Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Prompt-side safety risk classification on S-Eval

94.3F1 Score

Stage3-DPO

82.96485.90788.8591.793Jul 7, 2026
Updated 17d ago

Evaluation Results

MethodLinks
2026.07
94.3
2026.07
90.9
2026.07
90.7
2026.07
90.2
2026.07
90.1
2026.07
89.8
2026.07
88.9
2026.07
88.7
2026.07
85.1
2026.07
84.9
2026.07
84.2
2026.07
83.4