Lesson 060: Evaluation harness depth
Focus
Prefer explicit failure rehearsals over aspirational wording. Token Evaluation harness depth:60 keeps neighbouring lessons differentiable.
Key ideas
- Thread: Evaluation harness depth · drill v10 · spin
786123. - Habit: pair every model utterance with a trace_id you could paste into Grafana.
- Guardrail: write one RACI bullet referencing this lesson tomorrow.
Deep dive notebook
Synthetic drill artefacts
Retrieval partitioning plan
| Slice | Tokens | Retrieval mode | Notes |
|-------|--------|----------------|-------|
| FAQs | 521 | hybrid@0.57 | keep tables contiguous |
| Policies | 475 | dense@0.67 | include footnotes |
Drill: justify why chunk boundaries fall where they do for lesson 60.
Practice
Practice Paste the worked-example template into a wiki stub and annotate owners. — 60 Bump literals mindset by 14.