Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Agent Security Bench (ASB)

Benchmarks

Task NameDataset NameSOTA ResultTrend
Agent Security EvaluationAgent Security Bench (ASB) IPI Combined Attack (CA) and Context Ignoring (CI) averaged
ASR2.89
18
Security Attack DetectionAgent Security Bench (ASB) Structured Scenarios standardized (test)
ASR (Direct Injection)0.5
5
Agent Security EvaluationAgent Security Bench (ASB)
PNA84.31
4
Showing 3 of 3 rows