List your company

Jailbreak Safety Evaluation on PAP

2ASRSafePath
1.286.141115.86Jun 15, 2026

Evaluation results

MethodLinks
SafePathBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.062
ST-DBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.062
ST-SBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.062
ST-DBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.062
ST-DBackbone Model=Qwen3-8B2026.064
ST-SBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.064
ST-DBackbone Model=Qwen3-32B2026.064
ST-SBackbone Model=Qwen3-8B2026.0610
Star1Backbone Model=Qwen3-32B2026.0610
SafePathBackbone Model=Qwen3-32B2026.0610
ST-SBackbone Model=Qwen3-32B2026.0610
BaseBackbone Model=Qwen3-8B2026.0612
Star1Backbone Model=Qwen3-8B2026.0612
SafePathBackbone Model=Qwen3-8B2026.0612
BaseBackbone Model=Qwen3-32B2026.0612
SafePathBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.0612
Star1Backbone Model=DeepSeek-R1-Distill-Llama-8B2026.0614
Star1Backbone Model=DeepSeek-R1-Distill-Llama-70B2026.0614
BaseBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.0618
BaseBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.0620