Action Generation on Human Evaluation 48 instances (test)
58.3Preference RateDo-Undo
Evaluation Results
| Method | Links | |
|---|---|---|
| Do-Undo2025.12 | 58.3 | |
| BAGEL2025.12 | 41.6 | |
| Do-Undo(c)consistency=true2025.12 | 33.3 |
Loading the SOTA2 catalog…
| Method | Links | |
|---|---|---|
| Do-Undo2025.12 | 58.3 | |
| BAGEL2025.12 | 41.6 | |
| Do-Undo(c)consistency=true2025.12 | 33.3 |