Utility / Task Performance on Professional Domains Corpus
92.6Utilityspotlighting
Evaluation Results
| Method | Links | |
|---|---|---|
| spotlighting2026.06 | 92.6 | |
| baseline2026.06 | 91.8 | |
| sandwiching2026.06 | 91 | |
| PARSE2026.06 | 86.9 | |
| paraphrasing2026.06 | 82.8 | |
| Llama Guard2026.06 | 64.8 |
| Method | Links | |
|---|---|---|
| spotlighting2026.06 | 92.6 | |
| baseline2026.06 | 91.8 | |
| sandwiching2026.06 | 91 | |
| PARSE2026.06 | 86.9 | |
| paraphrasing2026.06 | 82.8 | |
| Llama Guard2026.06 | 64.8 |