Pretraining objectives, scaling, and checkpoints
Use pretraining objectives, scaling, and checkpoints to move the foundation models production brief toward a defensible release.
A private support assistant must answer from evidence and resist manipulation.
This lesson isolates pretraining objectives, scaling, and checkpoints as one decision inside that system. The people affected are product users and operators; the learning data must carry event time, availability time, ownership, and version; and the operating envelope is a declared latency, cost, and operator-capacity budget.
Choose whether and how to use decontamination at a declared prediction cutoff.
Measure decision utility alongside calibration, slice reliability, and system latency—not model score alone.
A private support assistant must answer from evidence and resist manipulation. An unsafe release must degrade to a named baseline or the last known-good version.
Build the mental model before the machinery.
The core move is to treat pretraining objectives, scaling, and checkpoints as a contract between data, a computation, and an action. Release model, tokenizer, retrieval index, prompt, policy, datasets, eval suite, and runtime config as one bundle. The implementation becomes easier to debug once you can state which inputs exist, which state is learned, what output means, and what must remain invariant after serialization.
decontamination
Define it in a hand-checkable form and name the prediction-time inputs.
tokenizer fertility
Connect it to the production metric and identify what it cannot guarantee.
SFT
Stress it with a slice, a temporal boundary, and a failure-safe alternative.
LoRA
Stress it with a slice, a temporal boundary, and a failure-safe alternative.
Lessons 1, 2 in this course.
Name every symbol. Check every shape.
Autoregressive next-token loss is the central invariant for this lesson. The formula is useful only when its inputs match the production cutoff and its output maps to an action.
Autoregressive next-token loss
| Symbol | Meaning / shape / unit |
|---|---|
T | sequence length |
x_t | target token |
θ | model parameters |
Open derivation and numerical substitution
Start from the production quantity being optimized, substitute the observed values with their declared units, then isolate the model-controlled term. Preserve shape annotations at each step so broadcasting or aggregation cannot silently change the result.
- Write the named inputs: T, x_t, θ.
- Substitute one small, hand-checkable batch before vectorizing.
- Calculate an independent reference value and compare within a declared tolerance.
# equation → code contract
inputs = validate_shapes_and_units(batch)
value = compute_c15(inputs)
assert is_finite(value)Calculate it small. Shape it realistically. Break it on purpose.
A result you can reproduce on paper
Train a tiny character model on repeated text and inspect how context changes the next-symbol distribution.
- Write every input and unit.
- Substitute values into the autoregressive next-token loss equation above.
- Compare the result to one simple baseline and explain the direction of the difference.
The same reasoning under real constraints
Compare prompt-only, retrieval, and parameter-efficient adaptation on versioned multilingual support data.
The production record includes the data snapshot, transformation state, artifact identity, cutoff, score, decision, and the version of the policy that consumed it.
The attractive result you should reject
Evaluation contamination and feedback only from successful chats hide refusals, unsupported languages, and hallucinated citations.
Diagnostic: replay the smallest failing slice from immutable inputs, then compare each boundary rather than retuning the model.
Change one assumption and make the tradeoff visible.
This lab runs predefined TypeScript only. It never executes learner code. Use the slider, numeric input, reset, live text, or table—the computation is the same.
Retrieval depth lab
Change top-k and estimate evidence recall, context cost, and distractor risk.
Assumption: Fixed retrieval curve; each passage contributes 180 tokens.
Open nonvisual data table
| Item | Computed state | Interpretation |
|---|---|---|
| passage 1 | 88.0% | 180 tokens |
| passage 2 | 79.0% | 192 tokens |
| passage 3 | 70.0% | 204 tokens |
| passage 4 | 61.0% | 216 tokens |
| passage 5 | 52.0% | 228 tokens |
Trace the complete operating path.
- 01
Validate and version decontamination.
- 02
Compute pretraining objectives, scaling, and checkpoints from prediction-time-safe inputs.
- 03
Persist model, feature, and configuration identities together.
- 04
Serve or materialize behind explicit a declared latency, cost, and operator-capacity budget.
- 05
Join telemetry to mature outcomes and retain a rollback path.
Observability
Join service health, input quality, prediction distributions, slice behavior, and mature outcomes by exact version.
Cost
Measure storage, preprocessing, compute, queueing, and human review under a representative arrival pattern.
Failure modes
Evaluation contamination and feedback only from successful chats hide refusals, unsupported languages, and hallucinated citations. Add a detector, owner, mitigation, and stop condition for this class of failure.
Alternatives
Compare a rule, a simpler statistical baseline, and a different system boundary before adding model complexity.
Explain the contract, not just the vocabulary.
Domain adaptation release package
Compare prompt, retrieval, and optional PEFT approaches under one provenance and evaluation contract.
- Approach decision
- Provenance ledger
- Slice evaluation
- Safety probes/model card
Read primary material with a purpose.
Return to the opening failure.
Release model, tokenizer, retrieval index, prompt, policy, datasets, eval suite, and runtime config as one bundle.
For this lesson, the release evidence is a hand-checked formal result, deterministic simulation output, a ≥80% checkpoint, the production rubric, and a named fallback. The course resolves when the system can produce a grounded system design spanning adaptation, rag, evaluation, and safety.