List your company

Harmful Refusal on OR-Bench vani harm

98.5RTAQwen2.5-14B-IT + SELF-REDTEAM
-2.48423.73349.9576.167Jun 9, 2025

Evaluation results

MethodLinks
Qwen2.5-14B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0698.5
Qwen2.5-3B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0697.2
Qwen2.5-7B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0696.8
Llama3.1-8B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0694.2
Qwen2.5-3B-ITTraining=Instruction-tuned base2025.0693
Qwen2.5-14B-ITTraining=Instruction-tuned base2025.0689.3
Qwen2.5-7B-ITTraining=Instruction-tuned base2025.0687.9
Llama3.1-8B-ITTraining=Instruction-tuned base2025.0686.4
Llama3.1-8B-IT-AB + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0684.6
Llama3.1-8B-IT-ABTraining=Instruction-tuned ablated base2025.061.4