List your company

Safety Evaluation on HH-RLHF (test)

1.02Harm ScoreSAILS
0.94561.44781.952.4522Dec 29, 2025

Evaluation results

MethodLinks
SAILSModel=Gemma-2-9B, # Params=0.19%2025.121.020.9960.4
SAILSModel=Llama-3.1-8B, # Params=0.21%2025.121.030.9920.8
IT+RLModel=Gemma-2-9B, # Params=100%2025.121.080.9820
IT+RLModel=Llama-3.1-8B, # Params=100%2025.121.130.9782.6
SAILSModel=Gemma-2-2B, # Params=0.24%2025.121.170.9682.6
IT+RLModel=Gemma-2-2B, # Params=100%2025.121.180.9462.6
PromptModel=Gemma-2-9B, # Params=-2025.121.290.9424.6
FFTModel=Llama-3.1-8B, # Params=100%2025.121.340.9286
LoRAModel=Llama-3.1-8B, # Params=0.21%2025.121.40.925.2
LoRAModel=Gemma-2-9B, # Params=0.19%2025.121.410.9244.8
FFTModel=Gemma-2-9B, # Params=100%2025.121.440.9227
PromptModel=Llama-3.1-8B, # Params=-2025.121.470.9067.6
PromptModel=Gemma-2-2B, # Params=-2025.121.510.888.8
FFTModel=Gemma-2-2B, # Params=100%2025.121.520.8928.8
DoRAModel=Gemma-2-2B, # Params=0.25%2025.121.540.897.6
DoRAModel=Llama-3.1-8B, # Params=0.22%2025.121.540.8967.6
LoRAModel=Gemma-2-2B, # Params=0.24%2025.121.560.8768.4
DoRAModel=Gemma-2-9B, # Params=0.20%2025.121.670.8610.2
OriginalModel=Llama-3.1-8B, # Params=-2025.122.540.6235.2
OriginalModel=Gemma-2-9B, # Params=-2025.122.730.57637.4
OriginalModel=Gemma-2-2B, # Params=-2025.122.880.52842.4