Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-turn Safety Risk Assessment on Multi-turn Safety 100 random sampled tasks
Loading...
0.15
ASR
ToolShield
0.1208
0.3179
0.515
0.7121
Feb 13, 2026
ASR
RR
Updated 1mo ago
Evaluation Results
Method
Method
Links
ASR
RR
ToolShield
Model=Claude-3.5-Sonnet
2026.02
0.15
0.83
Baseline
Model=Claude-3.5-Sonnet
2026.02
0.42
0.5
ToolShield
Model=Gemini-1.5-Flash
2026.02
0.57
0.31
Baseline
Model=Gemini-1.5-Flash
2026.02
0.69
0.16
Firewall
Model=Gemini-1.5-Flash
2026.02
0.79
0.11
w/o Defense
Model=Gemini-1.5-Flash
2026.02
0.81
0.08
Firewall
Model=Claude-3.5-Sonnet
2026.02
0.85
0.13
w/o Defense
Model=Claude-3.5-Sonnet
2026.02
0.88
0.1
Feedback
Search any
task
Search any
task