ML Intuition Notebook
These are the notes I took in attempts to learn something
Attention. Again
There should be hundreds of articles written on this topic, but I need to come up with my own, with my understanding and insights that I've gained. I go back to this mechanism again and again, asking new questions, and here is a portion of answers to them.
June 2025
Shape of everything inside a transformer
The purpose of this article is to go through the whole GPT model and thoroughly think of any dimension and its purpose. To add some organisation to these writings, let's go one by one in a logical order.
- • Part 1. Tokens and embeddings
- • Part 2. Normalisation and Attention
- • Part 3. MLP layer and model outputs
May 2025
"Vibe coding" notes
Everyone wonders how the world is going to look like after emerging of LLMs.\nMy main goal was to check how good LLM can follow instructions, how well can make tech decisions, what are the main limitations and what is a human part in such type of coding.
May 2025
Everything about tokenization
- • Part 1. Intro
- • Part 2. Tokenizer types
April 2025