Safety Evaluation on English Safety Evaluation
57.5Offensive Attack Success RateAceGPT-7B-chat
Evaluation results
| Method | Links | ||
|---|---|---|---|
| AceGPT-7B-chatModel=AceGPT-7B-chat2026.06 | 57.5 | 50.7 | |
| AceGPT-13B-chatModel=AceGPT-13B-chat2026.06 | 54.8 | 39.7 | |
| jais-30b-chat-v1Model=jais-30b-chat-v12026.06 | 53.4 | 17.8 | |
| jais-13b-chatModel=jais-13b-chat2026.06 | 42.5 | 17.8 | |
| Llama-3-70B-InstructModel=Llama-3-70B-Instruct2026.06 | 26 | 8.2 | |
| Llama-3-8B-InstructModel=Llama-3-8B-Instruct2026.06 | 21.9 | 0 |