Adam1010/gemma-4-31b-sae-features download history
Adam1010/gemma-4-31b-sae-features is a feature extraction dataset on the Hugging Face Hub. In the last 30 days it was downloaded 360 times (13 in the last 7 days), and 1,144 times in total. It ranks #42,659 among datasets by monthly downloads.
Gemma-4-31B Sparse Autoencoder Features 3,000 interpreted and verified SAE features across all 60 layers of Google's Gemma-4-31B-IT model. What's in this dataset? For each of the 60 transformer layers in Gemma-4-31B, we trained a TopK-64 Sparse Autoencoder with 43,008 features