A collection of lightweight interpretability scripts to understand how LLMs think

The filesystem for agents.

Staging repo for development of native port of TypeScript

Training VLM agents with multi-turn reinforcement learning

A simple, performant and scalable Jax LLM!

Run, manage, and scale AI workloads on any AI infrastructure. Use one system to access & manage all AI compute (Kubernetes, 20+ clouds, or on-prem).

Based on Nano-vLLM, a simple replication of vLLM with self-contained paged attention and flash attention implementation

JAX bindings for the flash-attention3 kernels

Official repository of DARE: dLLM Alignment and Reinforcement Executor

Fast and memory-efficient exact attention

Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen2.5, Qwen3, Llama, and more!

implementations and experimentation on mHC by deepseek - https://arxiv.org/abs/2512.24880

Tile-Based Runtime for Ultra-Low-Latency LLM Inference

⚙️🦀 Build modular and scalable LLM Applications in Rust

A collection of various llm pruning implementations, training code for GPUs & TPUs, and evaluation script.