List your company

Safety Evaluation on Llama-Guard 3-8B

2.38ASRMLP
-0.522419.068838.6658.2512May 6, 2026

Evaluation results

MethodLinks
MLPBase Model=Qwen2.5-72B, Condition=Cured2026.052.38
MLPBase Model=Llama-3.1-70B, Condition=Cured2026.054
MLPBase Model=Qwen2.5-7B, Condition=Cured2026.054.62
MoEBase Model=Gemma-2-27B, Condition=Cured2026.055.07
MoEBase Model=Gemma-2-9b, Condition=Cured2026.055.75
MoEBase Model=Llama-3.1-8B, Condition=Cured2026.057.19
MLPBase Model=Gemma-2-27B, Condition=Cured2026.059.12
MoEBase Model=Mistral-7B-v0.3, Condition=Cured2026.059.31
FMBase Model=Gemma-2-9b, Condition=Cured2026.0511.94
FMBase Model=Gemma-2-27B, Condition=Cured2026.0512.69
MLPBase Model=Gemma-2-9b, Condition=Cured2026.0516.94
MoEBase Model=Llama-3.1-70B, Condition=Cured2026.0517
MLPBase Model=Mistral-7B-v0.3, Condition=Cured2026.0518.38
FMBase Model=Llama-3.1-8B, Condition=Cured2026.0518.69
CVAEBase Model=Qwen2.5-72B, Condition=Cured2026.0518.86
MoEBase Model=Qwen2.5-72B, Condition=Cured2026.0519.06
MoEBase Model=Qwen2.5-7B, Condition=Cured2026.0522.2
MLPBase Model=Llama-3.1-8B, Condition=Cured2026.0524.12
CVAEBase Model=Gemma-2-9b, Condition=Cured2026.0530.25
Safe LoRABase Model=Qwen2.5-72B, Condition=Cured2026.0531.93
FMBase Model=Qwen2.5-7B, Condition=Cured2026.0534.11
FMBase Model=Mistral-7B-v0.3, Condition=Cured2026.0536.56
AEBase Model=Gemma-2-27B, Condition=Cured2026.0536.65
CVAEBase Model=Qwen2.5-7B, Condition=Cured2026.0536.74
Safe LoRABase Model=Gemma-2-9b, Condition=Cured2026.0539.06
FMBase Model=Qwen2.5-72B, Condition=Cured2026.0539.38
CVAEBase Model=Mistral-7B-v0.3, Condition=Cured2026.0540.75
AEBase Model=Gemma-2-9b, Condition=Cured2026.0541.21
FMBase Model=Llama-3.1-70B, Condition=Cured2026.0541.75
CVAEBase Model=Llama-3.1-8B, Condition=Cured2026.0542.12
Safe LoRABase Model=Llama-3.1-8B, Condition=Cured2026.0543.75
Safe LoRABase Model=Mistral-7B-v0.3, Condition=Cured2026.0544
CVAEBase Model=Gemma-2-27B, Condition=Cured2026.0544.97
AEBase Model=Qwen2.5-72B, Condition=Cured2026.0548.25
Safe LoRABase Model=Qwen2.5-7B, Condition=Cured2026.0549.75
AEBase Model=Llama-3.1-8B, Condition=Cured2026.0551.76
AEBase Model=Qwen2.5-7B, Condition=Cured2026.0553.04
CVAEBase Model=Llama-3.1-70B, Condition=Cured2026.0553.37
Safe LoRABase Model=Gemma-2-27B, Condition=Cured2026.0555.62
Safe LoRABase Model=Llama-3.1-70B, Condition=Cured2026.0556.38
AEBase Model=Llama-3.1-70B, Condition=Cured2026.0558.88
AEBase Model=Mistral-7B-v0.3, Condition=Cured2026.0559.88
Unsafe BaselineBase Model=Qwen2.5-72B, Condition=Unsafe (Finetuned)2026.0560.25
SaLoRABase Model=Qwen2.5-72B, Condition=Cured2026.0560.75
Unsafe BaselineBase Model=Gemma-2-9b, Condition=Unsafe (Finetuned)2026.0563.04
SaLoRABase Model=Gemma-2-9b, Condition=Cured2026.0563.23
Unsafe BaselineBase Model=Llama-3.1-70B, Condition=Unsafe (Finetuned)2026.0565.81
SaLoRABase Model=Llama-3.1-70B, Condition=Cured2026.0565.94
Unsafe BaselineBase Model=Gemma-2-27B, Condition=Unsafe (Finetuned)2026.0566.37
Unsafe BaselineBase Model=Llama-3.1-8B, Condition=Unsafe (Finetuned)2026.0566.85
SaLoRABase Model=Gemma-2-27B, Condition=Cured2026.0567.29
SaLoRABase Model=Llama-3.1-8B, Condition=Cured2026.0567.46
Unsafe BaselineBase Model=Qwen2.5-7B, Condition=Unsafe (Finetuned)2026.0573.36
SaLoRABase Model=Qwen2.5-7B, Condition=Cured2026.0573.47
Unsafe BaselineBase Model=Mistral-7B-v0.3, Condition=Unsafe (Finetuned)2026.0574.25
SaLoRABase Model=Mistral-7B-v0.3, Condition=Cured2026.0574.94