Skill · em Dados, IA e pesquisa

sparse-autoencoder-training

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.

Procedência

Antes de instalar

4 arquivos · 30,5 KB · só texto, nenhum script

Instalar na sua CLI

O comando baixa a versão fixada (commit 57f899e) direto da origem, para a pasta que a CLI lê. Precisa de curl (macOS e Linux); no Windows não há comando, porque o Rook Labs é para macOS.

Claude Code

Neste projeto: instala em .claude/skills/sparse-autoencoder-training/.

d=".claude/skills/sparse-autoencoder-training"
u="https://raw.githubusercontent.com/davila7/claude-code-templates/57f899e5394bb8ca166f38eacae8f0853cbfe033/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens"
curl -fsSL --create-dirs \
  -o "$d/SKILL.md" "$u/SKILL.md" \
  -o "$d/references/api.md" "$u/references/api.md" \
  -o "$d/references/README.md" "$u/references/README.md" \
  -o "$d/references/tutorials.md" "$u/references/tutorials.md"

Global: instala em ~/.claude/skills/sparse-autoencoder-training/.

d="$HOME/.claude/skills/sparse-autoencoder-training"
u="https://raw.githubusercontent.com/davila7/claude-code-templates/57f899e5394bb8ca166f38eacae8f0853cbfe033/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens"
curl -fsSL --create-dirs \
  -o "$d/SKILL.md" "$u/SKILL.md" \
  -o "$d/references/api.md" "$u/references/api.md" \
  -o "$d/references/README.md" "$u/references/README.md" \
  -o "$d/references/tutorials.md" "$u/references/tutorials.md"

Codex

Neste projeto: instala em .agents/skills/sparse-autoencoder-training/.

d=".agents/skills/sparse-autoencoder-training"
u="https://raw.githubusercontent.com/davila7/claude-code-templates/57f899e5394bb8ca166f38eacae8f0853cbfe033/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens"
curl -fsSL --create-dirs \
  -o "$d/SKILL.md" "$u/SKILL.md" \
  -o "$d/references/api.md" "$u/references/api.md" \
  -o "$d/references/README.md" "$u/references/README.md" \
  -o "$d/references/tutorials.md" "$u/references/tutorials.md"

Global: instala em ~/.agents/skills/sparse-autoencoder-training/.

d="$HOME/.agents/skills/sparse-autoencoder-training"
u="https://raw.githubusercontent.com/davila7/claude-code-templates/57f899e5394bb8ca166f38eacae8f0853cbfe033/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens"
curl -fsSL --create-dirs \
  -o "$d/SKILL.md" "$u/SKILL.md" \
  -o "$d/references/api.md" "$u/references/api.md" \
  -o "$d/references/README.md" "$u/references/README.md" \
  -o "$d/references/tutorials.md" "$u/references/tutorials.md"

Antigravity

Neste projeto: instala em .agents/skills/sparse-autoencoder-training/.

d=".agents/skills/sparse-autoencoder-training"
u="https://raw.githubusercontent.com/davila7/claude-code-templates/57f899e5394bb8ca166f38eacae8f0853cbfe033/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens"
curl -fsSL --create-dirs \
  -o "$d/SKILL.md" "$u/SKILL.md" \
  -o "$d/references/api.md" "$u/references/api.md" \
  -o "$d/references/README.md" "$u/references/README.md" \
  -o "$d/references/tutorials.md" "$u/references/tutorials.md"

Global: instala em ~/.gemini/antigravity-cli/skills/sparse-autoencoder-training/.

d="$HOME/.gemini/antigravity-cli/skills/sparse-autoencoder-training"
u="https://raw.githubusercontent.com/davila7/claude-code-templates/57f899e5394bb8ca166f38eacae8f0853cbfe033/cli-tool/components/skills/ai-research/mechanistic-interpretability-saelens"
curl -fsSL --create-dirs \
  -o "$d/SKILL.md" "$u/SKILL.md" \
  -o "$d/references/api.md" "$u/references/api.md" \
  -o "$d/references/README.md" "$u/references/README.md" \
  -o "$d/references/tutorials.md" "$u/references/tutorials.md"

Peça ao Rook

Já usa o Rook Labs? Cole no chat do Rook: instale a skill https://rooklabs.sh/marketplace/cct.sparse-autoencoder-training

Prévia do SKILL.md

---
name: sparse-autoencoder-training
description: 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.
version: 1.0.0
author: Orchestra Research
license: MIT
tags: [Sparse Autoencoders, SAE, Mechanistic Interpretability, Feature Discovery, Superposition]
dependencies: [sae-lens>=6.0.0, transformer-lens>=2.0.0, torch>=2.0.0]
---

# SAELens: Sparse Autoencoders for Mechanistic Interpretability

SAELens is the primary library for training and analyzing Sparse Autoencoders (SAEs) - a technique for decomposing polysemantic neural network activations into sparse, interpretable features. Based on Anthropic's groundbreaking research on monosemanticity.

**GitHub**: [jbloomAus/SAELens](https://github.com/jbloomAus/SAELens) (1,100+ stars)

## The Problem: Polysemanticity & Superposition

Individual neurons in neural networks are **polysemantic** - they activate in multiple, semantically distinct contexts. This happens because models use **superposition** to represent more features than they have neurons, making interpretability difficult.

**SAEs solve this** by decomposing dense activations into sparse, monosemantic features - typically only a small number of features activate for any given input, and each feature corresponds to an interpretable concept.

## When to Use SAELens

**Use SAELens when you need to:**
- Discover interpretable features in model activations
- Understand what concepts a model has learned
- Study superposition and feature geometry
- Perform feature-based steering or ablation
- Analyze safety-relevant features (deception, bias, harmful content)

**Consider alternatives when:**
- You need basic activation analysis → Use **TransformerLens** directly
- You want causal intervention experiments → Use **pyvene** or **TransformerLens**
- You need production steering → Consider direct activation engineering

## Installation

```bash
pip install sae-lens
…

Ver todo o marketplace