World History Maps vs Learning Transformer

This interactive atlas of the ancient western world gives you a close up view of the area from the Mediterranean to India from 3000 BCE to 500 BCE. The map allows zooming and panning. Pop-up information is available for each country shown. Each map shows all countries of the world as they were at the beginning of each year. View cities, battles, campaigns, ancient roads, and more. Links to Wikipedia articles provide more detail. Zoom in without losing any detail.

A Transformer, built by hand and shown one step at a time. Twenty-three screens follow "Attention Is All You Need" (Vaswani et al., 2017) in the paper's own order: a word becoming numbers, the three vectors, the scores, the softmax, four heads merged into one, the mask, the second attention, and the loop that writes one word at a time. What this is not: a trained model. Every weight here was set by hand, so that each one can be read in a single line. Nine numbers stand for a word instead of 512. One encoder layer and one decoder layer instead of six of each. Four heads, four words of vocabulary, one short sentence. What it is: the paper's arithmetic at a size you can check by hand. Every number on every screen is computed while you look at it — no videos, no quizzes, no account, no network. Steps 1 to 10 are free. Steps 11 to 23 are one purchase, made once and kept.