AI It becomes easier to understand when we can see it
Minh Khoa
Author
When learning about AI, I realized a problem:
There are concepts whose definitions make sense when you read them, but it is still very hard to picture how they actually work.
How does tokenization turn a sentence into tokens? How does embedding represent meaning? What is attention “paying attention” to? And RAG what does it do before generating an answer?
So, I built Visual Knowledge — a visual space that helps everyone explore concepts AI through images and interaction.
Instead of only reading theory, you can observe each process unfold, change the input data, and see the results update right on the screen.
The first version currently has 7 topics:
Tokenization · Neural Network · Embeddings · Attention · Transformer · LLM · RAG
My goal is very simple: turn abstract concepts into the moment:
“Ah, so that’s how it works.”
Visual Knowledge is still being refined and will have many more topics in the near future.