NDA non-compliance clause detection on NDA-3
92.2AccuracyPolicyGuard
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PolicyGuardBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 92.2 | 68.4 | |
| Majority ClassBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 89.5 | — | |
| Direct PromptingBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 78.9 | 40 | |
| Chain-of-ThoughtBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 78.4 | 42.5 | |
| Legal ReasoningBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 77.5 | 46 | |
| Legal SyllogismBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 75.3 | 40.2 |