Learning: Evaluation + Quality Assurance¶
How to Use This Page¶
- Build small eval sets first, then expand.
- Keep scoring deterministic where possible.
- Tie quality metrics to release decisions.
Lesson 1: Fundamentals¶
Learn case design, rubric scoring, and reproducible evaluation runs.
What You Learn¶
- How to define clear expected outcomes per case
- Why rubric criteria should be explicit and measurable
- How to separate correctness, latency, and safety metrics
Walkthrough¶
- Create an initial eval set with representative cases.
- Define rubric dimensions and scoring rules.
- Run baseline evaluation and store results.
Try It Yourself¶
- Add edge-case inputs and observe score shifts.
- Tighten one rubric criterion and compare pass rates.
- Ensure repeated runs produce stable scores.
Lesson 2: Intermediate Patterns¶
Learn baseline comparison and regression triage.
What You Learn¶
- How to compare candidate changes to fixed baselines
- Why delta analysis is better than raw score snapshots
- How failure categorization speeds remediation
Walkthrough¶
- Select a baseline run as reference.
- Evaluate a new prompt/model/version candidate.
- Classify regressions by failure type.
Try It Yourself¶
- Create one intentional regression and verify it is detected.
- Measure false alarm rate for your gate thresholds.
- Build a simple top-3 failure category report.
Lesson 3: Production Patterns¶
Learn release gating, adversarial evals, and drift monitoring.
What You Learn¶
- How to enforce hard quality thresholds before release
- Why adversarial subsets are required for safety-critical paths
- How judge/model drift can invalidate historical comparisons
Walkthrough¶
- Define gate thresholds for critical metrics.
- Add adversarial cases to eval suite.
- Monitor trend lines and drift indicators over time.
Try It Yourself¶
- Run gate checks on two candidate versions and choose one.
- Add a "block release" rule for severe safety regressions.
- Document a rollback trigger based on eval drift.
Code Examples¶
- Starter script:
docs/starter/stage08_eval_qa.py - Stage architecture notes:
docs/roadmap/stage-08-eval-design-and-regression-gates/architecture.md