Boss: place the vectors
You've watched the model place words. Could you place an unseen word where its meaning belongs?
The idea inside
Drag each mystery word onto the meaning map; the model reveals its true spot and scores how close you got.
After this lesson
You can place an unfamiliar word on a meaning map by reasoning from labelled axes and nearby words, and judge a placement by distance.
Where it leads
Words are vectors. Now: how does the model actually predict the next one?
Inside this lesson
That's the real lesson stage, paused. Claim your pass to operate it.
See how AI actually works, end to end.
This lesson is one stop on the full arc. Unlock all of it, and keep it for life.
What you get
- The 34-lesson main path, a finishable route from a word to agents
- Goal tracks for using AI at work and building AI features
- Boss labs that make you apply a whole act, not just recognize it
- Spaced recall that brings each idea back before you forget
- Course memory: every term defined, with links to where it first appears
- A shareable capability card when you finish the main path
- Lifetime access on every device, every future lesson included
Not videos to watch. You predict, operate the machine, then prove it. That is why it stays.
99 interactive lessons and challenges. No videos, no code.
Free launch pass: lifetime access, no card needed
New here? The first lessons are free to try. Start with lesson 0.1
What this lesson shows
Drag each mystery word onto the meaning map; the model reveals its true spot and scores how close you got.
The question it opens with
You've watched the model place words. Could you place an unseen word where its meaning belongs?
The walkthrough, in the lesson's own words
- Meaning is a position. You just placed words as vectors and scored by distance.
- Here's where the model keeps it. Closer to the model's spot is a better placement.
- Drag the mystery word to where its meaning sits, then lock it in.
- A hand-drawn map in the model's style; real maps are learned, see 2.5.
- Use the axes and the anchor words: where does its meaning fall?
- Right on the money. Your sense of its meaning matches the model's position.
- You placed each word by meaning, exactly what an embedding is.
- Meaning is a position; near positions mean similar things. Placing the words and scoring by distance is the whole of Act 1: a word is a vector (1.1), an embedding is its coordinates (1.2), similarity is distance (1.3).
- The anchors and labelled axes let you reason about an unseen word and place it, then distance scored how close you were. That is exactly how the model compares meanings: not by letters, but by how near two vectors sit. Every later idea, attention included, is built on this one.
- Spotify drops a song you never searched for into your mix, right next to ones you love. What did it just do?
- It placed the song by learned position. Your taste is a neighbourhood on a map like the one you just labeled, and recommendations are nearest-neighbour lookups in an embedding space.
Key takeaway
You placed words as vectors by meaning alone, reasoning from anchors and axes instead of letters.
What you can do after this lesson
You can place an unfamiliar word on a meaning map by reasoning from labelled axes and nearby words, and judge a placement by distance.
Where it leads: Words are vectors. Now: how does the model actually predict the next one?
This is the written summary. The lesson itself is interactive: you predict, drag and operate the mechanism above, and the reveal answers you.