Natural Language Understanding on GLUE (MNLI, QNLI, SST-2)
94.8Accuracy (SST-2)RoBERTa
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| RoBERTaModel Variant=Base2024.10 | 94.8 | 92.8 | 87.6 | |
| Poly-RoBERTaModel Variant=Base2024.10 | 93.35 | 91.62 | 86.93 | |
| Nexus (BERT)Model Variant=Base2024.10 | 92.11 | 89.9 | — |