Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Toxicity Steering on Eso-LM (512 sequences)
Loading...
-9.2
Toxicity Score
DDPO
-9.54
-7.245
-4.95
-2.655
Sep 25, 2025
Toxicity Score
Updated 1d ago
Evaluation Results
Method
Method
Links
Toxicity Score
DDPO
FLOPs x10^17=0.00
2025.09
-9.2
DDPO
FLOPs x10^17=0.25
2025.09
-9.2
d2-AnyOrder
FLOPs x10^17=0.00
2025.09
-9.2
DDPO
FLOPs x10^17=0.50
2025.09
-9.1
DDPO
FLOPs x10^17=0.75
2025.09
-8.9
DDPO
FLOPs x10^17=1.00
2025.09
-8.9
DDPO
FLOPs x10^17=1.25
2025.09
-8.6
d2-AnyOrder
FLOPs x10^17=0.25
2025.09
-8.5
d2-AnyOrder
FLOPs x10^17=0.50
2025.09
-7.3
d2-AnyOrder
FLOPs x10^17=0.75
2025.09
-5.5
d2-AnyOrder
FLOPs x10^17=1.00
2025.09
-2.7
d2-AnyOrder
FLOPs x10^17=1.25
2025.09
-0.7
Feedback
Search any
task
Search any
task