Reasoning on Instruction following
100Normalized ScoreInstr. following expert
Evaluation results
| Method | Links | |
|---|---|---|
| Instr. following expert2026.03 | 100 | |
| DUME2026.03 | 87.2 | |
| Oracle2026.03 | 85.3 | |
| DUME OOD2026.03 | 83.1 | |
| Model averaging2026.03 | 59 | |
| Random routing2026.03 | 57 | |
| Coding expert2026.03 | 51.1 | |
| Math expert2026.03 | 46.5 | |
| Multilingual expert2026.03 | 21.1 |