Skip to content
View DavidUdell's full-sized avatar
🧫
Alignment science
🧫
Alignment science
5 followers · 8 following

Achievements

Achievement: YOLOAchievement: Pull Sharkx2Achievement: Quickdraw

Achievements

Achievement: YOLOAchievement: Pull Sharkx2Achievement: Quickdraw

Block or report DavidUdell

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
DavidUdell/README.md

Work Showcase

sparse_circuit_discovery is a clean PyTorch reimplementation, for GPT-2-small, of the Marks et al. (2024) unsupervised circuit discovery algorithm. It also contains a simpler "naive ablations" algorithm for circuit discovery; some results from the ablations algorithm are presented here and here.

Pinned Loading

  1. sparse_circuit_discovery sparse_circuit_discovery Public

    Circuit discovery in GPT-2 small, using sparse autoencoding

    Python 8 1

  2. AI_2027_Widgets AI_2027_Widgets Public

    Widgets for AI 2027's blog post

    JavaScript 1

  3. montemac/activation_additions montemac/activation_additions Public

    Algebraic value editing in pretrained language models

    Jupyter Notebook 67 16

  4. feature-circuits feature-circuits Public

    Forked from JacksonKaunismaa/feature-circuits

    Discovering and editing interpretable causal graphs in language models

    Python 2

  5. persistent persistent Public

    A proof-of-concept persistent AI agent implementation

    Python 2

  6. simbrain/NeuralNetworksCogSciBook simbrain/NeuralNetworksCogSciBook Public

    Neural Networks in Cognitive Science

    TeX 16 2