The nnsight package enables interpreting and manipulating the internals of deep learned models.

Official Code for Stable Cascade

Generative Representational Instruction Tuning

The Art of Debugging

Machine Learning Engineering Open Book

A framework for few-shot evaluation of language models.

LLM vulnerability scanner

The hub for EleutherAI's work on interpretability and learning dynamics

Tools for understanding how transformer predictions are built layer-by-layer

Taming Transformers for High-Resolution Image Synthesis

A repo for distributed training of language models with Reinforcement Learning via Human Feedback (RLHF)

Toolkit for creating, sharing and using natural language prompts.

High-Resolution Image Synthesis with Latent Diffusion Models

A dataset of alignment research and code to reproduce it

CLOOB training (JAX) and inference (JAX and PyTorch)

MAGMA - a GPT-style multimodal model that can understand any combination of images and language. NOTE: The freely available model from this repo is only a demo. For the latest multimodal and multil…

Locating and editing factual associations in GPT (NeurIPS 2022)

Implementation of LogAvgExp for Pytorch

GPU implementation of a fast generalized ANS (asymmetric numeral system) entropy encoder and decoder, with extensions for lossless compression of numerical and other data types in HPC/ML applications.

An annotated implementation of the Transformer paper.

Annotated transformer blog

Implementation of Enformer, Deepmind's attention network for predicting gene expression, in Pytorch

v objective diffusion inference code for PyTorch.

Code and explanation for IEEE CoG paper "Predicting Human Card Selection in Magic: The Gathering with Contextual Preference Ranking"

State of the Art Magic: the Gathering Draft and DeckBuilder AI.

An efficient interactive zero-knowledge proof scheme based on GKR in terms of unlayered circuit.

A GKR-based zero-knowledge proof protocol for CNN model inference.