Instruction Following on SelfInt (ROUGE-L)
19.42ROUGE-LTeacher (Qwen1.5)
Evaluation Results
| Method | Links | |
|---|---|---|
| Teacher (Qwen1.5)Distillation Setting=Cross-Tokenizer KD2026.03 | 19.42 | |
| Teacher (GPT-2 XL)Distillation Setting=Same-Tokenizer KD2026.03 | 14.65 | |
| DSKD-CMA-GADistillation Setting=Cross-Tokenizer KD, Divergence Function=SKL2026.03 | 11.82 | |
| DSKD-CMADistillation Setting=Cross-Tokenizer KD, Divergence Function=SRKL2026.03 | 11.5 | |
| DSKD-CLADistillation Setting=Cross-Tokenizer KD2026.03 | 11.44 | |
| DSKD-CLPDistillation Setting=Cross-Tokenizer KD2026.03 | 11.35 | |
| DSKD-CMA-GADistillation Setting=Cross-Tokenizer KD, Divergence Function=KL2026.03 | 11.17 | |
| DSKD-CMA-CTDistillation Setting=Cross-Tokenizer KD, Divergence Function=KL2026.03 | 10.83 | |
| DSKDDistillation Setting=Same-Tokenizer KD2026.03 | 10.58 | |
| MinEDDistillation Setting=Cross-Tokenizer KD2026.03 | 9.78 | |
| DSKD-CMADistillation Setting=Cross-Tokenizer KD, Divergence Function=KL2026.03 | 9.73 | |
| SFTDistillation Setting=Cross-Tokenizer KD2026.03 | 9.49 | |
| ULDDistillation Setting=Cross-Tokenizer KD2026.03 | 9.28 | |
| Student (GPT-2)Distillation Setting=Baseline2026.03 | 4.72 |