vgel

llamafile

Distribute and run LLMs with a single file.

dspy

DSPy: The framework for programming—not prompting—foundation models

litgpt

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

InternLM

Official release of InternLM2.5 7B base and chat models. 1M context support

GPU-Puzzles

Solve puzzles. Learn CUDA.

tree-of-thought-llm

[NeurIPS 2023] Tree of Thoughts: Deliberate Problem Solving with Large Language Models

transformer-debugger

Eureka

Official Repository for "Eureka: Human-Level Reward Design via Coding Large Language Models" (ICLR 2024)

TransformerLens

A library for mechanistic interpretability of GPT-style language models

main-thread-scheduling

Fast and consistently responsive apps using a single function call

vlite

fast vector database made in numpy

godot-dodo

Finetuning large language models for GDScript generation.

tracr

maccarone

AI-managed code blocks in Python ⏪⏩

bash-stack

modern web framework in bash

SAELens

Training Sparse Autoencoders on Language Models

sae

Sparse autoencoders

llm_steer

Steer LLM outputs towards a certain topic/subject and enhance response capabilities using activation engineering by adding steering vectors

arckit

Tools for working with the Abstraction & Reasoning Corpus

arxiv-browse

Flask app for article abstract and listing pages

train-with-fsdp

simple_transformer

Simple Transformer in Jax

1L-Sparse-Autoencoder

raft

RAFT, or Retrieval-Augmented Fine-Tuning, is a method comprising of a fine-tuning and a RAG-based retrieval phase. It is particularly suited for the creation of agents that realistically emulate a specific human target.

onnx_embedding_models

utilities for loading and running text embeddings with onnx

sparse-dictionary-learning

An Open Source Implementation of Anthropic's Paper: "Towards Monosemanticity: Decomposing Language Models with Dictionary Learning"

twinkgrad

handmade-transformer

making GPT2 transformer weights by hand

corpus_toolkit

Python toolkit for textual analysis and visualization. Features include corpus tokenization, lexical diversity calculation, vocabulary growth prediction, entropy measures, and Zipf/Heaps law visualizations. Designed for computational linguistics research.

bsky-whatsmid