Exact recovery of injected triggers on Llama-3.1 70B
2,462Trigger Count
RuleFit+XGB
Updated 1mo ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2025.05 | 2,462 | 14 | 14 | 14 | |
2025.05 | 1,761 | 14 | 14 | 14 | |
2025.05 | 1,715 | 57 | 60 | 60 | |
2025.05 | 1,422 | 21 | 25 | 26 | |
2025.05 | 1,360 | 57 | 60 | 60 | |
2025.05 | 648 | 0 | 0 | 0 |