NDA non-compliance clause detection on NDA-5
92.1AccuracyPolicyGuard
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PolicyGuardBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 92.1 | 74.3 | |
| Majority ClassBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 85.3 | — | |
| Direct PromptingBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 69.6 | 36.5 | |
| Chain-of-ThoughtBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 69.1 | 34.7 | |
| Legal SyllogismBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 68.5 | 41 | |
| Legal ReasoningBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 67.1 | 39.1 |