Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Sparse Autoencoding on GPT-2 activations
Loading...
98
KL Score
Standard SAE
96.96
97.23
97.5
97.77
Jun 4, 2026
KL Score
CE Score
Explained Variance (frac)
Sparsity (L0)
Updated 1mo ago
Evaluation Results
Method
Method
Links
KL Score
CE Score
Explained Variance (frac)
Sparsity (L0)
Standard SAE
Training Token Budget=...
2026.06
98
98
99.1
60
SASA
Training Token Budget=...
2026.06
97
97
98.8
60
Feedback
Search any
task
Search any
task
