In-context Learning on 12 Downstream Classification Tasks
53AccuracyStateX
Evaluation Results
| Method | Links | |
|---|---|---|
| StateXArchitecture=State Space Models—Mamba2, Few-shot=24-shot2025.09 | 53 | |
| StateXArchitecture=Linear Attention—GLA, Few-shot=16-shot2025.09 | 52.4 | |
| VanillaArchitecture=State Space Models—Mamba2, Few-shot=24-shot2025.09 | 52.4 | |
| StateXArchitecture=State Space Models—Mamba2, Few-shot=16-shot2025.09 | 52.3 | |
| StateXArchitecture=Linear Attention—GLA, Few-shot=24-shot2025.09 | 51.9 | |
| VanillaArchitecture=Linear Attention—GLA, Few-shot=16-shot2025.09 | 49.7 | |
| VanillaArchitecture=State Space Models—Mamba2, Few-shot=16-shot2025.09 | 49.7 | |
| VanillaArchitecture=Linear Attention—GLA, Few-shot=24-shot2025.09 | 48.4 | |
| StateXArchitecture=Linear Attention—GLA, Few-shot=8-shot2025.09 | 48.1 | |
| VanillaArchitecture=State Space Models—Mamba2, Few-shot=8-shot2025.09 | 47.7 | |
| StateXArchitecture=State Space Models—Mamba2, Few-shot=8-shot2025.09 | 47.6 | |
| VanillaArchitecture=Linear Attention—GLA, Few-shot=8-shot2025.09 | 47.3 | |
| MoMArchitecture=Sparse Model—MoM, Few-shot=24-shot2025.09 | 42.9 | |
| MoMArchitecture=Sparse Model—MoM, Few-shot=8-shot2025.09 | 42.6 | |
| MoMArchitecture=Sparse Model—MoM, Few-shot=16-shot2025.09 | 42.2 |