Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Jailbreak attacks on MML-Mirror (MML-M)
Loading...
97
Safety Rate
TRR
-3.152
22.849
48.85
74.851
Dec 8, 2025
Safety Rate
Updated 4d ago
Evaluation Results
Method
Method
Links
Safety Rate
TRR
Backbone=Qwen2.5VL-7B,...
2025.12
97
TRR
Backbone=Qwen2.5VL-32B...
2025.12
80.5
TiS
Backbone=Qwen2.5VL-32B...
2025.12
65
MSR-Align
Backbone=Qwen2.5VL-32B...
2025.12
55.7
MSR-Align
Backbone=Qwen2.5VL-7B,...
2025.12
53.7
Claude-3.5-Sonnet
Model=Claude-3.5-Sonnet
2025.12
40
TiS
Backbone=Qwen2.5VL-7B,...
2025.12
28.8
Gemini-2.5-pro
Model=Gemini-2.5-pro
2025.12
15.7
Safework-R1-7B
Model=Safework-R1-7B
2025.12
12.5
Qwen2.5VL-7B
Backbone=Qwen2.5VL-7B,...
2025.12
6.5
GPT-4o
Model=GPT-4o
2025.12
1.9
Qwen2.5VL-32B
Backbone=Qwen2.5VL-32B...
2025.12
0.7
Feedback
Search any
task
Search any
task