List your company

Jailbreaking on Harmful prompt (test)

0Jailbreak Success RateUS
-3.7621.624772.38Apr 22, 2026Apr 30, 2026May 9, 2026May 18, 2026May 27, 2026Jun 5, 2026Jun 14, 2026

Evaluation results

MethodLinks
USBackbone=GPT-oss-120B, Jailbreaking Coefficient=NA2026.040
RepEBackbone=GPT-oss-120B, Jailbreaking Coefficient=NA2026.040
RepEBackbone=Phi4-14B, Jailbreaking Coefficient=72026.040
USBackbone=Phi4-3.8B, Jailbreaking Coefficient=2.752026.041
USBackbone=Qwen3-0.6B, Jailbreaking Coefficient=1.82026.042
GAS-Leak-LLMTarget Model=Llama 3.2 3B Instruct, Suffix Source Model=Qwen2.5-3B2026.0613.8
GAS-Leak-LLMTarget Model=Llama 3.2 3B Instruct, Suffix Source Model=Llama-3.2-3B Instruct2026.0617.76
USBackbone=Phi4-14B, Jailbreaking Coefficient=2.52026.0455
RepEBackbone=Llama-3.1-8B, Jailbreaking Coefficient=0.92026.0457
USBackbone=Qwen3-32B, Jailbreaking Coefficient=92026.0457
RepEBackbone=Qwen3-0.6B, Jailbreaking Coefficient=0.72026.0459
RepEBackbone=Phi4-3.8B, Jailbreaking Coefficient=32026.0467
RepEBackbone=GPT-oss-20B, Jailbreaking Coefficient=2002026.0476
RepEBackbone=Qwen3-32B, Jailbreaking Coefficient=82026.0482
RepEBackbone=Llama-3.3-70B-4bt, Jailbreaking Coefficient=0.52026.0483
USBackbone=Llama-3.1-8B, Jailbreaking Coefficient=0.62026.0486
GAS-Leak-LLMTarget Model=Qwen 2.5 3B, Suffix Source Model=Llama-3.2-3B Instruct2026.0688.99
USBackbone=Llama-3.3-70B-4bt, Jailbreaking Coefficient=0.32026.0491
GAS-Leak-LLMTarget Model=Qwen 2.5 3B, Suffix Source Model=Qwen2.5-3B2026.0691.63
USBackbone=GPT-oss-20B, Jailbreaking Coefficient=492026.0494