Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-attribute steering on Multi-attribute steering S+T+St (test)
Loading...
95.6
S Score
Adaptive
84.576
87.438
90.3
93.162
May 8, 2026
S Score
T Score
St Score
Perplexity (PPL)
d2 Score
Updated 21d ago
Evaluation Results
Method
Method
Links
S Score
T Score
St Score
Perplexity (PPL)
d2 Score
Adaptive
Backbone=DREAM-7B, Alp...
2026.05
95.6
98.3
73.7
74
77.6
Adaptive
Backbone=SEDD-124M, Al...
2026.05
95
100
82.5
63
83.7
Adaptive
Backbone=MDLM-124M, Al...
2026.05
90.1
95.8
58.1
84
92.8
Adaptive
Backbone=LLaDA-8B, Alp...
2026.05
85
86.8
45.2
7
46.1
Feedback
Search any
task
Search any
task
