Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safety Classification on Jailbreak prompts (test)
Loading...
100
Adv-suffix Accuracy
Llama Guard 3
90.64
93.07
95.5
97.93
Jul 16, 2025
Adv-suffix Accuracy
Persuasion Accuracy
Template Accuracy
Updated 18d ago
Evaluation Results
Method
Method
Links
Adv-suffix Accuracy
Persuasion Accuracy
Template Accuracy
Llama Guard 3
Base Model=LLAMA2-CHAT-7B
2025.07
100
0
76
Latent Guard
Base Model=LLAMA2-CHAT-7B
2025.07
100
41.6
100
Latent Guard
Base Model=QWEN2-INSTR...
2025.07
100
75
53.5
Llama Guard 3
Base Model=LLAMA3-INST...
2025.07
99.2
6.8
50
Llama Guard 3
Base Model=QWEN2-INSTR...
2025.07
97.8
17.8
91.4
Latent Guard
Base Model=LLAMA3-INST...
2025.07
91
65
100
Feedback
Search any
task
Search any
task