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

An unofficial Google Nearby Share app for macOS

RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference,…

Multi-LoRA inference server that scales to 1000s of fine-tuned LLMs

Mamba SSM architecture

AGI agent simulator

(Python, R, C) Fast approximations for the CDF of multivariate normal distributions

Data generator based on copulas

Reference implementation for DPO (Direct Preference Optimization)

Code accompanying our paper "Feature Learning in Infinite-Width Neural Networks" (https://arxiv.org/abs/2011.14522)

PyTorch-based library for various kinds of representational-similarity analysis

This repository includes code to reproduce the tables in "Loss Landscapes are All You Need: Neural Network Generalization Can Be Explained Without the Implicit Bias of Gradient Descent"

Code for the paper "How to catch an AI liar: Lie detection in black-box LLMs by asking unrelated questions"

A Python package for intrinsic dimension estimation

code for "Erasure of Unaligned Attributes from Neural Representations"

CEBaB: Estimating the Causal Effects of Real-World Concepts on NLP Model Behavior

Erasing concepts from neural representations with provable guarantees

Brave's Rust-based adblock engine

OpenAssistant is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so.

Simple, Elegant, Typed Argument Parsing with argparse

NL-Augmenter 🦎 → 🐍 A Collaborative Repository of Natural Language Transformations

Keeping language models honest by directly eliciting knowledge encoded in their activations.

Newton and Quasi-Newton optimization with PyTorch