List your company

Safety Evaluation on HEx-PHI (Harm & ASR)

1.26Harm ScoreBase
1.11242.10873.1054.1013Oct 17, 2025

Evaluation results

MethodLinks
BaseModel=Llama-3-8B-Instruct2025.101.263
BaseModel=Llama-3.1-8B-Instruct2025.101.395.67
BaseModel=Qwen2.5-7B-Instruct2025.101.585
BaseModel=Qwen3-8B2025.101.8510
BaseModel=DeepSeek-R1-Distill-Llama-8B2025.103.0933.67
SFTModel=Qwen2.5-7B-Instruct2025.104.6274.67
SFTModel=Qwen3-8B2025.104.7183
SFTModel=Llama-3-8B-Instruct2025.104.7579.33
SFTModel=DeepSeek-R1-Distill-Llama-8B2025.104.7780.67
SFTModel=Llama-3.1-8B-Instruct2025.104.882
HarmRLVRModel=Qwen3-8B2025.104.8389.67
HarmRLVRModel=Qwen2.5-7B-Instruct2025.104.8992.33
HarmRLVRModel=DeepSeek-R1-Distill-Llama-8B2025.104.9293.33
HarmRLVRModel=Llama-3-8B-Instruct2025.104.9394.67
HarmRLVRModel=Llama-3.1-8B-Instruct2025.104.9596