Harmful Refusal on XSTest vani harm
94.3RTA ScoreLlama3.1-8B-IT + SELF-REDTEAM
Evaluation Results
| Method | Links | |
|---|---|---|
| Llama3.1-8B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.06 | 94.3 | |
| Llama3.1-8B-ITTraining=Instruction-tuned base2025.06 | 92 | |
| Qwen2.5-14B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.06 | 91.5 | |
| Qwen2.5-3B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.06 | 91.3 | |
| Qwen2.5-7B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.06 | 91.2 | |
| Qwen2.5-3B-ITTraining=Instruction-tuned base2025.06 | 90 | |
| Qwen2.5-7B-ITTraining=Instruction-tuned base2025.06 | 89 | |
| Qwen2.5-14B-ITTraining=Instruction-tuned base2025.06 | 89 | |
| Llama3.1-8B-IT-AB + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.06 | 81.4 | |
| Llama3.1-8B-IT-ABTraining=Instruction-tuned ablated base2025.06 | 29 |