Skip to content
@apartresearch

apartresearch

Artificial intelligence will change the world. Our mission is to ensure this happens safely and to the benefit of everyone.
README.md

Apart facilitates new research in AI safety, towards reducing societal-scale risks from the technology.

We combine a community focus with a drive for high-quality security research.


Read more about our work:

  • Our Research — Foundational research for safe and beneficial advanced AI
  • Apart Lab — Our research fellowship program for aspiring researchers in AI safety
  • Apart Sprints — Weekend-long research sprints and hackathons for AI security and governance

Twitter Badge LinkedIn Badge YouTube Badge Discord Badge Alignment Jam RSS Badge

Pinned Loading

  1. interpretability-starter interpretability-starter Public

    🧠 Starter templates for doing interpretability research

    70 2

  2. Neuron2Graph Neuron2Graph Public

    Tools for exploring Transformer neuron behaviour, including input pruning and diversification.

    Jupyter Notebook 20 5

  3. deepdecipher deepdecipher Public

    🦠 DeepDecipher: An open source API to MLP neurons

    Rust 9

  4. specificityplus specificityplus Public

    👩‍💻 Code for the ACL paper "Detecting Edit Failures in LLMs: An Improved Specificity Benchmark"

    Python 20 4

  5. Integer_Addition Integer_Addition Public

    ✱ Understanding the underlying learning dynamics of simple tasks in Transformer networks

    Jupyter Notebook 16

  6. readingwhatwecan readingwhatwecan Public

    📚📚📚📚📚📚📚📚📚 Reading everything

    CSS 14 3

Repositories

Showing 10 of 41 repositories
View all repositories

Top languages

Loading…

Most used topics

Loading…