# Mixture of Experts Made Intrinsically Interpretable

arXiv: https://arxiv.org/abs/2503.07639 · HTML: https://mlsanity.com/paper/2503.07639/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- Placed in MoE large language models, one of 12 directions in Mixture-of-experts.
- Builds on 41 of its 52 references that are in the map.
- First posted on arXiv 5 Mar 2025.

## Placed in

- Mixture-of-experts / [MoE large language models](https://mlsanity.com/moe/moe-large-language-models/)

## Builds on (most cited first, 30 of 41)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [The LAMBADA dataset: Word prediction requiring a broad discourse context](https://mlsanity.com/paper/1606.06031/) (2016)
- [Toy Models of Superposition](https://mlsanity.com/paper/2209.10652/) (2022)
- [Mixture-of-Experts with Expert Choice Routing](https://mlsanity.com/paper/2202.09368/) (2022)
- [Scaling and evaluating sparse autoencoders](https://mlsanity.com/paper/2406.04093/) (2024)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [Neural Additive Models: Interpretable Machine Learning with Neural Nets](https://mlsanity.com/paper/2004.13912/) (2020)
- [The Alignment Problem from a Deep Learning Perspective](https://mlsanity.com/paper/2209.00626/) (2022)
- [An Overview of Catastrophic AI Risks](https://mlsanity.com/paper/2306.12001/) (2023)
- [k-Sparse Autoencoders](https://mlsanity.com/paper/1312.5663/) (2013)
- [Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints](https://mlsanity.com/paper/2212.05055/) (2022)
- [RedPajama: an Open Dataset for Training Large Language Models](https://mlsanity.com/paper/2411.12372/) (2024)
- [The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks](https://mlsanity.com/paper/2306.17844/) (2023)
- [Large Memory Layers with Product Keys](https://mlsanity.com/paper/1907.05242/) (2019)
- [Acquisition of Chess Knowledge in AlphaZero](https://mlsanity.com/paper/2111.09259/) (2021)
- [Automatically Interpreting Millions of Features in Large Language Models](https://mlsanity.com/paper/2410.13928/) (2024)
- [ReLU Strikes Back: Exploiting Activation Sparsity in Large Language Models](https://mlsanity.com/paper/2310.04564/) (2023)
- [Chess as a Testbed for Language Model State Tracking](https://mlsanity.com/paper/2102.13249/) (2021)
- [Mixture of A Million Experts](https://mlsanity.com/paper/2407.04153/) (2024)
- [ReLU2 Wins: Discovering Efficient Activation Functions for Sparse LLMs](https://mlsanity.com/paper/2402.03804/) (2024)
- [Emergent World Models and Latent Variable Estimation in Chess-Playing Language Models](https://mlsanity.com/paper/2403.15498/) (2024)
- [Measuring Progress in Dictionary Learning for Language Model Interpretability with Board Game Models](https://mlsanity.com/paper/2408.00113/) (2024)

