Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Safety Detection on MAS Qwen-2.5-7B (Current)
Loading...
97.46
Accuracy
AgentLens
47.2176
60.2613
73.305
86.3487
Jun 21, 2026
Accuracy
F1 Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
F1 Score
AgentLens
Method Category=Interp...
2026.06
97.46
96.25
LLaMA-Guard-4-12B
Method Category=Guardr...
2026.06
94.92
92.77
Deepseek-V4-Pro
Method Category=LLM-as...
2026.06
90.68
88.17
Gemini-3-Pro
Method Category=LLM-as...
2026.06
90.25
87.7
Deepseek-V4-Flash
Method Category=LLM-as...
2026.06
89.83
87.1
Claude Sonnet-4.6
Method Category=LLM-as...
2026.06
87.29
84.54
AGrail (GPT-4o)
Method Category=Guardr...
2026.06
85.17
78.76
Claude Opus-4.6
Method Category=LLM-as...
2026.06
82.63
80
GPT-4o
Method Category=LLM-as...
2026.06
68.22
68.62
GPT-5.2
Method Category=LLM-as...
2026.06
49.15
57.75
Feedback
Search any
task
Search any
task