Skip to content
Live
Trending storyPast story

ByteDance released the Qwen3-30B-A3B MoE-SAE model

1 reports1 outlets13 days agoUpdate

The story so far

AI overview

ByteDance released the Qwen3-30B-A3B MoE-Output Sparse Autoencoders model, which includes TopK sparse autoencoders (k=100, 2^20 latent variables) trained at the 4th, 30th, and 46th layers of the Qwen3-30B-A3B model. Each SAE is approximately 17.2 GB, with reconstruction variance unexplained scores of 12.6%, 25.1%, and 7.6% respectively.

AI-written from the reports · Updated 13 days ago

Coverage timeline

Follow the coverage to see every side of the story.

5 Nov
  1. ByteDance Seed · Hugging Face new models
    ByteDance-Seed/Qwen3-30B-A3B-MoE-SAE

    ByteDance-Seed/Qwen3-30B-A3B-MoE-SAE is released, containing TopK sparse autoencoders (k=100, 2^20 latent variables) trained on layers 4, 30, and 46 of the Qwen3-30B-A3B model. Each SAE is approximately 17.2 GB, with reconstruction variance unexplained scores of 12.6%, 25.1%, and 7.6% respectively.