List your company

Reading Comprehension on SQuAD (Usage/Attack Metrics)

75.91Attack AccuracyNo-shield
1.560420.862740.16559.4673Oct 16, 2024

Evaluation results

MethodLinks
No-shieldBackbone=LLaMA3-8B2024.1075.91——
NPLOBackbone=LLaMA3-8B2024.1075.2——
DarkneTZBackbone=LLaMA3-8B2024.1073.81——
SOTERBackbone=LLaMA3-8B2024.1069.12——
NPLOBackbone=ChatGLM3-6B2024.1068.21——
No-shieldBackbone=ChatGLM3-6B2024.1068.13——
SerdabBackbone=LLaMA3-8B2024.1067.92——
SerdabBackbone=ChatGLM3-6B2024.1066.28——
No-shieldBackbone=Gemma2-2B2024.1063.96——
NPLOBackbone=Gemma2-2B2024.1063.94——
DarkneTZBackbone=ChatGLM3-6B2024.1063.91——
SOTERBackbone=ChatGLM3-6B2024.1062.61——
DarkneTZBackbone=Gemma2-2B2024.1061.02——
SerdabBackbone=Gemma2-2B2024.1060.82——
SOTERBackbone=Gemma2-2B2024.1057.87——
NPLOBackbone=Qwen2-0.5B2024.1032.33——
No-shieldBackbone=Qwen2-0.5B2024.1030.54——
DarkneTZBackbone=Qwen2-0.5B2024.1029.89——
SerdabBackbone=Qwen2-0.5B2024.1028.42——
SOTERBackbone=Qwen2-0.5B2024.1026.34——
TLGBackbone=LLaMA3-8B2024.1011.94——
DTEBackbone=LLaMA3-8B2024.1010.48——
CoreGuardBackbone=LLaMA3-8B2024.1010.01073.02
Black-boxBackbone=LLaMA3-8B2024.109.71——
ShadowNetBackbone=LLaMA3-8B2024.109.64——
ShadowNetBackbone=ChatGLM3-6B2024.109.56——
DTEBackbone=ChatGLM3-6B2024.109.42——
Black-boxBackbone=ChatGLM3-6B2024.109.15——
CoreGuardBackbone=ChatGLM3-6B2024.108.98062.11
Black-boxBackbone=Gemma2-2B2024.108.81——
TLGBackbone=ChatGLM3-6B2024.108.61——
TLGBackbone=Gemma2-2B2024.107.91——
CoreGuardBackbone=Gemma2-2B2024.107.81043.21
DTEBackbone=Gemma2-2B2024.107.51——
CoreGuardBackbone=Qwen2-0.5B2024.107.35016.5
TLGBackbone=Qwen2-0.5B2024.106.81——
ShadowNetBackbone=Gemma2-2B2024.106.71——
ShadowNetBackbone=Qwen2-0.5B2024.105.93——
Black-boxBackbone=Qwen2-0.5B2024.105.66——
DTEBackbone=Qwen2-0.5B2024.104.42——