Text Summarization on CNN/DM (ROUGE-2)
16.95ROUGE-213B (Multi-Task Tuning)
Evaluation Results
| Method | Links | |
|---|---|---|
| 13B (Multi-Task Tuning)Tuning=Multi-task2024.06 | 16.95 | |
| 7B (Multi-Task Tuning)Context=Multi-task transfer from 7B2024.06 | 16.3 | |
| 7B-Expert (CNN/DM)Tuning=Single-task2024.06 | 15.21 | |
| 1.1B (Multi-Task Tuning)Context=Multi-task transfer from 1.1B2024.06 | 14.26 | |
| 1.1B-Expert (CNN/DM)Tuning=Single-task2024.06 | 14.22 | |
| 7B-Expert (GSM8K)Tuning=Single-task2024.06 | 11.19 | |
| 7B-Expert (TruthfulQA)Tuning=Single-task2024.06 | 11.02 | |
| Proxy TuningContext=Multi-task transfer from 7B2024.06 | 10.52 | |
| 7B-Expert (TriviaQA)Tuning=Single-task2024.06 | 10.3 | |
| Proxy TuningContext=Multi-task transfer from 1.1B2024.06 | 9.38 | |
| 13B (Base)Model Type=Pre-trained2024.06 | 8.94 | |
| 7B (Base)Model Type=Pre-trained2024.06 | 8.81 | |
| 1.1B-Expert (TriviaQA)Tuning=Single-task2024.06 | 8.03 | |
| 1.1B-Expert (GSM8K)Tuning=Single-task2024.06 | 7.75 | |
| 1.1B-Expert (TruthfulQA)Tuning=Single-task2024.06 | 7.23 | |
| 1.1B (Base)Model Type=Pre-trained2024.06 | 7.21 |