Thinking Collapse Analysis on AdvBench
0Thinking Collapse RateAutoDAN
Evaluation Results
| Method | Links | |
|---|---|---|
| AutoDANModel=Qwen3 1.7B2026.03 | 0 | |
| AutoDAN-TurboModel=Qwen3 1.7B2026.03 | 0 | |
| PAIRModel=Qwen3 4B2026.03 | 0 | |
| AutoDANModel=Qwen3 4B2026.03 | 0 | |
| AutoDAN-TurboModel=Qwen3 4B2026.03 | 0 | |
| PAIRModel=Qwen3 8B2026.03 | 0 | |
| AutoDANModel=Qwen3 8B2026.03 | 0 | |
| AutoDAN-TurboModel=Qwen3 8B2026.03 | 0 | |
| PAIRModel=Qwen3 1.7B2026.03 | 0.5 | |
| GCGModel=Qwen3 1.7B2026.03 | 0.8 | |
| GCGModel=Qwen3 8B2026.03 | 0.8 | |
| GCGModel=Qwen3 4B2026.03 | 1.6 | |
| MS_ReverseModel=Qwen3 1.7B2026.03 | 10 | |
| MS_ReverseModel=Qwen3 4B2026.03 | 13 | |
| MS_ReverseModel=Qwen3 8B2026.03 | 13.5 |