AI Paper Atlas

Research map / Architectures and efficiency / Mixture of experts

Not All Models Suit Expert Offloading: On Local Routing Consistency of Mixture-of-Expert Models

Jingcong Liang, Siyuan Wang, Miren Tian, Yitong Li, Duyu Tang, zhongyu wei

ICLR 2026 · Architectures and efficiency › Mixture of experts · approach: Expert routing and specialization · 8 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
remove the load-balancing loss (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 56.42 (load-balance SD 4.02 -> 13.21)Improved
prioritize load balance more strongly in training (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 36.42Hurt
add shared experts (one or two) (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 41.38 (1 shared), 38.79 (2 shared)Hurt
activate fewer experts per token (Top2) (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 27.13Hurt
activate more experts per token (Top16) (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 55.69Improved
use fewer total experts (32Exp) (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 41.62Hurt
make the first layer dense or interleave dense layers (ablation)toy MoE Baseline · toy models, m=16SRP: 43.56 -> 44.87 (Skip1), 43.67 (Sparse2)No clear effect
increase cache size from 1x to 2x active expertsoracle segment cache (SCH) · Baseline toy model, m=16relative hit rate: 80.97 -> 90.55 (3x: 96.23)Improved
use oracle segment cache instead of LRU or LFUSCH vs LRU/LFU · Baseline toy model, 2x cacherelative hit rate: 67.04 (LRU), 70.87 (LFU) -> 90.55 (SCH)Improved
apply SFT, DPO or instruction tuning after pretraining (ablation)OLMoE · REAL models, SRPSRP: +0.47 SRP at m=16 after SFTNo clear effect
domain-specialized experts vs vocabulary-specialized expertsMoE LLMs · REAL modelsSRP correlationImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy