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

Safety Evaluation on SORRY-Bench (expected HFR)

0.9Expected HFR

G

-0.72410.23821.232.162Oct 10, 2025
Updated 23h ago

Evaluation Results

MethodLinks
2025.10
0.9
2025.10
0.9
2025.10
5.1
2025.10
11.3
2025.10
14.2
2025.10
16.7
2025.10
20.7
2025.10
22
2025.10
22.3
2025.10
24.5
2025.10
26.3
2025.10
27.2
2025.10
27.3
2025.10
27.6
2025.10
29.5
2025.10
30.1
2025.10
30.4
2025.10
30.7
2025.10
30.8
2025.10
30.9
2025.10
31
2025.10
31.2
2025.10
31.2
2025.10
31.6
2025.10
31.6
2025.10
32.3
2025.10
32.8
2025.10
32.9
2025.10
32.9
2025.10
33.2
2025.10
33.5
2025.10
33.5
2025.10
33.5
2025.10
35.6
2025.10
41.5