| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Agent Security Bench (ASB) IPI Combined Attack (CA) and Context Ignoring (CI) averaged | ToolSafe | ASR2.89 | 18 | 1mo ago | |
| Agent Security Bench (ASB) DPI Combined Attack (CA) and Context Ignoring (CI) averaged | ReAct | ASR92.42 | 18 | 1mo ago | |
| ASB (Agent Security Benchmark) | gpt-4o + GBT-SE | ASR-d (ASB)7 | 8 | 4mo ago | |
| Agent Security Bench (ASB) | PNA84.31 | 4 | 3mo ago | ||
| Agent Security Benchmarks | - | - | 0 | 3mo ago |