LLM Jailbreaking on GPTFuzzer Scenario G2
77HypervolumeEvoJail
Evaluation results
| Method | Links | |
|---|---|---|
| EvoJailTarget Model=Llama-3.1-8B-Instruct2026.03 | 77 | |
| EvoJailTarget Model=gpt-4.1-Nano2026.03 | 75.5 | |
| CodeAttackTarget Model=gpt-4.1-Nano2026.03 | 71.3 | |
| EvoJailTarget Model=Llama-2-7b-chat-hf2026.03 | 67 | |
| CodeAttackTarget Model=Llama-3.1-8B-Instruct2026.03 | 59.5 | |
| ReNeLLMTarget Model=Llama-3.1-8B-Instruct2026.03 | 52.5 | |
| CodeChameleonTarget Model=gpt-4.1-Nano2026.03 | 52 | |
| CodeChameleonTarget Model=Llama-3.1-8B-Instruct2026.03 | 50 | |
| JailbrokenTarget Model=Llama-3.1-8B-Instruct2026.03 | 50 | |
| FlipAttackTarget Model=Llama-3.1-8B-Instruct2026.03 | 37.2 | |
| CodeChameleonTarget Model=Llama-2-7b-chat-hf2026.03 | 35.5 | |
| JailbrokenTarget Model=Llama-2-7b-chat-hf2026.03 | 23.6 | |
| ReNeLLMTarget Model=Llama-2-7b-chat-hf2026.03 | 16.3 | |
| FlipAttackTarget Model=gpt-4.1-Nano2026.03 | 12.8 | |
| ReNeLLMTarget Model=gpt-4.1-Nano2026.03 | 11 | |
| FlipAttackTarget Model=Llama-2-7b-chat-hf2026.03 | 4.7 | |
| CodeAttackTarget Model=Llama-2-7b-chat-hf2026.03 | 4.2 | |
| CipherTarget Model=Llama-3.1-8B-Instruct2026.03 | 2 | |
| JailbrokenTarget Model=gpt-4.1-Nano2026.03 | 1.4 | |
| CipherTarget Model=Llama-2-7b-chat-hf2026.03 | 0 | |
| CipherTarget Model=gpt-4.1-Nano2026.03 | 0 |