Learning

To turn graduate-level NLP and AI safety lectures into runnable Jupyter notebooks that anyone can break, watermark, or align.

A living archive of extracted system prompts from Anthropic, OpenAI, Google, and xAI that exposes the hidden instructions shaping tone, tools, and guardrails.

A readable reference for how classic machine learning actually works under the hood, from backprop to genetic algorithms.

An open-source Chinese textbook that binds math, code, and critical thinking into a single, living resource for deep learning.

It collects copy-paste role-play prompts that cast ChatGPT as a Linux terminal, proofreader, or interviewer for Chinese-speaking users.

It teaches how LLMs work by implementing tokenization, attention, pretraining, and finetuning in pure PyTorch, one notebook at a time.

Because reading about agents is easier than wiring them up correctly.

A massive collection of extracted system prompts and internal models from more than two dozen commercial AI coding agents and assistants.

Open-sources the exact LLM prompts and agent skills that researchers at MSRA, ByteDance Seed, and top Chinese universities use to write and polish papers.

A systematic Chinese tutorial for developers who want to stop treating LLMs as black boxes and hand-build a 215-million-parameter model from the ground up.

Curated tutorials and tool reviews covering vibe coding, DeepSeek, Cursor, and the rest of the generative-AI menagerie, maintained as a free, open-source knowledge base.

A single engineer translated the subtitles and wrote exhaustive Chinese notes for Andrew Ng’s 2014 ML course, then open-sourced them after tens of thousands of downloads.

Because 'just ask nicely' is not a production strategy, this is a curated, living reference for systematic LLM prompting.

A maintainer cataloged every Chinese NLP repo they touched into a single, obsessively categorized list so others wouldn’t have to hunt.

Official Python notebooks and guides for common OpenAI API tasks.

This repo exists because the LLM learning curve is a scattered mess of blog posts, so it organizes the field into three distinct tracks—fundamentals, scientist, and engineer—each paired with runnable Colab notebooks.

Jupyter notebooks that turn the fastai course and O'Reilly book into executable lessons.

A curated directory of software, plugins, and frameworks that integrate with the DeepSeek API, maintained by DeepSeek itself.

Because 'make it pretty' is not a prompt, and this repo treats image generation like a production pipeline, not a toy.

To teach agentic design patterns through videos, notebooks, and documentation auto-translated into more than 50 languages.