Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Agent Action Safety Verification on internal benchmark 300-scenario
Loading...
95
Verdict Accuracy
AgentTrust v0.5
42.688
56.269
69.85
83.431
May 6, 2026
Verdict Accuracy
False Positive Rate (FPR)
False Negative Rate (FNR)
Median Latency (ms)
Updated 27d ago
Evaluation Results
Method
Method
Links
Verdict Accuracy
False Positive Rate (FPR)
False Negative Rate (FNR)
Median Latency (ms)
AgentTrust v0.5
Evaluation Protocol=ru...
2026.05
95
2.3
5.4
1.72
AgentTrust v0.5 + LLM-Judge
Evaluation Protocol=hy...
2026.05
88
9
0.8
1,461
DeepSeek-V3
Evaluation Protocol=ze...
2026.05
82.3
7.5
2.3
1,345
Trivial regex blocklist
Pattern Count=50 patterns
2026.05
49.3
0
88.4
0.05
NeMo Guardrails
Backend LLM=DeepSeek-V3
2026.05
44.7
96.2
0
3,558
Feedback
Search any
task
Search any
task