Content Safety Detection on MOCHA
93.1F1 Scoreprompt–response verification framework
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| prompt–response verification framework2026.06 | 93.1 | 4.9 | |
| Single-Agent+CoTType=Strongest Applicable Baseline2026.06 | 86.4 | — |
Loading the SOTA2 catalog…