Summarization on CNN/DailyMail (RougeL, BLEURT)
23.23RougeLEM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EMBackbone=Llama-8B2025.09 | 23.23 | 44.78 | |
| SVSIBackbone=Llama-8B2025.09 | 20.72 | 45.53 | |
| EMBackbone=Llama-3B2025.09 | 20.69 | 38.97 | |
| SVSIBackbone=Qwen-7B2025.09 | 20.69 | 42.61 | |
| baseBackbone=Llama-8B2025.09 | 20.68 | 45.48 | |
| SVSIBackbone=Llama-1B2025.09 | 17.83 | 40.28 | |
| EMBackbone=Qwen-7B2025.09 | 17.65 | 33.75 | |
| baseBackbone=Llama-1B2025.09 | 16.86 | 37.93 | |
| SJBackbone=Llama-1B2025.09 | 16.53 | 38.02 | |
| baseBackbone=Qwen-7B2025.09 | 16.44 | 33.9 | |
| EMBackbone=Llama-1B2025.09 | 14.45 | 25.66 | |
| SVSIBackbone=Qwen-1.5B2025.09 | 13.72 | 30.63 | |
| SJBackbone=Qwen-1.5B2025.09 | 13.54 | 27.46 | |
| baseBackbone=Qwen-1.5B2025.09 | 13.5 | 29.62 | |
| SVSIBackbone=Llama-3B2025.09 | 9.18 | 20.36 | |
| SJBackbone=Qwen-7B2025.09 | 7.49 | 16.33 | |
| baseBackbone=Llama-3B2025.09 | 7.23 | 16.25 | |
| EMBackbone=Qwen-1.5B2025.09 | 1.63 | 4.41 | |
| SJBackbone=Llama-3B2025.09 | 0.18 | 1.92 | |
| SJBackbone=Llama-8B2025.09 | 0.08 | 1.73 | |
| baseBackbone=Qwen-3B2025.09 | 0 | 0 |