A community-tool-library lead wants this chapter’s risk item to teach embedding stores and a tuning job so invented loan rules stop. What is the correct Chapter 3 stance?
Select an answer to reveal the explanation.
Short Explanation
Chapter 3 names five in-chapter mitigations and only points ahead to RAG and fine-tuning. Embedding stores and tuning jobs are not the exam item here.
Full Explanation
Community tool-library leads sometimes want to jump straight to embedding stores and fine-tuning jobs. Within Chapter 3, the correct stance is to recognize RAG and fine-tuning as Chapter 4 complementary techniques, not procedures examined in this chapter’s risk-mitigation item. Chapter 3 focuses on context, decomposition, clear formats, model fit, and cross-model comparison.
Treating RAG chunking sizes as an examinable Chapter 3 procedure fails by syllabus scope: chunking belongs with retrieval architecture later. Requiring a full fine-tuning job before any mitigation is valid fails because in-chapter mitigations stand on their own without a tuning project. Replacing all five in-chapter mitigations with embedding-store setup fails by collapsing the chapter’s named set into a forward reference.
Exam caveat: “not examined here” does not mean RAG and fine-tuning are useless—only that this chapter’s exam items test the five in-chapter mitigations and the forward pointer, not embedding-store procedures.
Operational check: when training staff on Chapter 3, teach the five mitigations with civic examples first; label RAG and fine-tuning as upcoming complementary options; refuse to mark Chapter 3 quizzes wrong for not detailing chunk sizes or tuning hyperparameters.