Natural Language Inference on AX-b
58.42AccuracyOPUS
Evaluation results
| Method | Links | |
|---|---|---|
| OPUSModel Backbone=GPT-2 XL2026.02 | 58.42 | |
| GREATSModel Backbone=GPT-2 XL2026.02 | 57.34 | |
| FineWeb-EduModel Backbone=GPT-2 XL2026.02 | 55.25 | |
| PPLModel Backbone=GPT-2 XL2026.02 | 54.98 | |
| QuRatingModel Backbone=GPT-2 XL2026.02 | 54.35 | |
| DSIRModel Backbone=GPT-2 XL2026.02 | 53.53 | |
| RandomModel Backbone=GPT-2 XL2026.02 | 52.54 | |
| DCLM-FastTextModel Backbone=GPT-2 XL2026.02 | 52.08 | |
| UltraFinewebModel Backbone=GPT-2 XL2026.02 | 48.73 |