Jailbreak Defense on Emoji Attack ActorAttack
0Attack Success Rate (ASR)SelfGrader
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SelfGraderTarget LLM=LLama-3-8B Instruct2026.04 | 0 | 0 | |
| SelfDefendvariant=Direct, Target LLM=LLama-3-8B Instruct2026.04 | 2.66 | 76.16 | |
| Llama Guardvariant=Pre, Target LLM=LLama-3-8B Instruct2026.04 | 4 | 99.5 |