Share your thoughts, 1 month free Claude Pro on usSee more

Safety Evaluation on Harmfulness Evaluation Sequences

0.79Harmfulness Score

llama2-13b-chat

Updated 3mo ago

Evaluation Results

Method	Links
llama2-13b-chat 2024.05		0.79
Qwen-4B-chat 2024.05		0.77
llama2-13b-chat 2024.05		0.76
Qwen-1.8B-chat 2024.05		0.76
Qwen-1.8B-chat 2024.05		0.76
Qwen-4B-chat 2024.05		0.76
Qwen-7B-chat 2024.05		0.76
Qwen-7B-chat 2024.05		0.76
Qwen-7B-chat 2024.05		0.76
llama2-13b-chat 2024.05		0.75
llama2-13b-chat 2024.05		0.75
Qwen-1.8B-chat 2024.05		0.75
Qwen-1.8B-chat 2024.05		0.75
Qwen-4B-chat 2024.05		0.75
Qwen-4B-chat 2024.05		0.75
llama2-7b-chat 2024.05		0.74
llama2-7b-chat 2024.05		0.74
Qwen-1.8B-chat 2024.05		0.74
Qwen-4B-chat 2024.05		0.74
Qwen-7B-chat 2024.05		0.74
llama2-7b-chat 2024.05		0.73
llama2-7b-chat 2024.05		0.73
llama2-13b-chat 2024.05		0.73
llama2-13b-chat 2024.05		0.73
Qwen-0.5B-chat 2024.05		0.73
llama2-7b-chat 2024.05		0.72
Qwen-1.8B-chat 2024.05		0.72
Qwen-4B-chat 2024.05		0.72
Qwen-7B-chat 2024.05		0.72
Qwen-0.5B-chat 2024.05		0.71
Qwen-0.5B-chat 2024.05		0.7
Qwen-7B-chat 2024.05		0.7
Qwen-0.5B-chat 2024.05		0.68
Qwen-0.5B-chat 2024.05		0.66
Qwen-0.5B-chat 2024.05		0.5
llama2-7b-chat 2024.05		0.47
RepNoise (llama2-7b-chat) 2024.05		0.13
RepNoise (llama2-7b-chat) 2024.05		0.12
RepNoise (llama2-7b-chat) 2024.05		0.12
RepNoise (Qwen-0.5B-chat) 2024.05		0.12
RepNoise (llama2-7b-chat) 2024.05		0.11
RepNoise (llama2-7b-chat) 2024.05		0.1
RepNoise (Qwen-0.5B-chat) 2024.05		0.1
RepNoise (Qwen-4B-chat) 2024.05		0.1
RepNoise (Qwen-0.5B-chat) 2024.05		0.09
RepNoise (Qwen-4B-chat) 2024.05		0.09
RepNoise (llama2-7b-chat) 2024.05		0.08
Qwen-7B-chat 2024.05		0.08
llama2-13b-chat 2024.05		0.07
RepNoise (Qwen-0.5B-chat) 2024.05		0.07
RepNoise (Qwen-0.5B-chat) 2024.05		0.06
llama2-7b-chat 2024.05		0.05
RepNoise (llama2-7b-chat) 2024.05		0.05
RepNoise (llama2-13b-chat) 2024.05		0.05
RepNoise (Qwen-4B-chat) 2024.05		0.04
RepNoise (Qwen-1.8B-chat) 2024.05		0.02
RepNoise (Qwen-4B-chat) 2024.05		0.02
RepNoise (Qwen-4B-chat) 2024.05		0.02
RepNoise (Qwen-0.5B-chat) 2024.05		0.01
RepNoise (Qwen-1.8B-chat) 2024.05		0.01
RepNoise (Qwen-7B-chat) 2024.05		0.01
RepNoise (Qwen-7B-chat) 2024.05		0.01
RepNoise (Qwen-7B-chat) 2024.05		0.01
RepNoise (llama2-13b-chat) 2024.05		0
RepNoise (llama2-13b-chat) 2024.05		0
RepNoise (llama2-13b-chat) 2024.05		0
RepNoise (llama2-13b-chat) 2024.05		0
RepNoise (llama2-13b-chat) 2024.05		0
RepNoise (llama2-13b-chat) 2024.05		0
Qwen-0.5B-chat 2024.05		0
RepNoise (Qwen-0.5B-chat) 2024.05		0
Qwen-1.8B-chat 2024.05		0
RepNoise (Qwen-1.8B-chat) 2024.05		0
RepNoise (Qwen-1.8B-chat) 2024.05		0
RepNoise (Qwen-1.8B-chat) 2024.05		0
RepNoise (Qwen-1.8B-chat) 2024.05		0
RepNoise (Qwen-1.8B-chat) 2024.05		0
Qwen-4B-chat 2024.05		0
RepNoise (Qwen-4B-chat) 2024.05		0
RepNoise (Qwen-4B-chat) 2024.05		0
RepNoise (Qwen-7B-chat) 2024.05		0
RepNoise (Qwen-7B-chat) 2024.05		0
RepNoise (Qwen-7B-chat) 2024.05		0
RepNoise (Qwen-7B-chat) 2024.05		0