Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Sparse Autoencoder Evaluation on SAEBench Pythia-160M (test)

0.502RAVEL Score

Ensembling (Boosting)

0.4966960.4980730.499450.500827May 21, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.05
0.50210.2701
2025.05
0.49770.98290.0163
2025.05
0.4970.97830.2886
2025.05
0.49690.97710.2635