Describe a Variant by Its Changed Contract

Use a five-field architecture card to identify the changed component, invariant core, analytical consequence, evidence boundary, and later curriculum owner.

Compare architectures by their contracts, not only by their names. A useful architecture card states the changed component, the invariant core, an analytical consequence, the boundary of the empirical evidence, and the later subject that owns the missing depth.

Separate Five Questions

For any claimed variant, ask:

  1. What represents the input?
  2. What mixes information across positions?
  3. What transforms features at one position?
  4. What objective and data train the system?
  5. What work happens outside the model?

The first three describe architecture. The fourth describes training. The fifth may include retrieval, tools, serving, or product policy. Changing two of these at once prevents a clean architecture-only comparison.

Complete One Card

Consider causal local-window attention:

FieldStatement
Changed componentEach query sees only recent keys unless a declared global edge is added.
Invariant coreQ/K scores, normalized value mixing, output projection, residual path, and MLP remain.
Analytical consequenceThe number of allowed attention edges falls.
Evidence boundaryQuality depends on the sparse graph, model, data, task, and training setting tested.
Later ownerKernel and serving details belong to AI Engineering.

The card prevents two mistakes: treating every variant as a new foundation and treating one successful experiment as a universal result.

Q1. Classify a claimed change

A report keeps the decoder architecture fixed but adds retrieval before each prompt. Which contract changed: model architecture or the surrounding system?

Choose one

Select one choice, then check.

Hint
Ask whether any layer inside the decoder was replaced.
Solution
The surrounding system changed. Retrieval can alter the model input without changing the decoder's internal architecture.
Not attempted
Review

Not marked done.

Use the Card before Judging Results

An analytical consequence can often be derived from shapes or graph structure. An empirical consequence must be measured. Keep these two kinds of statement separate throughout this chapter.

An implementation can also change while the mathematical architecture remains fixed. FlashAttention, for example, reorders exact attention computation to reduce memory traffic; it does not by itself change which token pairs are visible. Its measured speed and memory results still depend on the stated kernels, hardware, shapes, and precision. Kernel implementation belongs to AI Engineering, while this subject preserves the attention contract that the kernel must calculate.

Pause and reflect

In your own words, note what you understood, what remains unclear, or what you want to revisit. The note stays with this lesson.

0 of 1 exercises marked done

Review

Not marked done.

LLM PrimerDescribe a Variant by Its Changed Contracthttps://llmprimer.com/transformers/variants-scaling-and-boundaries/describe-a-variant-by-its-changed-contract© 2026 LLM Primer