Natural Language Understanding on MultiNLI (Match)
1.02ECEMIR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MIRFramework=base, Backbone=RoBERTa-base2023.06 | 1.02 | 1.22 | 0.34 | 3.73 | |
| MIRFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 1.04 | 1.22 | 0.32 | 3.35 | |
| IRFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 1.23 | 1.54 | 0.45 | 3.66 | |
| IRFramework=base, Backbone=RoBERTa-base2023.06 | 1.3 | 1.71 | 0.3 | 3.7 | |
| TSFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 1.53 | 1.86 | 0.4 | 3.45 | |
| ETSFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 1.57 | 1.76 | 0.44 | 3.61 | |
| TSFramework=base, Backbone=RoBERTa-base2023.06 | 1.68 | 1.96 | 0.49 | 4.19 | |
| ConfFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 1.71 | 2.03 | 0.36 | 3.7 | |
| HBFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 1.8 | 2.1 | 0.67 | 3.67 | |
| ETSFramework=base, Backbone=RoBERTa-base2023.06 | 1.88 | 2.08 | 0.7 | 3.97 | |
| HBFramework=base, Backbone=RoBERTa-base2023.06 | 2.07 | 2.1 | 1.03 | 4.63 | |
| ConfFramework=base, Backbone=RoBERTa-base2023.06 | 2.39 | 2.68 | 0.54 | 4.07 | |
| PTSFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 3.48 | 3.49 | 1.13 | 5.29 | |
| PTSKFramework=PROCAL, Backbone=RoBERTa-base2023.06 | 6.1 | 6.15 | 4.03 | 6.8 | |
| PTSFramework=base, Backbone=RoBERTa-base2023.06 | 9.28 | 9.25 | 5.86 | 9.44 | |
| PTSKFramework=base, Backbone=RoBERTa-base2023.06 | 11.94 | 11.91 | 10.24 | 12.23 |