Cuda kernels for leveraging LLM sparsity to improve throughput and decrease the memory requirements during inference and training.

An agent-managed museum exhibit, built in Rust with Gajae-Code / LazyCodex — developed and maintained with no human intervention.

A Super AI Lab with massive AI Doctors as Assistants. Best IDE for Research via AI Power.

Code for the NPJ AI paper "How Large Language Models Encode Theory-of-Mind: A Study on Sparse Parameter Patterns"

A PyTorch native platform for training generative AI models

Train transformer language models with reinforcement learning.

A Heterogeneous Benchmark for Information Retrieval. Easy to use, evaluate your models across 15+ diverse IR datasets.

[EMNLP'25] Code for the paper "DEL-ToM: Inference-Time Scaling for Theory-of-Mind Reasoning via Dynamic Epistemic Logic"

A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.

PubMedQA: A Dataset for Biomedical Research Question Answering

Vortex: Programmable Sparse Attention for Agents as Algorithm Designers

🤖 MLE-Agent: Your intelligent companion for seamless AI engineering and research. 🔍 Integrate with arxiv and paper with code to provide better code/research plans 🧰 OpenAI, Anthropic, Gemini, Ollam…

Optimize prompts, code, and more with AI-powered Reflective Optimization

Open-source implementation of AlphaEvolve

[ICML 2026] Decoding Tree Sketching (DTS): a training-free & model agonistic & plug-in framework for LLM parallel reasoning.

This guide provides a setup for running `vLLM` with `Triton` on NVIDIA GH200.

RapidIn: Scalable Influence Estimation for Large Language Models (LLMs). The implementation for paper "Token-wise Influential Training Data Retrieval for Large Language Models" (Accepted on ACL 2024).

A simple and effective LLM pruning approach.

The fastest, litest AI Gateway. Rust core with Python SDK. Call 100+ LLM APIs in OpenAI (or native) format with cost tracking, guardrails, load balancing, and logging [Bedrock, Azure, OpenAI, Anthr…

H-Net: Hierarchical Network with Dynamic Chunking

A lightweight python-only library for reading and writing SMILES strings

Code repository for ICLR 2025 paper "LeanQuant: Accurate and Scalable Large Language Model Quantization with Loss-error-aware Grid"

Code Repo for EMNLP paper: Do LLMs Know to Respect Copyright Notice

A game theoretic approach to explain the output of any machine learning model.

[NeurIPS 2024] Simple and Effective Masked Diffusion Language Model

Production-tested AI infrastructure tools for efficient AGI development and community-driven innovation

A novel medical large language model family with 13/70B parameters, which have SOTA performances on various medical tasks

FlashMLA: Efficient Multi-head Latent Attention Kernels

MoBA: Mixture of Block Attention for Long-Context LLMs