## Built on by (most cited first, 24 of 24)

- [Multilingual Routing in Mixture-of-Experts](https://mlsanity.com/paper/2510.04694/) (2025)
- [Coupling Experts and Routers in Mixture-of-Experts via an Auxiliary Loss](https://mlsanity.com/paper/2512.23447/) (2025)
- [Abstracting Robot Manipulation Skills via Mixture-of-Experts Diffusion Policies](https://mlsanity.com/paper/2601.21251/) (2026)
- [The Expert Strikes Back: Interpreting Mixture-of-Experts Language Models at Expert Level](https://mlsanity.com/paper/2604.02178/) (2026)
- [The Illusion of Specialization: Unveiling the Domain-Invariant "Standing Committee" in Mixture-of-Experts Models](https://mlsanity.com/paper/2601.03425/) (2026)
- [RouteHijack: Routing-Aware Attack on Mixture-of-Experts LLMs](https://mlsanity.com/paper/2605.02946/) (2026)
- [Intrinsically Interpretable Attention via Sparse Post-Training](https://mlsanity.com/paper/2512.05865/) (2025)
- [Towards Intrinsic Interpretability of Large Language Models:A Survey of Design Principles and Architectures](https://mlsanity.com/paper/2604.16042/) (2026)
- [EMO: Pretraining Mixture of Experts for Emergent Modularity](https://mlsanity.com/paper/2605.06663/) (2026)
- [Routers Learn the Geometry of Their Experts: Geometric Coupling in Sparse Mixture-of-Experts](https://mlsanity.com/paper/2605.12476/) (2026)
- [What Gets Activated: Uncovering Domain and Driver Experts in MoE Language Models](https://mlsanity.com/paper/2601.10159/) (2026)
- [Sparsity and Superposition in Mixture of Experts](https://mlsanity.com/paper/2510.23671/) (2025)
- [Sparse Mixture-of-Experts Reward Models Learn Interpretable and Specialized Experts for Personalized Preference Modeling](https://mlsanity.com/paper/2606.04284/) (2026)
- [Position: Certifiable State Integrity Should Be Built from Local Validity, Not Global Scale](https://mlsanity.com/paper/2601.21249/) (2026)
- [Revisiting Transformers with Insights from Image Filtering and Boosting](https://mlsanity.com/paper/2506.10371/) (2025)
- [Hecto: Modular Sparse Experts for Adaptive and Interpretable Reasoning](https://mlsanity.com/paper/2506.22919/) (2025)
- [ReFIne: A Framework for Trustworthy Large Reasoning Models with Reliability, Faithfulness, and Interpretability](https://mlsanity.com/paper/2510.09062/) (2025)
- [Toward Structural Multimodal Representations: Specialization, Selection, and Sparsification via Mixture-of-Experts](https://mlsanity.com/paper/2605.03348/) (2026)
- [Beyond Routing: Characterising Expert Tuning and Representation in Vision Mixture-of-Experts](https://mlsanity.com/paper/2605.20610/) (2026)
- [Routing Divergence Is Not Evidence of Behavioral Influence in Same-Weight MoE Self-Distillation](https://mlsanity.com/paper/2608.15787/) (2026)
- [A Heterogeneous Mixture of Experts Framework for Interpretable Machine Learning](https://mlsanity.com/paper/2608.24195/) (2026)
- [A Persistent State for Auditable Mixture-of-Experts Routing](https://mlsanity.com/paper/2609.34634/) (2026)
- [Frequency Is Not Sensitivity Identifying Safety-Sensitive Experts in Sparse MoE LLM](https://mlsanity.com/paper/2610.02910/) (2026)
- [InstMoE: Adaptive Multimodal Routing with Specialized Experts](https://mlsanity.com/paper/2610.05111/) (2026)

## Abstract (arXiv)

Neurons in large language models often exhibit polysemanticity, simultaneously encoding multiple unrelated concepts and obscuring interpretability. Instead of relying on post-hoc methods, we present MoE-X, a Mixture-of-Experts (MoE) language model designed to be intrinsically interpretable. Our approach is motivated by the observation that, in language models, wider networks with sparse activations are more likely to capture interpretable factors. However, directly training such large sparse networks is computationally prohibitive. MoE architectures offer a scalable alternative by activating only a subset of experts for any given input, inherently aligning with interpretability objectives. In MoE-X, we establish this connection by rewriting the MoE layer as an equivalent sparse, large MLP. This approach enables efficient scaling of the hidden size while maintaining sparsity. To further enhance interpretability, we enforce sparse activation within each expert and redesign the routing mechanism to prioritize experts with the highest activation sparsity. These designs ensure that only the most salient features are routed and processed by the experts. We evaluate MoE-X on chess and natural language tasks, showing that it achieves performance comparable to dense models while significantly improving interpretability. MoE-X achieves a perplexity better than GPT-2, with interpretability surpassing even sparse autoencoder (SAE)-based approaches.