Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Harmlessness on HH-RLHF (test)
Loading...
2.772
Reward
MAVIS
2.44648
2.53099
2.6155
2.70001
Aug 19, 2025
Reward
KL Divergence
Updated 4d ago
Evaluation Results
Method
Method
Links
Reward
KL Divergence
MAVIS
Model Size=13B
2025.08
2.772
11.32
PPO
Model Size=13B
2025.08
2.762
11.4
MAVIS
Model Size=7B
2025.08
2.656
4.65
PPO
Model Size=7B
2025.08
2.459
4.23
Feedback
Search any
task
Search any
task