Commonsense Reasoning on OpenQA (eval)
75.4AccuracyT5-11B
Evaluation results
| Method | Links | |
|---|---|---|
| T5-11BModel Size=11B, Backbone=T52021.01 | 75.4 | |
| RoBERTa-L + MUPPETModel Size=Large, Backbone=RoBERTa, Training Strategy=MUPPET2021.01 | 74.4 | |
| RoBERTa-LModel Size=Large, Backbone=RoBERTa2021.01 | 73.6 | |
| BARTBackbone=BART2021.01 | 71.4 | |
| BART + MUPPETBackbone=BART, Training Strategy=MUPPET2021.01 | 70.8 | |
| T5-LModel Size=Large, Backbone=T52021.01 | 70.4 | |
| RoBERTa-B + MUPPETModel Size=Base, Backbone=RoBERTa, Training Strategy=MUPPET2021.01 | 64.6 | |
| RoBERTa-BModel Size=Base, Backbone=RoBERTa2021.01 | 63.8 |