List your company

Safety Evaluation on PAP

93.65Safety ScoreREFLECTOR (+GDPO)
34.047649.521364.99580.4687May 20, 2026

Evaluation results

MethodLinks
REFLECTOR (+GDPO)Backbone=Llama-3.1-8B-Instruct2026.0593.65
REFLECTOR (+SFT)Backbone=Llama-3.1-8B-Instruct2026.0592.69
REFLECTOR (+GDPO)Backbone=Qwen-2.5-7B-Instruct2026.0591.34
REFLECTOR (SFT)Backbone=Qwen-2.5-7B-Instruct2026.0586.96
STAIRBackbone=Llama-3.1-8B-Instruct2026.0585.35
STAIRBackbone=Qwen-2.5-7B-Instruct2026.0582.53
Shallow-AlignBackbone=Llama-3.1-8B-Instruct2026.0578.2
Shallow-AlignBackbone=Qwen-2.5-7B-Instruct2026.0576.1
Self-CritiqueBackbone=Qwen-2.5-7B-Instruct2026.0548.07
Self-CritiqueBackbone=Llama-3.1-8B-Instruct2026.0545.6
DPOBackbone=Qwen-2.5-7B-Instruct2026.0544.03
DPOBackbone=Llama-3.1-8B-Instruct2026.0541.7
SFTBackbone=Llama-3.1-8B-Instruct2026.0540.1
SFTBackbone=Qwen-2.5-7B-Instruct2026.0538.46
OriginalBackbone=Llama-3.1-8B-Instruct2026.0538.28
OriginalBackbone=Qwen-2.5-7B-Instruct2026.0536.34