Knowledge Distillation Robustness on Qwen2.5-7B teacher vs. Qwen2.5-7B-Open-R1-Distill student (test)
0.9998Similarity ScorePCS
Evaluation results
| Method | Links | |
|---|---|---|
| PCSDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.9998 | |
| ICSDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.9992 | |
| REEFDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.9926 | |
| AttnDiffDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.9873 | |
| LogitsDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.699 | |
| LLMmapDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.6744 | |
| ProFlingoDeployable distilled model=Qwen2.5-7B-Open-R1-Distill2026.04 | 0.56 |