Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Sparse Autoencoder Disentanglement and Interpretability on Llama-3-8B (layer 20)
Loading...
0.986
KL Score
Batch TopK
0.985272
0.985461
0.98565
0.985839
Jun 29, 2026
KL Score
Interp
Composition Score
Absorption Score
Split Score
Updated 25d ago
Evaluation Results
Method
Method
Links
KL Score
Interp
Composition Score
Absorption Score
Split Score
Batch TopK
layer=20, k=100
2026.06
0.986
0.9405
0.3334
0.147
1.1923
Ort
layer=20, k=100
2026.06
0.9859
0.9265
0.12
0.0979
1.0385
C^2R
layer=20, k=100
2026.06
0.9858
0.9443
0.0816
0.041
1.0385
Matryoshka
layer=20, k=100
2026.06
0.9853
0.9552
0.1803
0.0349
1.1154
Feedback
Search any
task
Search any
task