Toxicity Assessment on Red Teaming Evaluation B=1000
0.0054Kruskal-Wallis p-valueToxSearch vs ToxSearch-S
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ToxSearch vs ToxSearch-SOutcome=AUC@B, B (Budget)=1000, n (runs per method)=72026.06 | 0.0054 | 0.0262 | 0.035 | 0.714 | |
| ToxSearch vs RainbowPlusOutcome=AUC@B, B (Budget)=1000, n (runs per method)=72026.06 | 0.0054 | 0.0175 | 0.035 | 0.755 | |
| ToxSearch-S vs RainbowPlusOutcome=AUC@B, B (Budget)=1000, n (runs per method)=72026.06 | 0.0054 | 0.0111 | 0.0332 | 0.796 | |
| ToxSearch vs ToxSearch-SOutcome=Best@B, B (Budget)=1000, n (runs per method)=72026.06 | 0.0586 | 0.096 | 0.192 | 0.551 | |
| ToxSearch vs RainbowPlusOutcome=Best@B, B (Budget)=1000, n (runs per method)=72026.06 | 0.0586 | 0.2 | 0.2 | 0.429 | |
| ToxSearch-S vs RainbowPlusOutcome=Best@B, B (Budget)=1000, n (runs per method)=72026.06 | 0.0586 | 0.055 | 0.165 | 0.633 |