Multi-Document Summarization on Multi-News 256 (test)
46ROUGE-1PRIMERA (reported)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PRIMERA (reported)Evaluation Setting=100-shot, Maximum Length=2562022.08 | 46 | 16.8 | 22.9 | |
| CentrumEvaluation Setting=100-shot, Maximum Length=2562022.08 | 45.7 | 16.8 | 23.2 | |
| PRIMERA (reproduced)Evaluation Setting=100-shot, Maximum Length=2562022.08 | 45.2 | 16.3 | 22.7 | |
| PRIMERA (reported)Evaluation Setting=10-shot, Maximum Length=2562022.08 | 44 | 15.5 | 22 | |
| CentrumEvaluation Setting=Zero-shot, Maximum Length=2562022.08 | 43.5 | 15.7 | 22.4 | |
| PRIMERA (reproduced)Evaluation Setting=10-shot, Maximum Length=2562022.08 | 43.4 | 15.3 | 21.6 | |
| CentrumEvaluation Setting=10-shot, Maximum Length=2562022.08 | 43.4 | 16.6 | 22.2 | |
| Pegasus (reported)Evaluation Setting=100-shot, Maximum Length=2562022.08 | 43 | 13.5 | 21.1 | |
| PRIMERA (reported)Evaluation Setting=Zero-shot, Maximum Length=2562022.08 | 42 | 13.6 | 20.8 | |
| PRIMERA (reproduced)Evaluation Setting=Zero-shot, Maximum Length=2562022.08 | 41.6 | 13.1 | 19.9 | |
| Pegasus (reported)Evaluation Setting=10-shot, Maximum Length=2562022.08 | 39 | 12.1 | 20.3 | |
| Pegasus (reported)Evaluation Setting=Zero-shot, Maximum Length=2562022.08 | 32 | 10.1 | 16.7 |