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

Safety and Helpfulness Evaluation on RRB

1HarmR

GRPO SafeSearch

-1.01612.59226.239.808Oct 19, 2025
Updated 26d ago

Evaluation Results

MethodLinks
2025.10
12.7
2025.10
2.22.69
2025.10
2.61.81
2025.10
31.84
2025.10
8.92.42
2025.10
14.32.57
2025.10
232.29
2025.10
29.12.52
2025.10
30.52.49
2025.10
31.72.53
2025.10
32.42.55
2025.10
39.82.6
2025.10
49.52.6
2025.10
50.52.54
2025.10
51.42.58