Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Attack Success Rate Evaluation on 40 diverse smartphone tasks Persuade subcategory
Loading...
100
Target ASR
HG-IDA*
-4
23
50
77
Oct 9, 2025
Target ASR
Reduction ASR
Updated 9d ago
Evaluation Results
Method
Method
Links
Target ASR
Reduction ASR
HG-IDA*
Target Model=Gemini-2....
2025.10
100
93.3
HG-IDA*
Target Model=GPT-4o
2025.10
80
73.3
HG-IDA*
Target Model=Deepseek-VL2
2025.10
80
20
DA
Target Model=Gemini-2....
2025.10
66.7
33.3
Prefix
Target Model=Gemini-2....
2025.10
53.3
33.3
GCG
Target Model=Gemini-2....
2025.10
40
13.3
HG-IDA*
Target Model=LLaVA-One...
2025.10
40
33.3
DA
Target Model=LLaVA-One...
2025.10
20
20
DA
Target Model=Deepseek-VL2
2025.10
6.7
6.7
DA
Target Model=GPT-4o
2025.10
0
0
Prefix
Target Model=GPT-4o
2025.10
0
0
GCG
Target Model=GPT-4o
2025.10
0
0
Prefix
Target Model=Deepseek-VL2
2025.10
0
0
GCG
Target Model=Deepseek-VL2
2025.10
0
0
Prefix
Target Model=LLaVA-One...
2025.10
0
0
GCG
Target Model=LLaVA-One...
2025.10
0
0
Feedback
Search any
task
Search any
task