NDA non-compliance clause detection on Internal NDA 40-policy subset (Average)
96.3Accuracy (Average Subset)PolicyGuard
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PolicyGuardModel=GPT-4.1, Protocol=zero-shot2026.06 | 96.3 | 90.7 | |
| PolicyGuardModel=Sonnet-4.5, Protocol=zero-shot2026.06 | 87.1 | 70.8 | |
| Claude CoworkModel=Opus-4.6, Guidelines=raw policy guidelines, Protocol=zero-shot2026.06 | 85 | 65.4 | |
| Claude CoworkModel=Sonnet-4.6, Guidelines=raw policy guidelines, Protocol=zero-shot2026.06 | 75.6 | 52.6 |