jinensetpal

flash-attention

Fast and memory-efficient exact attention

my-chess

My on-going attempt at making a chess engine. V1 is playable in WASM via the link below. V2 is a refactor. V3 is in progress.

a-stack

creating my own web dev stack lol

vsynth

vector program synthesis

dlbook_notation

LaTeX files for the Deep Learning book notation

Liger-Kernel

Efficient Triton Kernels for LLM Training

nn

open_lm

A repository for research on medium sized language models.

awesome-world-models-manipulation

Awesome world models for manipulation

sae-auto-interp

NeMo

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

sae

Sparse autoencoders

ladybird

Truly independent web browser

nanoGPT

The simplest, fastest repository for training/finetuning medium-sized GPTs.

statistically-correct-python

An approximate Python interpreter

pykan

Kolmogorov Arnold Networks

unicodeit

Converts LaTeX tags to unicode: \mathcal{H} → ℋ. Available on the web or as Automator script for the Mac.

wayvnc

A VNC server for wlroots based Wayland compositors

penzai

A JAX research toolkit for building, editing, and visualizing neural networks.

othello_world

Emergent world representations: Exploring a sequence model trained on a synthetic task

chess_llm_interpretability

Visualizing the internal board state of a GPT trained on chess PGN strings, and performing interventions on its internal board state and representation of player Elo.

simple-evals

ollama

Get up and running with Llama 3.2, Mistral, Gemma 2, and other large language models.

evolutionary-model-merge

Official repository of Evolutionary Optimization of Model Merging Recipes

llm.c

LLM training in simple, raw C/CUDA

einops

Flexible and powerful tensor operations for readable and reliable code (for pytorch, jax, TF and others)

SAELens

Training Sparse Autoencoders on Language Models

schedule_free

Schedule-Free Optimization in PyTorch

guarantees-based-mechanistic-interpretability

Automatic-Circuit-Discovery