Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Helpfulness Evaluation on Helpfulness (evaluation set)
Loading...
84.05
Win Rate
SafeDPO
35.9188
48.4144
60.91
73.4056
May 26, 2025
Win Rate
Tie Rate
Loss Rate
Updated 1mo ago
Evaluation Results
Method
Method
Links
Win Rate
Tie Rate
Loss Rate
SafeDPO
Judge Model=GPT-4, Eva...
2025.05
84.05
9.42
6.53
SafeRLHF
Judge Model=GPT-4, Eva...
2025.05
77.74
11.19
11.07
DPO-HARMLESS
Judge Model=GPT-4, Eva...
2025.05
65.12
21.08
13.8
DPO-SAFEBETTER
Judge Model=GPT-4, Eva...
2025.05
55.65
31.66
12.69
DPO-HELPFUL
Judge Model=GPT-4, Eva...
2025.05
37.77
39.15
23.09
Feedback
Search any
task
Search any
task