Search Results - RepositoryStats

34

605

unknown

5

A complete end-to-end pipeline for LLM interpretability with sparse autoencoders (SAEs) using Llama 3.2, written in pure PyTorch and fully reproducible.

llama3 pytorch open-research feature-steering feature-extraction sparse-autoencoder llm-interpretability

This repository has been archived (exclude archived)

Created 2024-11-21

2 commits to main branch, last one about a month ago

repeng vgel

42

537

mit

5

A library for making RepE control vectors

sae saes transformers language-model machine-learning sparse-autoencoder sparse-autoencoders representation-engineering

Created 2024-01-21

28 commits to main branch, last one 19 days ago

Awesome-Interpretability-in-Large-Language-Models ruizheliUOA

19

309

cc0-1.0

5

This repository collects all relevant resources about interpretability in LLMs

sparse-autoencoder dictionary-learning mechanistic-interpretability interpretability-and-explainability

Created 2024-06-30

56 commits to main branch, last one 2 months ago

sansa glami

4

40

apache-2.0

8

SANSA - sparse EASE for millions of items

sparse-matrix recommender-system sparse-autoencoder approximate-inverse collaborative-filtering

Created 2023-07-11

71 commits to main branch, last one 19 days ago

ravel explanare

6

36

mit

1

Evaluate interpretability methods on localizing and disentangling concepts in LLMs.

probing intervention interpretability sparse-autoencoder causal-intervention disentangled-representations

Created 2024-02-17

13 commits to main branch, last one 3 months ago