Jailbreak Attack on Gemini Flash 2.5 (test)
0ASRGPTFUZZER
Evaluation Results
| Method | Links | |
|---|---|---|
| GPTFUZZERDefense Strategy=Self-Reminder2026.02 | 0 | |
| ICRTDefense Strategy=ICD (1-shot) + Llama Guard (Output)2026.02 | 0 | |
| ICRTDefense Strategy=ICD (2-shot) + Llama Guard (Output)2026.02 | 0 | |
| ICRTDefense Strategy=Self-Reminder + Llama Guard (Output)2026.02 | 0 | |
| GPTFUZZERDefense Strategy=Self-Reminder + Llama Guard (Output)2026.02 | 0 | |
| ICRTDefense Strategy=ICD (2-shot) + Self-Reminder + LG (Output)2026.02 | 0 | |
| GPTFUZZERDefense Strategy=ICD (2-shot) + Self-Reminder + LG (Output)2026.02 | 2 | |
| ICRTDefense Strategy=ICD (2-shot) + Self-Reminder2026.02 | 6 | |
| GPTFUZZERDefense Strategy=ICD (2-shot) + Self-Reminder2026.02 | 6 | |
| ICRTDefense Strategy=Self-Reminder2026.02 | 8 | |
| CC-BOSDefense Strategy=ICD (2-shot) + Llama Guard (Output)2026.02 | 16 | |
| GPTFUZZERDefense Strategy=ICD (2-shot) + Llama Guard (Output)2026.02 | 16 | |
| CC-BOSDefense Strategy=ICD (2-shot) + Self-Reminder + LG (Output)2026.02 | 16 | |
| CC-BOSDefense Strategy=Self-Reminder + Llama Guard (Output)2026.02 | 18 | |
| ICRTDefense Strategy=ICD (2-shot)2026.02 | 20 | |
| GPTFUZZERDefense Strategy=ICD (1-shot) + Llama Guard (Output)2026.02 | 20 | |
| GPTFUZZERDefense Strategy=ICD (2-shot)2026.02 | 22 | |
| CC-BOSDefense Strategy=ICD (1-shot) + Llama Guard (Output)2026.02 | 22 | |
| GPTFUZZERDefense Strategy=ICD (1-shot)2026.02 | 24 | |
| CC-BOSDefense Strategy=ICD (2-shot) + Self-Reminder2026.02 | 24 | |
| GPTFUZZERDefense Strategy=No Defense2026.02 | 28 | |
| CC-BOSDefense Strategy=Self-Reminder2026.02 | 28 | |
| ICRTDefense Strategy=ICD (1-shot)2026.02 | 46 | |
| CC-BOSDefense Strategy=ICD (2-shot)2026.02 | 54 | |
| CC-BOSDefense Strategy=ICD (1-shot)2026.02 | 76 | |
| ICRTDefense Strategy=No Defense2026.02 | 92 | |
| CC-BOSDefense Strategy=No Defense2026.02 | 100 |