Summarization on CNN/DM (test) using ROUGE, BERTScore, and QAFactEval
44.05ROUGE-1GloSA-sum
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GloSA-sumEncoder=GloSA-sum2026.02 | 44.05 | 21.22 | 41.06 | 0.88 | 0.78 | |
| FT-LLaMA-3 8BEncoder=FT-LLaMA-3 8B2026.02 | 39.14 | 17.94 | 36.1 | 0.87 | 0.776 | |
| Claude-3 SumEncoder=Claude-3 Sum2026.02 | 38.66 | 16.01 | 35.24 | 0.863 | 0.771 | |
| GPT-4 Prompt-SumEncoder=GPT-4 Prompt-Sum2026.02 | 37.91 | 15.42 | 34.83 | 0.858 | 0.762 |