logoalt Hacker News

growingsweyesterday at 6:46 AM6 repliesview on HN

Great stuff! I wrote an interactive blogpost that walks through the code and visualizes it: https://growingswe.com/blog/microgpt


Replies

O4epegbyesterday at 8:07 PM

> By the end of training, the model produces names like "kamon", "karai", "anna", and "anton". None of them are copies from the dataset.

All 4 are in the dataset, btw

evntdrvnyesterday at 1:11 PM

You should totally submit that to HN as an article, if you haven't already.

show 1 reply
joenot443yesterday at 1:17 PM

This is awesome! Normally I'm pretty critical of LLM-assisted-blogging, but this one's a real winner.

spinningslateyesterday at 10:25 AM

That’s beautifully done, thanks for posting. As helpful again to an ML novice like me as Karpathy’s original.

hei-limayesterday at 11:17 AM

Great!

evntdrvnyesterday at 1:11 PM

really nice, thanks