| Benchmark | Model | Rank | Results |
|---|---|---|---|
| common-sense-reasoning-on-record | AlexaTM 20B | #19 | F1: 88.4 |
| coreference-resolution-on-winograd-schema | AlexaTM 20B | #33 | Accuracy: 68.3 |
| natural-language-inference-on-commitmentbank | AlexaTM 20B | #13 | Accuracy: 67.9 |
| natural-language-inference-on-rte | AlexaTM 20B | #53 | Accuracy: 68.6% |
| question-answering-on-boolq | AlexaTM 20B | #41 | Accuracy: 69.4 |
| question-answering-on-copa | AlexaTM 20B | #36 | Accuracy: 78.0 |
| question-answering-on-multirc | AlexaTM 20B | #19 | F1: 59.6 |
| word-sense-disambiguation-on-words-in-context | AlexaTM 20B | #14 | Accuracy: 53.3 |