Prompt Injection Defense on Mind2Web
84Benign Utility
Vanilla
Updated 4d ago
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
2025.11 | 84 | - | 61.2 | 46.2 | 59.8 | 48.8 | 54.3 | 63.5 | 0 | 100 | |
2025.11 | 84 | - | 63.1 | 44.7 | 61.2 | 46.5 | 55.8 | 58.9 | 31.7 | 71.5 | |
2025.11 | 84 | 0 | 70.4 | 25.1 | 68.9 | 29.6 | 64.1 | 28.7 | 35.4 | 67.2 | |
2025.11 | 81 | 0 | 75.6 | 5.2 | 74.9 | 4.7 | 73.3 | 5.6 | 69.5 | 8.5 | |
2025.11 | 74.5 | 0 | 69.2 | 6.1 | 68.1 | 5.8 | 66.7 | 6.4 | 64.9 | 10.9 | |
2025.11 | 74 | - | 53.4 | 48.3 | 52.8 | 49.7 | 48.9 | 62.1 | 31.2 | 72.6 | |
2025.11 | 74 | - | 55.8 | 46.2 | 54.1 | 47.9 | 50.3 | 57.4 | 32.8 | 68.9 | |
2025.11 | 74 | 0 | 61.3 | 26.6 | 59.7 | 31.1 | 56.2 | 29.6 | 33.5 | 66.1 |