Enhancing Neural Network Interpretability with Feature-Aligned Sparse Autoencoders
This work addresses interpretability challenges in neural networks for researchers and practitioners, though it is incremental as it builds on existing SAE methods with a novel regularization approach.
The paper tackles the problem of sparse autoencoders (SAEs) learning features not aligned with input features, limiting interpretability, by proposing Mutual Feature Regularization (MFR) to encourage similar feature learning across parallel SAEs, resulting in up to 21.21% improvement in reconstruction loss on GPT-2 Small and 6.67% on EEG data.
Sparse Autoencoders (SAEs) have shown promise in improving the interpretability of neural network activations, but can learn features that are not features of the input, limiting their effectiveness. We propose \textsc{Mutual Feature Regularization} \textbf{(MFR)}, a regularization technique for improving feature learning by encouraging SAEs trained in parallel to learn similar features. We motivate \textsc{MFR} by showing that features learned by multiple SAEs are more likely to correlate with features of the input. By training on synthetic data with known features of the input, we show that \textsc{MFR} can help SAEs learn those features, as we can directly compare the features learned by the SAE with the input features for the synthetic data. We then scale \textsc{MFR} to SAEs that are trained to denoise electroencephalography (EEG) data and SAEs that are trained to reconstruct GPT-2 Small activations. We show that \textsc{MFR} can improve the reconstruction loss of SAEs by up to 21.21\% on GPT-2 Small, and 6.67\% on EEG data. Our results suggest that the similarity between features learned by different SAEs can be leveraged to improve SAE training, thereby enhancing performance and the usefulness of SAEs for model interpretability.