Jailbreak Attack on HADES Violence
0.3ASRHADES
Evaluation Results
| Method | Links | |
|---|---|---|
| HADESModel=GPT-4o, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.3 | |
| HADESModel=InternVL 2.5, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.3333 | |
| HADESModel=Qwen2.5-VL, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.4533 | |
| HADESModel=Grok-2-vision, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.5333 | |
| VisCRAModel=GPT-4o, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.6533 | |
| SI-AModel=GPT-4o, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.66 | |
| VisCRAModel=InternVL 2.5, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.6867 | |
| SI-AModel=Qwen2.5-VL, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.6933 | |
| SI-AModel=Grok-2-vision, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.7133 | |
| SI-AModel=InternVL 2.5, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.72 | |
| VisCRAModel=Qwen2.5-VL, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.9067 | |
| GAMBITModel=InternVL 2.5, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.92 | |
| GAMBITModel=GPT-4o, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.92 | |
| GAMBITModel=Qwen2.5-VL, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.94 | |
| GAMBITModel=Grok-2-vision, Evaluation Protocol=Pass@5, Evaluator=Llama-Guard-32026.01 | 0.94 |