Tiny optimized Stable-diffusion that can run on GPUs with just 1GB of VRAM. (Beta)

Tiny AutoEncoder for Stable Diffusion

Command-line tools for building high-performance VAEs, latent space Transcoders, and more. Based on Tiny AutoEncoder models.

Efficient Triton Kernels for LLM Training

Stanford NLP Python library for benchmarking the utility of LLM interpretability methods

Unified Concept Editing in Diffusion Models

DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

Janus-Series: Unified Multimodal Understanding and Generation Models

Fully open reproduction of DeepSeek-R1

Clean, minimal, accessible reproduction of DeepSeek R1-Zero

Medusa: Simple Framework for Accelerating LLM Generation with Multiple Decoding Heads

Stanford NLP Python library for understanding and improving PyTorch models via interventions

A resource repository for machine unlearning in large language models

Erasing concepts from neural representations with provable guarantees

Recipes to scale inference-time compute of open models

Minimal GPT (~350 lines with a simple task to test it)

Publication-ready visualization of 3D objects and point clouds in seconds.

A library for efficient patching and automatic circuit discovery.

Mass-editing thousands of facts into a transformer memory (ICLR 2023)

Locating and editing factual associations in GPT (NeurIPS 2022)

Erasing Concepts from Diffusion Models

LTXVideo Q8

Because we don't want a jupyter notebook mess...

Simple implementation of Gwern's AUNN proposal

D2 is a modern diagram scripting language that turns text to diagrams.