FaithfulSAE: Towards Capturing Faithful Features with Sparse Autoencoders without External Datasets Dependency

Seonglae Cho, Harryn Oh, Donghyun Lee, Luis Eduardo Rodrigues Vieira, Andrew Bermingham, Ziad El Sayed

Keywords: Sparse Autoencoder, Mechanistic Interpretability, Synthetic Dataset, LLM Feature, Dictionary learning, Unsupervised Learning

TL;DR: FaithfulSAEs train SAEs on self-generated synthetic data, removing external dataset dependency; Reduce “Fake Features,” and increase proving performance.

Abstract: Sparse Autoencoders (SAEs) have emerged as a promising solution for decomposing large language model representations into interpretable features. However, Paulo & Belrose (2025) have highlighted instability across different initialization seeds, and Heap et al. (2025) have pointed out that SAEs may not capture model-internal features. These problems likely stem from training SAEs on external datasets—either collected from the Web or generated by another model—which may contain out-of-distribution (OOD) data beyond the model’s generalisation capabilities. This can result in hallucinated SAE features, which we term ”Fake Features”, that misrepresent the model’s internal activations. To address these issues, we propose FaithfulSAE, a method that trains SAEs on the model’s own synthetic dataset. Using FaithfulSAEs, we demonstrate that training SAEs on less-OOD instruction datasets results in SAEs being more stable across seeds. Notably, FaithfulSAEs outperform SAEs trained on webbased datasets in the SAE probing task and exhibit a lower Fake Feature Ratio in 5 out of 7 models. Overall, our approach eliminates the dependency on external datasets, advancing interpretability by better capturing model-internal features while highlighting the often neglected importance of SAE training datasets.

Submission Number: 69