Doctrine Consistency Analysis on 336 publications, 12,847 statements dataset
91.2PrecisionGIPFEL-I
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GIPFEL-I2026.01 | 91.2 | 87.6 | 89.4 | 8.8 | 1.2 | 12 | |
| Claude-22026.01 | 76.8 | 71.3 | 73.9 | 23.2 | 5.9 | 16 | |
| GPT-42026.01 | 73.4 | 69.2 | 71.2 | 26.6 | 7.8 | 18 | |
| Def. Llama2026.01 | 68.9 | 64.7 | 66.7 | 31.1 | 11.3 | 22 | |
| Human2026.01 | 68.3 | 82.1 | 74.5 | 31.7 | 4.7 | 240 |