AI engineering glossary · Transformers LLM architecture
What is SAE (sparse autoencoder)?
Sparse autoencoder (also called SAE): A small neural network trained to reconstruct a language model's internal activations through a wide layer where only a few units are active at once. The active units often correspond to human-interpretable features, making it a main tool for interpretability research.
Sparse autoencoder: an example
An autoencoder trained on layer-20 activations finds a feature that fires on Indian currency amounts such as rupee prices.
Related terms
More Transformers LLM architecture terms
- State space model
- Switch Transformer
- T5
- Temperature
- Test-time compute
- Top-k expert routing
- Top-k sampling
- Top-p sampling
Where you learn it
Sparse autoencoder is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.