STEM Knowledge and Reasoning on GPQA Diamond MMLU Reduced
58.86GPQA Diamond AccuracyTeacher
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TeacherDistillation Setting=Multi-Domain Distillation, Teacher Model=Qwen3-Nemotron-4B2026.05 | 58.86 | 77.03 | |
| TrOPDDistillation Setting=Multi-Domain Distillation, Student Model=Qwen3-SFT-1.7B, Teacher Model=Qwen3-Nemotron-4B2026.05 | 35.98 | 68.74 | |
| EOPDDistillation Setting=Multi-Domain Distillation, Student Model=Qwen3-SFT-1.7B, Teacher Model=Qwen3-Nemotron-4B2026.05 | 33.84 | 68.26 | |
| REOPOLDDistillation Setting=Multi-Domain Distillation, Student Model=Qwen3-SFT-1.7B, Teacher Model=Qwen3-Nemotron-4B2026.05 | 30.56 | 68.3 | |
| Entropy OPDDistillation Setting=Multi-Domain Distillation, Student Model=Qwen3-SFT-1.7B, Teacher Model=Qwen3-Nemotron-4B2026.05 | 29.92 | 68.51 | |
| OPDDistillation Setting=Multi-Domain Distillation, Student Model=Qwen3-SFT-1.7B, Teacher Model=Qwen3-Nemotron-4B2026.05 | 29.8 | 68.6 | |
| Qwen3-SFT-1.7BDistillation Setting=Baseline (None), Student Model=Qwen3-SFT-1.7B2026.05 | 25.25 | 66.6 |