| Benchmark | Model | Rank | Results |
|---|---|---|---|
| social-commonsense-reasoning-on-siqa | Token Rounding (1.8B MoE) | #11 | Accuracy: 42.1 |
| common-sense-reasoning-on-arc-easy | Token Rounding (1.8B MoE) | #74 | Accuracy: 63.0 |
| question-answering-on-boolq | Token Rounding (1.8B MoE) | #93 | Accuracy: 61.4 |
| question-answering-on-piqa | Token Rounding (1.8B MoE) | #147 | Accuracy: 69.6 |