# Measuring Sparse Autoencoder Feature Sensitivity

arXiv: https://arxiv.org/abs/2509.23717 · HTML: https://mlsanity.com/paper/2509.23717/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in SAE features and their applications, one of 13 directions in Mechanistic interpretability.
- Builds on 21 of its 29 references that are in the map.
- First posted on arXiv 28 Sep 2025.

## Placed in

- Mechanistic interpretability / [SAE features and their applications](https://mlsanity.com/mechinterp/sae-features-and-their-applications/)

## Builds on (most cited first, 21 of 21)

- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Scaling and evaluating sparse autoencoders](https://mlsanity.com/paper/2406.04093/) (2024)
- [Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models](https://mlsanity.com/paper/2403.19647/) (2024)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [Transcoders Find Interpretable LLM Feature Circuits](https://mlsanity.com/paper/2406.11944/) (2024)
- [A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders](https://mlsanity.com/paper/2409.14507/) (2024)
- [Learning Multi-Level Features with Matryoshka Sparse Autoencoders](https://mlsanity.com/paper/2503.17547/) (2025)
- [SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability](https://mlsanity.com/paper/2503.09532/) (2025)
- [Automatically Interpreting Millions of Features in Large Language Models](https://mlsanity.com/paper/2410.13928/) (2024)
- [Sparse Autoencoders Do Not Find Canonical Units of Analysis](https://mlsanity.com/paper/2502.04878/) (2025)
- [Towards Principled Evaluations of Sparse Autoencoders for Interpretability and Control](https://mlsanity.com/paper/2405.08366/) (2024)
- [RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations](https://mlsanity.com/paper/2402.17700/) (2024)
- [Explaining black box text modules in natural language with language models](https://mlsanity.com/paper/2305.09863/) (2023)
- [A Multimodal Automated Interpretability Agent](https://mlsanity.com/paper/2404.14394/) (2024)
- [Applying sparse autoencoders to unlearn knowledge in language models](https://mlsanity.com/paper/2410.19278/) (2024)
- [Measuring Progress in Dictionary Learning for Language Model Interpretability with Board Game Models](https://mlsanity.com/paper/2408.00113/) (2024)
- [Rigorously Assessing Natural Language Explanations of Neurons](https://mlsanity.com/paper/2309.10312/) (2023)
- [Sparse Autoencoders for Hypothesis Generation](https://mlsanity.com/paper/2502.04382/) (2025)
- [Enhancing Neural Network Interpretability with Feature-Aligned Sparse Autoencoders](https://mlsanity.com/paper/2411.01220/) (2024)

## Built on by (most cited first, 5 of 5)

- [Monitoring Emergent Reward Hacking During Generation via Internal Activations](https://mlsanity.com/paper/2603.04069/) (2026)
- [Persistent Sparse Autoencoders: Learning Feature-Specific Timescales in Language Model Representations](https://mlsanity.com/paper/2607.17117/) (2026)
- [A Unifying Perspective on Language Model Representations: From Filler-Role Structure to Mechanistic Interpretability](https://mlsanity.com/paper/2608.29034/) (2026)
- [Active Budget Can Kill Sensitivity: Diagnosing and Repairing TopK Sparse Autoencoder Reliability](https://mlsanity.com/paper/2609.37857/) (2026)
- [Representation Redistribution in Sparse Autoencoders](https://mlsanity.com/paper/2610.21240/) (2026)

## Abstract (arXiv)

Sparse Autoencoder (SAE) features have become essential tools for mechanistic interpretability research. SAE features are typically characterized by examining their activating examples, which are often "monosemantic" and align with human interpretable concepts. However, these examples don't reveal feature sensitivity: how reliably a feature activates on texts similar to its activating examples. In this work, we develop a scalable method to evaluate feature sensitivity. Our approach avoids the need to generate natural language descriptions for features; instead we use language models to generate text with the same semantic properties as a feature's activating examples. We then test whether the feature activates on these generated texts. We demonstrate that sensitivity measures a new facet of feature quality and find that many interpretable features have poor sensitivity. Human evaluation confirms that when features fail to activate on our generated text, that text genuinely resembles the original activating examples. Lastly, we study feature sensitivity at the SAE level and observe that average feature sensitivity declines with increasing SAE width across 7 SAE variants. Our work establishes feature sensitivity as a new dimension for evaluating both individual features and SAE architectures.