SAE 解讀 MoE:稀疏自編碼器揭示 MoE 專傢俱有人類可解讀的細粒度專門化
SAE Interprets MoE: Sparse AutoEncoders Reveal That MoE Experts Exhibit Human-Interpretable Fine-Grained Specialization
研究以稀疏自編碼器分析 Qwen3-30B-A3B,發現 MoE 專家對潛在特徵的貢獻比路由選擇更集中。第 4、30 和 46 層中,大多數潛在特徵只有一至兩個主導專家;專家分工細至 Java 與前端網頁程式碼、幾何、算術和多項式等領域。稀疏自編碼器及程式碼已公開。
Published on Nov 5
Authors:
,
,
Abstract
A sparse autoencoder trained on mixture-of-experts layer outputs reveals that individual experts drive specific latent features, showing finer-grained and more concentrated specialization than routing or domain labels alone.
Mixture-of-Experts (MoE) models route each token to a few of many experts, yet what an individual expert is responsible for remains unsettled. Existing analyses count how often an expert is selected on domain-labeled corpora, an instrument that fixes the taxonomy in advance, stays coarse, and records selection rather than contribution. We instead study specialization in feature space by training a sparse autoencoder (SAE) on the output of an MoE layer. Since this output is a router-weighted sum of expert outputs and the SAE encoder is affine, the pre-activation of every latent decomposes exactly over the experts. Aggregated over a corpus, this gives each expert a dominant set of latents whose activation it principally drives. On layers 4, 30 and 46 of Qwen3-30B-A3B, contribution is far more concentrated than gating, most latents have only one or two dominators, and the overlap between dominant sets concentrates on a few expert pairs. Auto-interpreting these sets reveals a division of labor below coarse domains, with distinct experts for Java and front-end web code, or for geometry and for arithmetic and polynomials, while the coarse domains themselves take very unequal numbers of experts. Our SAEs and code are available at https://github.com/Bytedance-Seed/SAE-Interprets-MoE.
View arXiv page View PDF Project page GitHub 1 Add to collection
Models citing this paper 1
Updated about 2 hours ago • 1ByteDance-Seed/Qwen3-30B-A3B-MoE-SAE
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2611.08089 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2611.08089 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co