← Curriculum track ← Learn hub
Quanta GenAI Curriculum · LLMOps · Basic

LLMOps Basic — 153: simulate empty retrieval on `Benchmark scepticism rituals` — memo `300042 [153]`

Lesson 153: Benchmark scepticism rituals

Focus

Bias toward observable metrics: latency, cost, escalation rate. Token Benchmark scepticism rituals:153 keeps neighbouring lessons differentiable.

Key ideas

Deep dive notebook

Synthetic drill artefacts

Eval harness snippet

case_id: LO-26484
route: support_rag_v3
must_include_patterns:
  - "\[chunk_"
forbid_patterns:
  - "guaranteed SLA"
judge_profile: tempered_4

Practice

Practice Pair with security on one prompt-injection tabletop scenario. — 153 Bump 30.