登录 注册
登录 注册

EMO: Pretraining Mixture of Experts for Emergent Modularity

🔗 访问原文
🔗 Access Paper

📝 摘要
Abstract

Large language models are typically deployed as monolithic systems, requiring the full model even when applications need only a narrow subset of capabilities, e.g., code, math, or domain-specific knowledge. Mixture-of-Experts (MoEs) seemingly offer a potential alternative by activating only a subset of experts per input, but in practice, restricting inference to a subset of experts for a given domain leads to severe performance degradation. This limits their practicality in memory-constrained settings, especially as models grow larger and sparser. We introduce EMO, an MoE designed for modularity-the independent use and composition of expert subsets-without requiring human-defined priors. Our key idea is to encourage tokens from similar domains to rely on similar experts. Since tokens within a document often share a domain, EMO restricts them to select experts from a shared pool, while allowing different documents to use different pools. This simple constraint enables coherent expert groupings to emerge during pretraining using document boundaries alone. We pretrain a 1B-active, 14B-total EMO on 1T tokens. As a full model, it matches standard MoE performance. Crucially, it enables selective expert use: retaining only 25% (12.5%) of experts incurs just a 1% (3%) absolute drop, whereas standard MoEs break under the same setting. We further find that expert subsets in EMO specialize at semantic levels (e.g., domains such as math or code), in contrast to the low-level syntactic specialization observed in standard MoEs. Altogether, our results demonstrate a path toward modular, memory-efficient deployment of large, sparse models and open new opportunities for composable architectures.

📊 文章统计
Article Statistics

基础数据
Basic Stats

190 浏览
Views
0 下载
Downloads
23 引用
Citations

引用趋势
Citation Trend

阅读国家分布
Country Distribution

阅读机构分布
Institution Distribution

月度浏览趋势
Monthly Views

相关关键词
Related Keywords

影响因子分析
Impact Analysis

7.00 综合评分
Overall Score
引用影响力
Citation Impact
浏览热度
View Popularity
下载频次
Download Frequency

📄 相关文章
Related Articles

🌊