One-Time Edit on CounterFact
99.88EfficacyMEMIT (aligned)
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MEMIT (aligned)Model=Qwen2.52026.05 | 99.88 | 94.41 | 83.2 | 625.26 | 32.91 | |
| AlphaEdit (aligned)Model=Qwen2.52026.05 | 99.88 | 94.03 | 83.38 | 625.12 | 32.73 | |
| MEMIT (aligned)Model=GPT-J2026.05 | 99.83 | 94.34 | 79.13 | 620.66 | 41.03 | |
| PRUNE (aligned)Model=Qwen2.52026.05 | 99.82 | 94.3 | 83.3 | 625.1 | 32.7 | |
| MEMIT (aligned)Model=LLaMA32026.05 | 99.8 | 88.33 | 87.5 | 632.76 | 32.08 | |
| PRUNE (aligned)Model=GPT-J2026.05 | 99.8 | 94.17 | 79.34 | 621.28 | 40.95 | |
| AlphaEdit (aligned)Model=GPT-J2026.05 | 99.77 | 94.56 | 78.79 | 619.7 | 41.09 | |
| PRUNE (aligned)Model=LLaMA32026.05 | 99.57 | 86.8 | 87.41 | 632.28 | 31.7 | |
| AlphaEdit (aligned)Model=GPT-2 XL2026.05 | 99.57 | 92.72 | 69.4 | 617.77 | 41.35 | |
| PRUNE (aligned)Model=GPT-2 XL2026.05 | 99.23 | 91.96 | 71.09 | 623.39 | 41.23 | |
| MEMIT (aligned)Model=GPT-2 XL2026.05 | 99.22 | 92.19 | 70.97 | 623.37 | 41.26 | |
| RECT (aligned)Model=Qwen2.52026.05 | 99.12 | 86.45 | 84.57 | 625.57 | 31.84 | |
| RECT (aligned)Model=GPT-J2026.05 | 98.27 | 77.26 | 81.31 | 622.26 | 37.64 | |
| AlphaEdit (aligned)Model=LLaMA32026.05 | 95.3 | 77.12 | 87.37 | 632.04 | 29.96 | |
| RECT (aligned)Model=GPT-2 XL2026.05 | 95.12 | 79.13 | 75.04 | 626.95 | 38.96 | |
| RECT (aligned)Model=LLaMA32026.05 | 83.23 | 59.41 | 88.69 | 634.34 | 27.99 | |
| Pre-editedModel=GPT-2 XL2026.05 | 22.23 | 24.34 | 78.53 | 626.64 | 31.88 | |
| Pre-editedModel=GPT-J2026.05 | 16.22 | 18.56 | 83.11 | 621.81 | 29.74 | |
| Pre-editedModel=Qwen2.52026.05 | 13.95 | 16.75 | 86.02 | 625.68 | 25.49 | |
| Pre-editedModel=LLaMA32026.05 | 7.85 | 10.58 | 89.48 | 635.23 | 24.14 |