Starred repositories
Measuring and evolving with the frontier of agent work
A Python port of Pi’s minimalist coding agent.
Production-grade client-side tracing, profiling, and analysis for complex software systems.
A curated list of resources for Japanese natural language processing (NLP): Python libraries, LLMs, dictionaries, corpora, and datasets. Includes Claude Code skills to search resources.
日本語LLMまとめ - Overview of Japanese LLMs
Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels
A library that generates an interactive radar, inspired by https://thoughtworks.com/radar/.
I replicated Ng's RYS method and found that duplicating 3 specific layers in Qwen2.5-32B boosts reasoning by 17% and duplicating layers 12-14 in Devstral-24B improves logical deduction from 0.22→0.…
Train the smallest LM you can that fits in 16MB. Best model wins!
A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.
AI agents running research on single-GPU nanochat training automatically
Podman: A tool for managing OCI containers and pods.
100M tokens. Infinite compute. Lowest val loss wins.
Resources for skilling up in AI alignment research engineering. Covers basics of deep learning, mechanistic interpretability, and RL.
Practical Clean Architecture backend example built with FastAPI. No stateful globals (DI), low coupling (DIP), tactical DDD, CQRS, proper UoW usage. REST API, per-route error handling, session-base…
An alignment auditing agent capable of quickly exploring alignment hypothesis
Interpretable Causal Diffusion Language Models
High-efficiency LLM inference engine in C++/CUDA. Run Llama 70B on RTX 3090.
A fast, free JS & TS playground — runs entirely in your browser with instant execution, npm imports, and code export
Multi-platform SDK for integrating GitHub Copilot Agent into apps and services
The most atomic way to train and inference a GPT in pure, dependency-free JavaScript. This repository covers the complete algorithm. Everything else is just efficiency. Inspired by Karpathy's micro…
DSPy: The framework for programming—not prompting—language models