Document Contradiction Detection on ContraDoc 1.0 (test)
63.4EHRRnR
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RnRModel=GPT-4o2026.01 | 63.4 | 68 | 64.2 | 69.7 | 76.3 | 76.3 | 39.9 | 60.1 | 23.7 | 83.1 | 20.3 | 26.6 | 13.3 | 6.4 | 2.92 | |
| RnR+CFModel=GPT-4o2026.01 | 57.1 | — | — | — | — | — | — | — | — | 75.7 | 28.9 | 38.3 | 5.4 | 2.31 | — | |
| DPModel=GPT-4o2026.01 | 53.6 | — | — | — | — | — | — | — | — | 70.3 | 26.7 | 34.9 | 5.1 | 2.15 | — | |
| RnRModel=LLaMA3.2-90B2026.01 | 47.5 | 68.1 | 70.4 | 65.7 | 61.6 | 61.6 | 25.4 | 74.6 | 38.4 | 77.1 | 15.5 | 25.2 | 11.8 | 6.55 | 2.5 | |
| DPModel=LLaMA3.2-90B2026.01 | 41.2 | — | — | — | — | — | — | — | — | 66.9 | 19.7 | 31.9 | 4.7 | 2.41 | — | |
| RnR+CFModel=LLaMA3.2-90B2026.01 | 40.8 | — | — | — | — | — | — | — | — | 70.7 | 20.1 | 34.9 | 4.6 | 2.6 | — | |
| RnRModel=LLaMA3-70B2026.01 | 20.3 | 60.1 | 86.9 | 35.7 | 22.5 | 22.5 | 3.3 | 96.7 | 77.5 | 90.3 | 9.5 | 42.4 | 2.2 | 3.08 | 1.41 | |
| DPModel=LLaMA3-70B2026.01 | 20 | — | — | — | — | — | — | — | — | 89.2 | 9.9 | 44.1 | 1.6 | 2.09 | — | |
| RnR+CFModel=LLaMA3-70B2026.01 | 19.4 | — | — | — | — | — | — | — | — | 90.8 | 9.6 | 45.2 | 1.5 | 2.25 | — |