| AgentHazard Strongest | BraveGuard-Qwen3-Guard-4B | Accuracy90.87 | | 56 | 1mo ago |
| Alpaca + XSTest full (val-selected) | wopt | AUROC0.991 | | 42 | 3mo ago |
| chat 1m (test) | MLP-Cls. | MCA Accuracy100 | | 21 | 3mo ago |
| Harmbench (test) | ReGA | MCA Accuracy100 | | 21 | 3mo ago |
| Polyguard Social Media | T3 | AUROC96.73 | | 18 | 4mo ago |
| Polyguard HR | T3 | AUROC0.9982 | | 18 | 4mo ago |
| Polyguard Education | T3 | AUROC99.43 | | 18 | 4mo ago |
| Polyguard Cyber | T3 | AUROC0.9886 | | 18 | 4mo ago |
| Polyguard Code | T3 | AUROC0.9959 | | 18 | 4mo ago |
| ATBench-500 | | Accuracy90 | | 14 | 1mo ago |
| SafeDialBench (full) | Nemotron | Recall99 | | 12 | 4mo ago |
| MAS Gemma-2-9B (Current) | AgentLens | Accuracy95.35 | | 10 | 1mo ago |
| MAS Qwen-2.5-7B (Current) | AgentLens | Accuracy97.46 | | 10 | 1mo ago |
| MAS LLaMA-3.1-8B (Current) | AgentLens | Accuracy99.15 | | 10 | 1mo ago |
| MAS Qwen-2.5-7B (Lookahead) | AgentLens | Accuracy96.77 | | 8 | 1mo ago |
| MAS LLaMA-3.1-8B (Lookahead) | AgentLens | Accuracy92.31 | | 8 | 1mo ago |
| XSafety | T3+OCSVM | Safety Score (De)0.9815 | | 8 | 4mo ago |
| RTP LX | DUOGUARD | Safety Score (De)98.76 | | 8 | 4mo ago |
| Alpaca and AdvBench Conversation Level (test) | ReGA | MCA Accuracy100 | | 7 | 3mo ago |
| Alpaca and AdvBench Prompt Level (test) | ReGA | Accuracy (MCA)100 | | 7 | 3mo ago |
| DoNotAnswer (held-out) | Geometry-Lite | AUROC97.4 | | 5 | 2mo ago |
| WildJailbreak (held-out) | Geometry-Lite | AUROC99 | | 5 | 2mo ago |
| XSTest (held-out) | Geometry-Lite | AUROC99.6 | | 5 | 2mo ago |
| JBB-Behaviors (held-out) | MultiLayer-DIM | AUROC93.7 | | 5 | 2mo ago |
| PKU-SafeRLHF (held-out) | MultiLayer-DIM | AUROC90.6 | | 5 | 2mo ago |