| Benchmark | Model | Rank | Results |
|---|---|---|---|
| common-sense-reasoning-on-winogrande | A0.3B-1.3B LoopMoE (R=4) | – | Accuracy: 64.4 |
| multiple-choice-question-answering-mcqa-on-arc-c | A0.3B-1.3B LoopMoE (R=5) | – | Accuracy: 47.5 |
| multiple-choice-question-answering-mcqa-on-arc-e | A0.3B-1.3B LoopMoE (R=5) | – | Accuracy: 64.3 |
| question-answering-on-natural-questions | A0.3B-1.3B LoopMoE (R=5) | – | EM: 9.9 |
| question-answering-on-piqa | A0.3B-1.3B LoopMoE (R=5) | – | Accuracy: 74.0 |
| sentence-completion-on-hellaswag | A0.3B-1.3B LoopMoE (R=5) | – | Accuracy: 63.1 |