Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Safety Evaluation on JailbreakLLMs Orig.
Loading...
0
Unsafe Rate
VLGuard PH
-0.8608
4.9496
10.76
16.5704
Oct 11, 2024
Unsafe Rate
Updated 4d ago
Evaluation Results
Method
Method
Links
Unsafe Rate
VLGuard PH
Backbone=LLaVA-v1.5-13B
2024.10
0
VLGuard Mixed
Backbone=LLaVA-v1.5-13B
2024.10
0.38
VLGuard Mixed
Backbone=LLaVA-v1.5-7B
2024.10
0.76
CMRM_sample
Backbone=ShareGPT4V
2024.10
1.14
LLaVA-v1.5-13B
Input Protocol=Caption
2024.10
1.82
VLGuard PH
Backbone=LLaVA-v1.5-7B
2024.10
2.27
LLaVA-v1.5-13B
Input Protocol=Query
2024.10
3.03
CMRM_sample
Backbone=LLaVA-v1.5-13B
2024.10
3.03
CMRM_dataset
Backbone=ShareGPT4V
2024.10
3.79
CMRM_sample
Backbone=LLaVA-v1.5-7B
2024.10
4.55
LLaVA-v1.5-7B
Input Protocol=Caption
2024.10
4.85
CMRM_dataset
Backbone=LLaVA-v1.5-13B
2024.10
4.92
CMRM_dataset
Backbone=LLaVA-v1.5-7B
2024.10
8.33
ShareGPT4V
Input Protocol=Caption
2024.10
8.33
ShareGPT4V
Input Protocol=Query
2024.10
10.23
LLaVA-v1.5-13B
Input Protocol=Visual...
2024.10
12.42
LLaVA-v1.5-7B
Input Protocol=Query
2024.10
12.73
ShareGPT4V
Input Protocol=Visual...
2024.10
19.32
LLaVA-v1.5-7B
Input Protocol=Visual...
2024.10
21.52
Feedback
Search any
task
Search any
task