| common-sense-reasoning-on-winogrande | RoBERTa-Winogrande 355M (fine-tuned) | #17 | Accuracy: 79.1 |
| common-sense-reasoning-on-winogrande | BERT-Winogrande 345M (fine-tuned) | #41 | Accuracy: 64.9 |
| common-sense-reasoning-on-winogrande | RoBERTa-DPR 355M (0-shot) | #49 | Accuracy: 58.9 |
| common-sense-reasoning-on-winogrande | BERT-large 345M (0-shot) | #58 | Accuracy: 51.9 |
| common-sense-reasoning-on-winogrande | BERT-DPR 345M (0-shot) | #59 | Accuracy: 51 |
| common-sense-reasoning-on-winogrande | RoBERTa-large 355M (0-shot) | #60 | Accuracy: 50 |
| coreference-resolution-on-winograd-schema | RoBERTa-WinoGrande 355M | #7 | Accuracy: 90.1 |
| coreference-resolution-on-winograd-schema | RoBERTa-DPR 355M | #16 | Accuracy: 83.1 |
| coreference-resolution-on-winograd-schema | WKH | #53 | Accuracy: 57.1 |
| coreference-resolution-on-winograd-schema | KEE+NKAM on WinoGrande | #58 | Accuracy: 52.8 |
| question-answering-on-copa | RoBERTa-Winogrande-ft 355M (fine-tuned) | #15 | Accuracy: 90.6 |
| question-answering-on-copa | RoBERTa-ft 355M (fine-tuned) | #22 | Accuracy: 86.4 |
| question-answering-on-copa | RoBERTa-Winogrande 355M (fine-tuned) | #27 | Accuracy: 84.4 |
| question-answering-on-copa | Causal Strength w/multi-word predicates (presumably on WinoGrande?) | #39 | Accuracy: 76.4 |
| question-answering-on-copa | Pointwise Mutual Information (on 10M stories) | #44 | Accuracy: 65.4 |