Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Jailbreak Defense on AdvBench (GCG Attack)
Loading...
100
DSR
SmoothLLM
4.32
29.16
54
78.84
Feb 26, 2024
DSR
Updated 1mo ago
Evaluation Results
Method
Method
Links
DSR
SmoothLLM
Target Model=GPT-3.5-t...
2024.02
100
Paraphrasing
Target Model=GPT-3.5-t...
2024.02
100
Backtranslation
Target Model=GPT-3.5-t...
2024.02
100
Response check
Target Model=Llama-2-13B
2024.02
100
Backtranslation
Target Model=Llama-2-13B
2024.02
100
SmoothLLM
Target Model=Llama-2-13B
2024.02
98
Paraphrasing
Target Model=Llama-2-13B
2024.02
98
Backtranslation
Target Model=Vicuna-13B
2024.02
98
No defense
Target Model=GPT-3.5-t...
2024.02
94
Response check
Target Model=GPT-3.5-t...
2024.02
94
SmoothLLM
Target Model=Vicuna-13B
2024.02
92
Paraphrasing
Target Model=Vicuna-13B
2024.02
84
No defense
Target Model=Llama-2-13B
2024.02
66
Response check
Target Model=Vicuna-13B
2024.02
30
No defense
Target Model=Vicuna-13B
2024.02
8
Feedback
Search any
task
Search any
task