← all repositories

amitshekhariitbhu/llm-internals

An educational blog and video series that teaches how Large Language Models work internally, from tokenization to attention to inference optimization.

1k stars Learning
llm-internals
Velocity · 7d
+18
★ / day
Trend
steady
star history

This repository provides structured educational content to help developers understand how LLMs work under the hood. It covers fundamental concepts including tokenization and Byte Pair Encoding, the mathematics behind attention mechanisms with Query, Key, and Value matrices, and practical topics like quantization and fine-tuning. Content is delivered through blog posts and accompanying YouTube videos.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.