Model Editing on zsRE
99.6EfficacyROME
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ROMEBase Model=GPT-2 Large (774M)2022.02 | 99.6 | 84.7 | 22.5 | |
| FT+LBase Model=GPT-2 Large (774M)2022.02 | 98.3 | 56.8 | 22.4 | |
| FT+LBase Model=GPT-2 Medium (345M)2022.02 | 97.2 | 59.4 | 20.9 | |
| ROMEBase Model=GPT-2 Medium (345M)2022.02 | 96.6 | 79.8 | 21.3 | |
| GPT-2 LargeModel Size=774M, Configuration=Vanilla Model2022.02 | 20.6 | 19.8 | 22.5 | |
| GPT-2 MediumModel Size=345M, Configuration=Vanilla Model2022.02 | 18.8 | 18.1 | 21.3 |