List your company

Harmful Refusal on StrongREJECT vani harm

99.5RTA Rate (vani harm)Qwen2.5-14B-IT + SELF-REDTEAM
8.60432.20255.879.398Jun 9, 2025

Evaluation results

MethodLinks
Qwen2.5-14B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0699.5
Qwen2.5-7B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0697.9
Qwen2.5-3B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0697.1
Qwen2.5-14B-ITTraining=Instruction-tuned base2025.0697.1
Llama3.1-8B-ITTraining=Instruction-tuned base2025.0697.1
Llama3.1-8B-IT + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0695.8
Qwen2.5-3B-ITTraining=Instruction-tuned base2025.0692
Qwen2.5-7B-ITTraining=Instruction-tuned base2025.0692
Llama3.1-8B-IT-AB + SELF-REDTEAMTraining=SELF-REDTEAM fine-tuned2025.0691.2
Llama3.1-8B-IT-ABTraining=Instruction-tuned ablated base2025.0612.1