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

Safety Dialogue Evaluation on SafeDialBench

8.42Score

Full-Context

6.0286.6497.277.891Jan 29, 2026
Updated 3mo ago

Evaluation Results

MethodLinks
2026.01
8.42
2026.01
8.33
2026.01
8.31
2026.01
8.3
2026.01
8.29
2026.01
8.28
2026.01
8.27
2026.01
8.18
2026.01
8.11
2026.01
8.1
2026.01
8.1
2026.01
8.07
2026.01
7.99
2026.01
7.9
2026.01
7.88
2026.01
7.85
2026.01
7.8
2026.01
7.66
2026.01
7.62
2026.01
7.61
2026.01
7.6
2026.01
7.52
2026.01
6.66
2026.01
6.61
2026.01
6.56
2026.01
6.55
2026.01
6.52
2026.01
6.52
2026.01
6.51
2026.01
6.44
2026.01
6.44
2026.01
6.39
2026.01
6.12