Sparse Autoencoders
Collection
SAEs are tools for understanding the internal representations of neural networks. These can be loaded using https://github.com/EleutherAI/sae
•
9 items
•
Updated
These SAEs were trained on the outputs of each of the MLPs in EleutherAI/pythia-160m. We used 8.2 billion tokens from the Pile training set at a context length of 2049. The number of latents is 32,768.
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no library tag.