List your company
ResearchFormerly WizWand

Jailbreaking on JBB-Behaviors (test)

99ASR (GPT-4o)HMNS
39.254.72570.2585.775Apr 11, 2026
Updated 6mo ago

Evaluation results

MethodLinks
HMNSTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0499941.9
HMNSTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.0499961.8
HMNSTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0494881.9
ArrAttackTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.049496.27.2
ArrAttackTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0493887.3
ArrAttackTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0489887.8
TempestTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.0486.481.28.9
TempestTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0485.279.49.1
TempestTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0483.277.19.4
PrisonBreakTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.0479.874.311.1
PrisonBreakTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0478.572.611.2
AutoDANTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.0475.269.312
PrisonBreakTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0474.468.212.1
AutoDANTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0473.467.512.5
MasterKeyTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.0472.967.110.1
MasterKeyTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0471.866.110.2
ADCTarget LLM=LLaMA-3.1-70B, Zero-shot=true2026.0471.565.89.3
ADCTarget LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0469.363.89.7
Many-shot JB (MSJ)Target LLM=LLaMA-3.1-70B, Zero-shot=true2026.0468.462.810.3
AutoDANTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0466.359.913.1
Many-shot JB (MSJ)Target LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.046660.110.6
MasterKeyTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0463.457.510.9
ADCTarget LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0462.556.810.5
Many-shot JB (MSJ)Target LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0460.154.411.5
Foot-In-The-Door (FITD)Target LLM=LLaMA-3.1-70B, Zero-shot=true2026.0447.641.915.2
Foot-In-The-Door (FITD)Target LLM=LLaMA-2-7B-Chat, Zero-shot=true2026.0445.139.215.9
Foot-In-The-Door (FITD)Target LLM=Phi-3-Medium-14B (Instruct), Zero-shot=true2026.0441.535.916.4