Plan Generation on RecipeNLG (test)
91.3BLEUSelf-Ask
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Self-Askreveal_range=k in [1, 4]2026.01 | 91.3 | 15.7 | |
| ReActreveal_range=k in [1, 4]2026.01 | 91.2 | 59.9 | |
| SQ-BCPreveal_range=k in [1, 4]2026.01 | 90.7 | 5.8 | |
| CoTreveal_range=k in [1, 4]2026.01 | 90 | 64.1 | |
| ToSreveal_range=k in [1, 4]2026.01 | 89.8 | 64.2 | |
| Direct Promptreveal_range=k in [1, 4]2026.01 | 89.7 | 65.7 | |
| ToTreveal_range=k in [1, 4]2026.01 | 89.2 | 66.5 |