Logical Reasoning on FOLIO FOL fields (val)
68.1AccuracyCGD-PD
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CGD-PDModel=GPT-5.2, Method=CGD-PD2026.03 | 68.1 | 53.9 | 36.3 | 4.36 | |
| GPT-5.2 (Single)Model=GPT-5.2, Method=Single2026.03 | 63.7 | 57.4 | 41.5 | 1 | |
| CGD-PDModel=Claude Sonnet 4.5, Method=CGD-PD2026.03 | 49 | 58.8 | 53.3 | 4.91 | |
| Claude Sonnet 4.5 (Single)Model=Claude Sonnet 4.5, Method=Single2026.03 | 42.2 | 75.5 | 72.6 | 1 |