Explore visualization tools for understanding Transformer-based large language models (LLMs):
-
Transformer Explainer
Learn How Transformer Models Work with Interactive Visualization
Georgia Tech and IBM, 2024.08
Demo / GitHub / arXiv -
Gemma Scope
Help the safety community shed light on the inner workings of language models
Google DeepMind, 2024.07
Demo / Blog / PDF -
LLM Transparency Tool
An open-source interactive toolkit for analyzing internal workings of Transformer-based language models.
Meta, 2024.04
Demo / GitHub / arXiv -
Neuronpedia
Neuronpedia is a platform for mechanistic interpretability research. Its goal is to accelerate researchers for Sparse Autoencoders (SAEs) by hosting models, feature dashboards, data visualizations, tooling, and more.
Johnny Lin and Joseph Bloom, 2024.03
Demo -
CircuitsVis
Mechanistic Interpretability visualizations, that work both in both Python (e.g. with Jupyter Lab) and JavaScript (e.g. React or plain HTML).
Alan Cooney and Neel Nanda, 2023.10
Demo / GitHub -
LLM Visualization
A visualization and walkthrough of the LLM algorithm that backs OpenAI's ChatGPT. Explore the algorithm down to every add & multiply, seeing the whole process in action.
Brendan Bycroft, 2023.05
Demo / GitHub -
TransformerLens
A library for mechanistic interpretability of GPT-style language models
Neel Nanda and Joseph Bloom, 2022.08
GitHub / Distill / Documentation -
BertViz
Visualize Attention in NLP Models (BERT, GPT2, BART, etc.)
Jesse Vig, 2019.07
GitHub / ACL Anthology