Week 09 · overview
Week 9: Generative Networks, Attention, and Transformers
Generative networks produce sequences one step at a time. Attention changes the mechanism: each query can weight relevant context positions directly, while positional information preserves order for transformer processing.
| Day | Role | Evidence |
|---|---|---|
| 1 | Lesson 1: Generative networks | Trace prompting, next-token sampling, and temperature |
| 2 | Lesson 2: Attention and transformers | Explain queries, keys, weights, position, and parallel processing |
| 3 | Investigation | Calculate a complete single-head attention example |
| 4 | Lab | Compare attention allocations under two queries |
| 5 | Assessment + Reflection | Repair an attention-explains-the-model claim |
Your Attention Allocation Record preserves queries, keys, dot products, softmax terms, normalized weights, comparisons, claim repair, and limitation. The Lab uses invented two-dimensional vectors and contacts no model.
The complete pinned Microsoft source and credits remain in upstream/.
Review the source record.
Chapter question
How does Generative Networks, Attention, and Transformers change what evidence a student should trust?
What you will understand
You will connect the named mechanism to the evidence it can and cannot support.
What you will do
You will inspect supplied cases, trace the important calculation, and test one bounded claim.
What you will produce
You will produce the week evidence record named in the Lab or reflection.
How the week connects
This week carries the previous AI mechanism forward into a stricter evidence check.
Key vocabulary
mechanism; evidence; boundary; test; claim
Approximate time
Five class meetings at about 210 instructional minutes total.
Final evidence required
Submit the completed week record with at least two exact labels from the artifact.