Unveiling Super Experts in Mixture-of-Experts Large Language Models

• URL: http://arxiv.org/abs/2507.23279v1
• Date: Thu, 31 Jul 2025 06:35:33 GMT
• Title: Unveiling Super Experts in Mixture-of-Experts Large Language Models
• Authors: Zunhai Su, Qingyuan Li, Hao Zhang, YuLei Qian, Yuchen Xie, Kehong Yuan,
• Abstract summary: Sparsely activated Mixture-of-Experts (MoE) models have shown promise in enhancing the learning capacity of large language models (LLMs)<n>We present the first discovery and investigation of a distinct subset of experts that play a crucial role in the underlying mechanisms during the model's forward inference.<n>Our findings confirm that MoE LLMs rely on SEs to induce attention sinks, which are crucial for the distribution of attention scores but are significantly disrupted by SE pruning.
• Score: 5.36587814108884
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Sparsely activated Mixture-of-Experts (MoE) models have shown promise in enhancing the learning capacity of large language models (LLMs). Leveraging the intrinsic importance differences among experts, recent research has explored expert-level compression techniques to improve the efficiency of MoE LLMs. However, existing approaches often rely on empirical criteria to identify critical experts, lacking a deeper exploration and understanding of the heterogeneous importance of experts. In this study, we present the first discovery and investigation of a distinct subset of experts that play a crucial role in the underlying mechanisms during the model's forward inference. These experts are prevalent in open-source MoE LLMs, and despite their limited number, pruning them leads to a significant decline in model performance (e.g., pruning three causes Qwen3-30B-A3B to produce repetitive and uninformative outputs). We refer to these experts as Super Experts (SEs). Our comprehensive analysis provides progressively deeper insights into SEs. (i) SEs are characterized by rare but extreme activation outliers in the output of the down_proj, which give rise to massive activations in the hidden states between decoder layers. Moreover, the distribution of SEs remains model-specific and is unaffected by post-training processes. (ii) By pruning SEs, we assess their significance across a variety of tasks, revealing their considerable impact on the model's overall performance, particularly in mathematical reasoning. (iii) We further enhance our understanding of the influence of SEs compression. Our findings confirm that MoE LLMs rely on SEs to induce attention sinks, which are crucial for the distribution of attention scores but are significantly disrupted by SE pruning. The code is available at https://github.com/ZunhaiSu/Super-Experts-Profilling.

Related papers

• Unveiling Hidden Collaboration within Mixture-of-Experts in Large Language Models [5.211806751260724]
  We propose a hierarchical sparse dictionary learning (HSDL) method that uncovers the collaboration patterns among experts.<n>We also introduce the Contribution-Aware Expert Pruning (CAEP) algorithm, which effectively prunes low-contribution experts.
  arXiv  Detail & Related papers  (2025-04-16T04:06:15Z)
• Domain-Specific Pruning of Large Mixture-of-Experts Models with Few-shot Demonstrations [48.890534958441016]
  This study investigates domain specialization and expert redundancy in large-scale MoE models.<n>We propose a simple yet effective pruning framework, EASY-EP, to identify and retain only the most relevant experts.<n>Experiments on DeepSeek-R1 and DeepSeek-V3-0324 show that our method can achieve comparable performances and $2.99times$ throughput under the same memory budget with full model with only half the experts.
  arXiv  Detail & Related papers  (2025-04-09T11:34:06Z)
• Finding Fantastic Experts in MoEs: A Unified Study for Expert Dropping Strategies and Observations [86.90549830760513]
  Sparsely activated Mixture-of-Experts (SMoE) has shown promise in scaling up the learning capacity of neural networks.<n>We propose MoE Experts Compression Suite (MC-Suite) to provide a benchmark for estimating expert importance from diverse perspectives.<n>We present an experimentally validated conjecture that, during expert dropping, SMoEs' instruction-following capabilities are predominantly hurt.
  arXiv  Detail & Related papers  (2025-04-08T00:49:08Z)
• Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time [1.1655046053160683]
  We present a method that extends the Mixture-of-Experts architecture of Large Language Models (LLMs)<n>MoTE enables meaningful and focused behavior changes in LLMs on-the-fly during inference time.
  arXiv  Detail & Related papers  (2025-02-16T12:24:39Z)
• Unveiling and Consulting Core Experts in Retrieval-Augmented MoE-based LLMs [64.9693406713216]
  Internal mechanisms that contribute to the effectiveness of RAG systems remain underexplored. Our experiments reveal that several core groups of experts are primarily responsible for RAG-related behaviors. We propose several strategies to enhance RAG's efficiency and effectiveness through expert activation.
  arXiv  Detail & Related papers  (2024-10-20T16:08:54Z)
• Diversifying the Expert Knowledge for Task-Agnostic Pruning in Sparse Mixture-of-Experts [75.85448576746373]
  We propose a method of grouping and pruning similar experts to improve the model's parameter efficiency.<n>We validate the effectiveness of our method by pruning three state-of-the-art MoE architectures.<n>The evaluation shows that our method outperforms other model pruning methods on a range of natural language tasks.
  arXiv  Detail & Related papers  (2024-07-12T17:25:02Z)
• Let the Expert Stick to His Last: Expert-Specialized Fine-Tuning for Sparse Architectural Large Language Models [24.915387910764082]
  Expert-Specialized Fine-Tuning, or ESFT, tunes the experts most relevant to downstream tasks while freezing the other experts and modules. MoE models with finer-grained experts are more advantageous in selecting the combination of experts that are most relevant to downstream tasks.
  arXiv  Detail & Related papers  (2024-07-02T03:11:13Z)
• Generalization Error Analysis for Sparse Mixture-of-Experts: A Preliminary Study [65.11303133775857]
  Mixture-of-Experts (MoE) computation amalgamates predictions from several specialized sub-models (referred to as experts) Sparse MoE selectively engages only a limited number, or even just one expert, significantly reducing overhead while empirically preserving, and sometimes even enhancing, performance.
  arXiv  Detail & Related papers  (2024-03-26T05:48:02Z)
• MoEC: Mixture of Expert Clusters [93.63738535295866]
  Sparsely Mixture of Experts (MoE) has received great interest due to its promising scaling capability with affordable computational overhead. MoE converts dense layers into sparse experts, and utilizes a gated routing network to make experts conditionally activated. However, as the number of experts grows, MoE with outrageous parameters suffers from overfitting and sparse data allocation.
  arXiv  Detail & Related papers  (2022-07-19T06:09:55Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.