Red-teaming on HarmBench (test)
85.1ASR@1
DIALTREE
Updated 1mo ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2025.10 | 85.1 | 98.6 | 99.5 | |
2025.10 | 44.9 | 69.6 | 78.9 | |
2025.10 | 22.6 | 38.5 | 45.1 |
DIALTREE
| Method | Links | |||
|---|---|---|---|---|
2025.10 | 85.1 | 98.6 | 99.5 | |
2025.10 | 44.9 | 69.6 | 78.9 | |
2025.10 | 22.6 | 38.5 | 45.1 |