Sparse Autoencoder Training

davila7

Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable features, analyzing superposition, or studying monosemantic representations in language models.

Install

npx skills add davila7/claude-code-templates --skill sparse-autoencoder-training

npx skills add davila7/claude-code-templates --skill sparse-autoencoder-training -a claude-code -g

npx skills add davila7/claude-code-templates --skill sparse-autoencoder-training -a codex -g

gemini skills install https://github.com/davila7/claude-code-templates.git --path cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens

gh skill install davila7/claude-code-templates sparse-autoencoder-training

git clone --depth 1 https://github.com/davila7/claude-code-templates /tmp/davila7__claude-code-templates
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/davila7__claude-code-templates/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens ~/.claude/skills/sparse-autoencoder-training   # Claude Code
cp -R /tmp/davila7__claude-code-templates/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens ~/.agents/skills/sparse-autoencoder-training   # Codex, Cursor, Gemini CLI, Copilot, OpenCode

Install

Related