Science Question Answering on ARC-C (Accuracy, TFLOPs)
93.6Accuracy (ARC-C)Ours-context-unaware
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Ours-context-unawarePair=8B → 4B, Context Setting=context-unaware2026.06 | 93.6 | 6.56 | |
| Ours-context-unawarePair=4B → 8B, Context Setting=context-unaware2026.06 | 91.38 | 9.42 | |
| Ours-context-unawarePair=14B → 8B, Context Setting=context-unaware2026.06 | 88.99 | 13.28 | |
| Ours-context-unawarePair=8B → 14B, Context Setting=context-unaware2026.06 | 88.48 | 15.85 | |
| Ours-context-unawarePair=14B → 4B, Context Setting=context-unaware2026.06 | 88.48 | 8.9 | |
| Ours-context-unawarePair=4B → 14B, Context Setting=context-unaware2026.06 | 86.86 | 17.15 | |
| T2T-context-unawarePair=8B → 4B, Context Setting=context-unaware2026.06 | 23.74 | 32.75 | |
| T2T-context-unawarePair=4B → 14B, Context Setting=context-unaware2026.06 | 23.39 | 60.83 | |
| T2T-context-unawarePair=8B → 14B, Context Setting=context-unaware2026.06 | 22.96 | 72.06 | |
| T2T-context-unawarePair=14B → 8B, Context Setting=context-unaware2026.06 | 22.87 | 53.8 | |
| C2C-context-unawarePair=4B → 8B, Context Setting=context-unaware2026.06 | 22.52 | 14.94 | |
| T2T-context-unawarePair=14B → 4B, Context Setting=context-unaware2026.06 | 20.61 | 41.95 | |
| T2T-context-unawarePair=4B → 8B, Context Setting=context-unaware2026.06 | 19.48 | 37.86 | |
| C2C-context-unawarePair=4B → 14B, Context Setting=context-unaware2026.06 | 10.7 | 20.05 | |
| C2C-context-unawarePair=8B → 4B, Context Setting=context-unaware2026.06 | 10.35 | 14.76 | |
| C2C-context-unawarePair=8B → 14B, Context Setting=context-unaware2026.06 | 7.39 | 9.84 | |
| C2C-context-unawarePair=14B → 4B, Context Setting=context-unaware2026.06 | 4.96 | 17.93 | |
| C2C-context-unawarePair=14B → 8B, Context Setting=context-unaware2026.06 | 2.78 | 24.77 |