NDA non-compliance clause detection on NDA-1
93AccuracyPolicyGuard
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PolicyGuardBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 93 | 85.3 | |
| Direct PromptingBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 76.9 | 62.1 | |
| Majority ClassBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 76.8 | — | |
| Chain-of-ThoughtBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 76 | 61.9 | |
| Legal SyllogismBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 74.7 | 60.7 | |
| Legal ReasoningBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 72.8 | 58.4 |