token by token is a technical blog where we post weekly about topics we learn. All posts discuss ML/AI topics.
"If you can't explain something to a first year student, then you haven't really understood."
We decided to start this project to encourage each other to learn and dive deep into topics in the ML field. By publishing our work, we are held accountable to present concepts in a format that could be easily understood by other learners.
It is called token by token because we hope to break down and present the information in bits that are digestible, similar to the way we learned them.
Writing is how we learn. If we let AI write for us, we skip the part where understanding actually happens. The struggle of putting a concept into our own words is the whole point.
Curiosity. Whatever catches our interest that week: a paper we stumbled on, a concept that confused us, or something we saw and wanted to understand deeper.
We're open to guest posts. If you have a topic you'd like to write about, pitch it to us via GitHub. We'll review it together and if it fits, you're in.
Because that's how language models generate text, and that's how learning works, one piece at a time.
Spot an error? Want to discuss something? Open an issue on GitHub.