Skip to content

ML Application · Python

silico-e4b-sae-run

What it is

The repository trains a 42-layer JumpReLU Sparse Autoencoder (SAE) atlas for the Gemma 4 E4B language model using the E2B recipe and an event-aware trainer with automatic sparsity control. It generates a private Hugging Face dataset containing the SAEs and includes scripts for training, verification, and result aggregation.

Scored from the code at its last scan, 22 Oct 2026. How merit is scored

Where it counts

  • Machine Learning80%
  • NLP & LLMs80%
  • MLOps30%
  • Data Science20%

How much of the project is about each field, as scoring counts it. It ranks in the field it is most about.

Around it on the Machine Learning board

Whole board
  1. 09, shared rankmamba-viewAI Developer Tool · Python62.2/100AdvancedRick
  2. 09, shared rankResNet_Family-from-scratchArchitecture Reimplementation · Jupyter Notebook62.2/100AdvancedPop Alexandru
  3. 11, shared rankvid-factoryML Application · JavaScript62.0/100AdvancedJarad Peckenpaugh
  4. 11, shared ranksilico-e4b-sae-runML Application · Python62.0/100AdvancedRick
  5. 13mamba3-atlas-imageAI Developer Tool · Python61.5/100AdvancedRick

More from Rick