← LectureMoment
Transformers & Attention · Advanced
Transformers — Deep Dive
20 practice questions on Transformers & Attention. Every question is written from a specific moment in a real lecture, and after you answer it links to that exact timestamp so you can check it yourself.
What this pack asks
The free questions in this pack. Answer choices and explanations appear as you play.
Splitting text on whitespace makes "bear" and "bears" two distinct tokens. Why is this a problem for the learned representations?
Answer this one →Self-attention builds a token's representation as a weighted average over the sequence. Which statement correctly assigns the roles of queries, keys, and values?
Answer this one →
Written from lectures by Stanford Online. Not affiliated with or endorsed by any university, channel, creator or certification program.