LLMOps Intermediate — 105: compress prompt surface on `Synthetic eval data ethics` — memo `555876 [105]`

Lesson 105: Synthetic eval data ethics

Focus

Document interfaces between retrieval, prompts, and policy engines. Token Synthetic eval data ethics:105 keeps neighbouring lessons differentiable.

Key ideas

• Thread: Synthetic eval data ethics · drill v5 · spin 300418.
• Habit: attach a trace_id to every completion you would paste into an ops dashboard.
• Guardrail: add one RACI bullet for prompt or index changes before tomorrow's standup.

Deep dive notebook

Synthetic drill artefacts

Ops review capsule

Pilot L-105 checkpoint
- Grounding accuracy Δ `0.75`
- Escalation rate Δ `0.053`
- Spend guardrail `$824/day`
Risk: Token burn spike
Owner: Applied

Practice

Practice List five adversarial prompts using your org's domain nouns. — 105 Bump 34.