🎯 Mixture of experts

The items that name it 17

• [AINews] Google admits AI Mode hurts Search revenue; -15% $500B drop worst since 2004 IPO. AI Bear Market coming?
  article2026-10-31

  Architecture of Mistral Large 4 (128 routed experts, 4 active plus 1 shared) and Nemotron 3.5.

• Multi-View MoE: Letting Every Router Shape Expert Computation
  paper2026-10-29

  Sparse models expand parameter capacity by activating a small subset of experts per token.

• On the Fundamental Limits of Prefill-Based Decoding Length Prediction
  paper2026-10-29It is written here as Mixture-of-Experts.

  A predictor variant tested for decode length; it converges to similar performance near the bound.

• 🚨 AI News | TestingCatalog (@testingcatalog) on X
  thread2026-10-29It is written here as Mixture of Experts (MoE).

  Nemotron 3.5 Transcribe is a 30B MoE model with about 4B active parameters.

• Assisted Turbulence Modeling via Data Assimilation and Machine Learning for Turbomachinery Flows
  paper2026-10-19It is written here as Mixture of Experts (MoE).

  MoE

• DeepSeek V4.1 Flash 1017 (Max)
  other2026-10-19It is written here as Mixture of Experts (MoE).

  MoE

• Master the captain (@mollicksoif9) on X
  thread2026-10-17It is written here as Mixture-of-Experts (MoE).

  MoE

• Rethinking Routing Exploration in Modern Mixture-of-Experts
  paper2026-10-15It is written here as Mixture-of-experts (MoE).

  MoE

• Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta
  article2026-10-15
• Mistral AI Drops 'Le Chonk': A Massive AI Model Named After a Cat Meme
  article2026-10-06
• Sahil Verma (@sahilvermaai) on X
  thread2026-10-04
• Aleph Alpha Kolibri: How the sovereign German LLM works
  article2026-10-03It is written here as Mixture of experts (MoE).

  MoE

• Harnessing Domain Specialists in Multimodal Mixture-of-Experts for Efficient Adaptation
  paper2026-10-01It is written here as Mixture-of-Experts (MoE).

  MoE

• Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs
  article2026-10-01It is written here as Mixture-of-experts (MoE).

  MoE

• How to Loop MoE: Flatten the Experts, Untie the Attention
  paper2026-09-28It is written here as Mixture-of-experts (MoE).

  MoE

• Computation and Data Movement for Inference
  article2026-09-21It is written here as Mixture of Experts (MoE).

  Changes which tensors are active per token; a router picks a few experts per token, so serving, memory movement and batching differ from dense models.

• Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data
  paper2026-09-10It is written here as Mixture-of-Experts (MoE).

  MoE