Sentiment Steering LLaMA2-13B-Chat (test)
79.7BadNets
No Defense
Updated 1mo ago
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
2026.03 | 79.7 | 94.76 | 3.05 | 6.5 | 77.85 | 52.37 | |
2026.03 | 79.63 | 93.27 | 4.32 | 5.2 | 78.52 | 52.18 | |
2026.03 | 77.69 | 81.32 | 1.01 | 6.32 | 48.31 | 42.93 | |
2026.03 | 66.89 | 87.45 | 2.05 | 7.23 | 56.94 | 44.11 | |
2026.03 | 63.75 | 93.22 | 3.05 | 5.66 | 77.23 | 48.58 | |
2026.03 | 23.91 | 29.94 | 0.53 | 9.05 | 58.93 | 24.47 | |
2026.03 | 23.04 | 53.64 | 0 | 2.32 | 48.28 | 25.45 | |
2026.03 | 23.04 | 37.89 | 0 | 0 | 27.23 | 17.63 | |
2026.03 | 4.66 | 6.45 | 0 | 0 | 6.35 | 3.49 | |
2026.03 | 3.11 | 7.69 | 0 | 0 | 5.18 | 3.2 | |
2026.03 | 2.72 | 39.32 | 0 | 1.01 | 46.33 | 17.87 |