skills.rest
robertbr123
robertbr123COMMUNITY
Explore all skills by this author →
WebsiteGitHub

sparse-autoencoder-training

Train and analyze Sparse Autoencoders to decompose neural network activations into interpretable features.

Updated May 9, 2026
Search Tags:#machine learning#feature discovery#SAE#sparse autoencoder#neural network interpretability
One-click install
npx skills add https://github.com/robertbr123/Linket-Agent --skill sparse-autoencoder-training-robertbr123
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sparse-autoencoder-training
Source: https://github.com/robertbr123/Linket-Agent/tree/main/optional-skills/mlops/saelens
Command: npx skills add https://github.com/robertbr123/Linket-Agent --skill sparse-autoencoder-training-robertbr123

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires sae-lens>=6.0.0, transformer-lens>=2.0.0, torch>=2.0.0, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill addresses the challenge of interpretability in neural networks by providing a framework for training and analyzing Sparse Autoencoders (SAEs), which decompose neural network activations into interpretable features.

Core Features & Use Cases

  • Feature Discovery: Discover interpretable features in model activations.
  • Superposition Analysis: Understand the superposition of features in neural networks.
  • Use Case: For a language model, identify the features that activate when the model predicts a specific word or concept.

Quick Start

Train a Sparse Autoencoder on the GPT-2 model to analyze feature activation for the input 'The capital of France is Paris'.

Frequently Asked Questions about sparse-autoencoder-training

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I train a sparse autoencoder to find interpretable features in neural network activations?▼

You can analyze feature activation for specific inputs like 'The capital of France is Paris' by training a sparse autoencoder on the GPT-2 model. This helps identify which features activate when the model predicts a specific word or concept.

What is superposition analysis in neural networks and how does a sparse autoencoder help?▼

You need Python 3.10 or higher. The framework specifically requires transformer-lens version 2.0.0 or higher and torch version 2.0.0 or higher to run sparse autoencoder training and analysis.

Does this sparse autoencoder framework work with transformer-lens and GPT-2 models?▼

You need Python 3.10 or higher. The framework specifically requires transformer-lens version 2.0.0 or higher and torch version 2.0.0 or higher to run sparse autoencoder training and analysis.

Do I need sae-lens to discover interpretable features in language model activations?▼

Yes, sae-lens version 6.0.0 or higher is required. It works alongside transformer-lens and torch to train sparse autoencoders for decomposing neural network activations into interpretable features.

Related Skills

Explore More Software Engineering Skills→
letta-ailetta-ai
134

gpt2-codegolf

Implement minimal GPT-2 inference with checkpoint parsing and BPE tokenization.

Official
Advanced
tianhao909tianhao909
1

sparse-autoencoder-training

Train and analyze Sparse Autoencoders to decompose neural network activations into interpretable features.

Community
Advanced
AissenLiuAissenLiu

sparse-autoencoder-training

Train and analyze Sparse Autoencoders with SAELens on transformer activations.

Community
Advanced
hochoa13hochoa13

sparse-autoencoder-training

Train and analyze Sparse Autoencoders to decompose neural network activations into interpretable features.

Community
Advanced