# Interpreting and Steering LLMs with Mutual Information-based Explanations on Sparse Autoencoders

arXiv: https://arxiv.org/abs/2502.15576 · HTML: https://mlsanity.com/paper/2502.15576/

## Key facts

- Cited by 21 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 23 of its 45 references that are in the map.
- First posted on arXiv 21 Feb 2025.

## Placed in

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

## Builds on (most cited first, 23 of 23)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](https://mlsanity.com/paper/1502.01852/) (2015)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Mixed Precision Training](https://mlsanity.com/paper/1710.03740/) (2017)
- [WizardLM: Empowering large pre-trained language models to follow complex instructions](https://mlsanity.com/paper/2304.12244/) (2023)
- [Enhancing Chat Language Models by Scaling High-quality Instructional Conversations](https://mlsanity.com/paper/2305.14233/) (2023)
- [A Primer in BERTology: What we know about how BERT works](https://mlsanity.com/paper/2002.12327/) (2020)
- [SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks](https://mlsanity.com/paper/2310.03684/) (2023)
- [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)
- [Jumping Ahead: Improving Reconstruction Fidelity with JumpReLU Sparse Autoencoders](https://mlsanity.com/paper/2407.14435/) (2024)
- [HelpSteer2: Open-source dataset for training top-performing reward models](https://mlsanity.com/paper/2406.08673/) (2024)
- [SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models](https://mlsanity.com/paper/2402.05044/) (2024)
- [k-Sparse Autoencoders](https://mlsanity.com/paper/1312.5663/) (2013)
- [Defending Against Alignment-Breaking Attacks via Robustly Aligned LLM](https://mlsanity.com/paper/2309.14348/) (2023)
- [Linear Algebraic Structure of Word Senses, with Applications to Polysemy](https://mlsanity.com/paper/1601.03764/) (2016)
- [WebGLM: Towards An Efficient Web-Enhanced Question Answering System with Human Preferences](https://mlsanity.com/paper/2306.07906/) (2023)
- [Evaluating Layers of Representation in Neural Machine Translation on Part-of-Speech and Semantic Tagging Tasks](https://mlsanity.com/paper/1801.07772/) (2018)
- [From Language Modeling to Instruction Following: Understanding the Behavior Shift in LLMs after Instruction Tuning](https://mlsanity.com/paper/2310.00492/) (2023)
- [Evaluating Open-Source Sparse Autoencoders on Disentangling Factual Knowledge in GPT-2 Small](https://mlsanity.com/paper/2409.04478/) (2024)
- [Neuron to Graph: Interpreting Language Model Neurons at Scale](https://mlsanity.com/paper/2305.19911/) (2023)

## Built on by (most cited first, 21 of 21)

- [A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models](https://mlsanity.com/paper/2503.05613/) (2025)
- [Towards Trustworthy GUI Agents: A Survey](https://mlsanity.com/paper/2503.23434/) (2025)
- [SAE-SSV: Supervised Steering in Sparse Representation Spaces for Reliable Control of Language Models](https://mlsanity.com/paper/2505.16188/) (2025)
- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [Feature Extraction and Steering for Enhanced Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2505.15634/) (2025)
- [Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety](https://mlsanity.com/paper/2506.05451/) (2025)
- [SAGE: An Agentic Explainer Framework for Interpreting SAE Features in Language Models](https://mlsanity.com/paper/2511.20820/) (2025)
- [Concept-Level Explainability for Auditing & Steering LLM Responses](https://mlsanity.com/paper/2505.07610/) (2025)
- [Beyond Input Activations: Identifying Influential Latents by Gradient Sparse Autoencoders](https://mlsanity.com/paper/2505.08080/) (2025)
- [A Comparative Analysis of Sparse Autoencoder and Activation Difference in Language Model Steering](https://mlsanity.com/paper/2510.01246/) (2025)
- [SegCompass: Exploring Interpretable Alignment with Sparse Autoencoders for Enhanced Reasoning Segmentation](https://mlsanity.com/paper/2605.22658/) (2026)
- [Can GPT tell us why these images are synthesized? Empowering Multimodal Large Language Models for Forensics](https://mlsanity.com/paper/2504.11686/) (2025)
- [ProtSAE: Disentangling and Interpreting Protein Language Models via Semantically-Guided Sparse Autoencoders](https://mlsanity.com/paper/2509.05309/) (2025)
- [Towards Interpretable Hallucination Analysis and Mitigation in LVLMs via Contrastive Neuron Steering](https://mlsanity.com/paper/2602.00621/) (2026)
- [EmoPerso: Enhancing Personality Detection with Self-Supervised Emotion-Aware Modelling](https://mlsanity.com/paper/2509.02450/) (2025)
- [Rethinking the Understanding Ability across LLMs through Mutual Information](https://mlsanity.com/paper/2505.23790/) (2025)
- [ConceptViz: A Visual Analytics Approach for Exploring Concepts in Large Language Models](https://mlsanity.com/paper/2509.20376/) (2025)
- [What's on My Network? Using Large Language Models to Identify Real-World IoT Devices at Scale](https://mlsanity.com/paper/2510.13817/) (2025)
- [FlowNIB: An Information Bottleneck Analysis of Bidirectional vs. Unidirectional Language Models](https://mlsanity.com/paper/2506.00859/) (2025)
- [Steered Generation via Gradient-Based Optimization on Sparse Query Features](https://mlsanity.com/paper/2605.23040/) (2026)
- [SAERec: Constructing Fine-grained Interpretable Intents Priors via Sparse Autoencoders for Recommendation](https://mlsanity.com/paper/2606.18897/) (2026)

## Abstract (arXiv)

Large language models (LLMs) excel at handling human queries, but they can occasionally generate flawed or unexpected responses. Understanding their internal states is crucial for understanding their successes, diagnosing their failures, and refining their capabilities. Although sparse autoencoders (SAEs) have shown promise for interpreting LLM internal representations, limited research has explored how to better explain SAE features, i.e., understanding the semantic meaning of features learned by SAE. Our theoretical analysis reveals that existing explanation methods suffer from the frequency bias issue, where they emphasize linguistic patterns over semantic concepts, while the latter is more critical to steer LLM behaviors. To address this, we propose using a fixed vocabulary set for feature interpretations and designing a mutual information-based objective, aiming to better capture the semantic meaning behind these features. We further propose two runtime steering strategies that adjust the learned feature activations based on their corresponding explanations. Empirical results show that, compared to baselines, our method provides more discourse-level explanations and effectively steers LLM behaviors to defend against jailbreak attacks. These findings highlight the value of explanations for steering LLM behaviors in downstream applications. We will release our code and data once accepted.