Summarization on CNN/DM (Factuality, Informativeness, Fluency, Overall)
4.87FactualityFENICE-0.75
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| FENICE-0.75Framework=ConSUM, Threshold=0.752026.05 | 4.87 | 4.3 | 4.73 | 4.63 | |
| BaselineSystem Type=Standard Baseline2026.05 | 4.79 | 4.19 | 4.71 | 4.56 | |
| FIZZ-0.75Framework=ConSUM, Threshold=0.752026.05 | 4.77 | 3.99 | 4.67 | 4.48 | |
| MBR-1.0Threshold=1.02026.05 | 4.74 | 4.31 | 4.66 | 4.57 | |
| GoldSystem Type=Human Reference2026.05 | 4.57 | 3.32 | 3.87 | 3.92 |