NDA non-compliance clause detection on NDA-4
94.7Accuracy (NDA-4)Majority Class
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Majority ClassBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 94.7 | — | |
| PolicyGuardBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 94.4 | 55.3 | |
| Chain-of-ThoughtBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 82.4 | 29.5 | |
| Direct PromptingBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 80.4 | 26.8 | |
| Legal ReasoningBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 78.8 | 24.7 | |
| Legal SyllogismBackbone=GPT-4.1, Zero-shot protocol=true2026.06 | 77.7 | 26.9 |