTarget Model=DS-V3.22025.051004.94.89
StegoAttackTarget Model=Qwen32025.0599.334.994.44
DRATarget Model=DS-V3.22025.05984.914.84
StegoAttackTarget Model=GPT-52025.0582.673.982.44
DRATarget Model=Qwen32025.05824.123.92
JailbrokenTarget Model=DS-V3.22025.0572.673.333.72
DrAttackTarget Model=DS-V3.22025.0556.672.623.69
DrAttackTarget Model=Gemini-32025.0539.332.891.81
FFATarget Model=Qwen32025.0534.671.533.4
FFATarget Model=DS-V3.22025.0530.671.473.77
AutoDANTarget Model=Qwen32025.0529.331.753.57
DrAttackTarget Model=Qwen32025.0527.331.593.1
DRATarget Model=Gemini-32025.05201.623.27
JailbrokenTarget Model=Gemini-32025.05141.352.85
ArtPromptTarget Model=DS-V3.22025.0513.331.791.75
ArtPromptTarget Model=Gemini-32025.05121.272.41
CipherTarget Model=DS-V3.22025.05121.392.2
FFATarget Model=Gemini-32025.0510.671.253.25
JailbrokenTarget Model=Qwen32025.0510.671.242.69
CipherTarget Model=Qwen32025.057.331.232.89
ArtPromptTarget Model=Qwen32025.054.671.133.03
VanillaTarget Model=Gemini-32025.0541.053.88
GCGTarget Model=Gemini-32025.053.331.053.36
GCGTarget Model=Qwen32025.052.671.053.52
DrAttackTarget Model=GPT-52025.052.671.054.03
GCGTarget Model=DS-V3.22025.0521.033.6
AutoDANTarget Model=Gemini-32025.0521.033.57
DRATarget Model=GPT-52025.0521.134.09
CipherTarget Model=Gemini-32025.051.331.061.87
VanillaTarget Model=GPT-52025.050.671.014.4
JailbrokenTarget Model=GPT-52025.050.671.014.04
VanillaTarget Model=DS-V3.22025.05013.56
VanillaTarget Model=Qwen32025.05013.72
GCGTarget Model=GPT-52025.0501.493.89
FFATarget Model=GPT-52025.05012.97
AutoDANTarget Model=GPT-52025.05014
AutoDANTarget Model=DS-V3.22025.05013.41
ArtPromptTarget Model=GPT-52025.05012.92
CipherTarget Model=GPT-52025.05013.17