Lesson 087: Embedding probes on tiny corpora
Focus
Bias toward observable metrics, not model marketing. Token Embedding probes on tiny corpora:87 keeps neighbouring lessons differentiable.
Key ideas
- Thread: Embedding probes on tiny corpora · drill v7 · spin
249919. - 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
Logging field note
| Field | Retention |
|-------|-----------|
| trace_id | 10 days |
| prompt_hash_sha256 | permanent |
| completion_excerpt_redacted | 24h hot, then cold vault |
Heuristic `pii-mask-v2` tags sensitive spans before persistence.
Practice
Practice List five adversarial prompts unique to your org’s nouns. — 87 Bump literals mindset by 11.