List your company

Jailbreak Attack on JailbreakBench (ASR Variants and Query Count)

97ASR1LASH
-0.7624.625075.38May 20, 2026

Evaluation results

MethodLinks
LASHTarget Model=Llama2 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05978025.84
LASHTarget Model=Llama3 8B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05948920.79
LASHTarget Model=GPT-4o mini, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05918525.53
LASHTarget Model=Gemini 2-flash, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05907927.99
FlipAttackTarget Model=GPT-4o mini, Temperature=0, Top-p=1, Maximum target-query budget=1602026.058953—
AutoDAN-TTarget Model=GPT-4o mini, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05886228.55
FlipAttackTarget Model=Llama3 8B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.058751—
FlipAttackTarget Model=Gemini 2-flash, Temperature=0, Top-p=1, Maximum target-query budget=1602026.058681—
LASHTarget Model=Vicuna 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05858317.52
LASHTarget Model=Average, Temperature=0, Top-p=1, Maximum target-query budget=1602026.0584.574.530.01
AutoDAN-TTarget Model=Llama3 8B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05825038.62
PAIRTarget Model=GPT-4o mini, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05815638.4
AutoDANTarget Model=Vicuna 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.058170—
FlipAttackTarget Model=Vicuna 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.058130—
PAIRTarget Model=Vicuna 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05774934.24
AutoDAN-TTarget Model=Gemini 2-flash, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05774020.39
PAIRTarget Model=Gemini 2-flash, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05734046.18
TAPTarget Model=Llama3 8B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.0572684.57
FlipAttackTarget Model=Average, Temperature=0, Top-p=1, Maximum target-query budget=1602026.0566.8338.83—
AutoDAN-TTarget Model=Average, Temperature=0, Top-p=1, Maximum target-query budget=1602026.0561.6734.8336.21
TAPTarget Model=GPT-4o mini, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05585333.87
PAIRTarget Model=Llama3 8B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.055718107.91
TAPTarget Model=Gemini 2-flash, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05575110.7
AutoDAN-TTarget Model=Vicuna 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05573826
PAIRTarget Model=Average, Temperature=0, Top-p=1, Maximum target-query budget=1602026.055227.590.51
AutoDAN-TTarget Model=GPT-5.4 nano, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05501268.21
LASHTarget Model=GPT-5.4 nano, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05503162.36
TAPTarget Model=Average, Temperature=0, Top-p=1, Maximum target-query budget=1602026.054740.6755.11
TAPTarget Model=Llama2 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05453161.85
FlipAttackTarget Model=GPT-5.4 nano, Temperature=0, Top-p=1, Maximum target-query budget=1602026.054110—
AutoDANTarget Model=Average, Temperature=0, Top-p=1, Maximum target-query budget=1602026.053629.67—
TAPTarget Model=GPT-5.4 nano, Temperature=0, Top-p=1, Maximum target-query budget=1602026.052924107.73
PAIRTarget Model=Llama2 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05211156.31
TAPTarget Model=Vicuna 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.052117111.93
AutoDANTarget Model=Llama2 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.051711—
FlipAttackTarget Model=Llama2 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05178—
AutoDAN-TTarget Model=Llama2 7B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.0516735.46
AutoDANTarget Model=Llama3 8B, Temperature=0, Top-p=1, Maximum target-query budget=1602026.05108—
PAIRTarget Model=GPT-5.4 nano, Temperature=0, Top-p=1, Maximum target-query budget=1602026.0531160