Text Summarization Long Q, Short A (test)
31.8ROUGE-1WorldCup
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| WorldCupBackbone=GEMMA2-9B-IT, k=12026.02 | 31.8 | 67.1 | 62.7 | |
| BiMarkBackbone=GEMMA2-9B-IT2026.02 | 31.7 | 66.9 | 59.8 | |
| GEMMA2-9B-ITWatermark=None2026.02 | 31.2 | — | — | |
| WorldCupBackbone=GEMMA2-9B-IT, k=22026.02 | 31.2 | 66.8 | 66.4 | |
| MPACBackbone=GEMMA2-9B-IT2026.02 | 30.9 | 67.1 | 61.9 | |
| WorldCupBackbone=LLAMA3.1-8B-IT, k=12026.02 | 26.1 | 91.2 | 86 | |
| LLAMA3.1-8B-ITWatermark=None2026.02 | 24.9 | — | — | |
| MPACBackbone=LLAMA3.1-8B-IT2026.02 | 24.9 | 88 | 86.9 | |
| WorldCupBackbone=LLAMA3.1-8B-IT, k=22026.02 | 24.7 | 96.3 | 91.9 | |
| BiMarkBackbone=LLAMA3.1-8B-IT2026.02 | 24.2 | 96 | 89.8 | |
| WorldCupBackbone=MINISTRAL-8B-IT, k=12026.02 | 23.1 | 96.7 | 94.8 | |
| MINISTRAL-8B-ITWatermark=None2026.02 | 22.1 | — | — | |
| MPACBackbone=MINISTRAL-8B-IT2026.02 | 21.5 | 95.1 | 92.9 | |
| WorldCupBackbone=MINISTRAL-8B-IT, k=22026.02 | 21.3 | 98.3 | 95.5 | |
| BiMarkBackbone=MINISTRAL-8B-IT2026.02 | 20.8 | 97.7 | 94.3 |