| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| OpenClaw Injection Tool | Harm Score (HS)3.66 | 8 | 23d ago | ||
| OpenClaw ZH | Harm Score (HS)3.82 | 4 | 23d ago | ||
| OpenClaw EN | Harm Score (HS)3.76 | 4 | 23d ago | ||
| OpenClaw Average across conversation variants | Harm Score (HS)4.35 | 4 | 23d ago | ||
| OpenClaw Web Content | Kimi k2.5 | Harm Score (HS)4.03 | 4 | 23d ago | |
| OpenClaw Log Replay | Kimi k2.5 | Harm Score (HS)4.34 | 4 | 23d ago | |
| OpenClaw Routine | Harm Score (HS)3.9 | 4 | 23d ago |
