Ica Lens

Use this skill when applying Independent Component Analysis as a training-free interpretability lens — decomposing a target activation site (residual stream, MLP output, attention-head output, or any cached hook point) into maximally non-Gaussian directions and treating each direction as a candidate monosemantic component for sparse-probing, targeted perturbation, annotation, and cross-comparison against trained SAE / transcoder / crosscoder features.

zjunlp Updated 1.5k repo stars

File contents

zjunlp/mechanist commit 56ab71883c

Frequently asked questions

npx get-skill add zjunlp/ica-lens