Summarization on XL-Sum Chinese (test)
33.62BERTScore-F1GRASP LoRA Merge + pruning
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GRASP LoRA Merge + pruningAdaptation Method=GRASP LoRA Merge, Training Languages=Chinese+English, Pruning Strategy=Learned prune ratio2026.01 | 33.62 | 56.94 | 16.54 | 30.95 | |
| Adaptive LoRA MergeAdaptation Method=Adaptive LoRA Merge, Training Languages=Chinese+English, Pruning Strategy=Without Pruning2026.01 | 32.01 | — | 14.33 | 29.14 | |
| Adaptive LoRA Merge + pruningAdaptation Method=Adaptive LoRA Merge, Training Languages=Chinese+English, Pruning Strategy=Best grid search over [10% – 80%]2026.01 | 32 | 50 | 14.81 | 29.5 | |
| LoRAAdaptation Method=LoRA, Training Languages=Chinese, Pruning Strategy=Without Pruning2026.01 | 31.09 | — | 13.73 | 28.15 | |
| Zero shotAdaptation Method=Zero-shot, Training Languages=Chinese, Pruning Strategy=Without Pruning2026.01 | 26.48 | — | 9.39 | 20.87 | |
| LoRAHubAdaptation Method=LoRAHub, Training Languages=Chinese+English, Pruning Strategy=Without Pruning2026.01 | 26.42 | — | 10.35 | 24.16 | |
| TIES mergeAdaptation Method=TIES merge, Training Languages=Chinese+English, Pruning Strategy=Without Pruning2026.01 | 26.23 | — | 9.37 | 21.3 |