Module 08 Knowledge Check
5 questions. Pass mark 4 out of 5. Answer every question before checking the answer key below, then retry after reading the feedback.
1. The most important boundary for conversational memory is…
- A. Token count
- B. Which user and session the memory belongs to
- C. The model
- D. Latency
2. Unbounded conversation history eventually causes…
- A. Better answers
- B. Context-window overflow, rising cost, and diluted attention on what matters
- C. Faster responses
- D. Nothing
3. Summarising older turns risks…
- A. Nothing
- B. Losing a specific detail the user expects the assistant to still know
- C. Higher latency only
- D. Format errors
4. Stored conversation memory containing personal data requires…
- A. No special handling
- B. Retention limits, access control, and a deletion path
- C. Only encryption
- D. Compression
5. Long-term memory differs from conversation history in that it…
- A. Is shorter
- B. Persists selected facts across sessions, which raises consent and correction questions
- C. Is faster
- D. Uses no storage
---
Answer Key and Explanations
Check these only after attempting every question.
1. B - Which user and session the memory belongs to
Leaking one user's memory into another's session is a severe and very achievable bug.
2. B - Context-window overflow, rising cost, and diluted attention on what matters
History must be summarised, windowed, or selectively retrieved as it grows.
3. B - Losing a specific detail the user expects the assistant to still know
Summarisation is lossy compression; decide deliberately what must be preserved verbatim.
4. B - Retention limits, access control, and a deletion path
Memory stores are user-data stores and must be governed as such.
5. B - Persists selected facts across sessions, which raises consent and correction questions
Remembering a user across sessions is a product and privacy decision, not just a technical one.
Practical Check
Apply this module to your own work: complete the module activity for *Memory and Session State Boundaries*, then write one sentence naming what your result shows and one naming what it does not.
Strong Answer Pattern
A strong answer names the task, the evidence used, the check performed, and the remaining limitation. It avoids "proved", "guaranteed", or "always" unless the evidence genuinely supports it.