Jailbreak Attack on JBB Behaviors (ASR)
100ASRAMIS
Evaluation Results
| Method | Links | |
|---|---|---|
| AMISTarget LLM=Llama-3.1-8B2025.11 | 100 | |
| PAPTarget LLM=Llama-3.1-8B2025.11 | 99 | |
| AMISTarget LLM=GPT-4o-mini2025.11 | 99 | |
| TAPTarget LLM=Llama-3.1-8B2025.11 | 93 | |
| SeqARTarget LLM=Llama-3.1-8B2025.11 | 93 | |
| AMISTarget LLM=GPT-4o2025.11 | 93 | |
| PAIRTarget LLM=Llama-3.1-8B2025.11 | 92 | |
| AMISTarget LLM=Sonnet-42025.11 | 88 | |
| PAPTarget LLM=GPT-4o-mini2025.11 | 87 | |
| AutoDAN-TurboTarget LLM=Llama-3.1-8B2025.11 | 86 | |
| PAPTarget LLM=GPT-4o2025.11 | 83 | |
| TAPTarget LLM=GPT-4o-mini2025.11 | 81 | |
| PAIRTarget LLM=GPT-4o-mini2025.11 | 78 | |
| AMISTarget LLM=Haiku-3.52025.11 | 78 | |
| PAPTarget LLM=Haiku-3.52025.11 | 74 | |
| PAIRTarget LLM=GPT-4o2025.11 | 70 | |
| TAPTarget LLM=GPT-4o2025.11 | 65 | |
| AutoDAN-TurboTarget LLM=GPT-4o-mini2025.11 | 61 | |
| PAIRTarget LLM=Haiku-3.52025.11 | 52 | |
| AutoDAN-TurboTarget LLM=GPT-4o2025.11 | 51 | |
| VanillaTarget LLM=Llama-3.1-8B2025.11 | 47 | |
| TAPTarget LLM=Haiku-3.52025.11 | 45 | |
| AutoDAN-TurboTarget LLM=Sonnet-42025.11 | 36 | |
| AutoDAN-TurboTarget LLM=Haiku-3.52025.11 | 35 | |
| PAPTarget LLM=Sonnet-42025.11 | 32 | |
| PAIRTarget LLM=Sonnet-42025.11 | 30 | |
| TAPTarget LLM=Sonnet-42025.11 | 23 | |
| SeqARTarget LLM=Sonnet-42025.11 | 14 | |
| VanillaTarget LLM=GPT-4o-mini2025.11 | 11 | |
| VanillaTarget LLM=GPT-4o2025.11 | 7 | |
| VanillaTarget LLM=Sonnet-42025.11 | 5 | |
| VanillaTarget LLM=Haiku-3.52025.11 | 4 | |
| SeqARTarget LLM=Haiku-3.52025.11 | 3 | |
| SeqARTarget LLM=GPT-4o-mini2025.11 | 0 | |
| SeqARTarget LLM=GPT-4o2025.11 | 0 |