Compacted Retrieval

A summary instead of raw chunks

Library › Context Problems

1. The Context Window

2. Window Occupancy

3. Growth per Turn

4. Remaining Space

5. Turn Budget

6. Retrieval Footprint

7. Retrieval per Turn

8. Search Rounds

9. Window Sizing

10. Truncation

11. Compaction

12. The Compaction Bill

13. Compacted Retrieval

14. The Compaction Threshold

15. Verbatim Tail

16. Stateless API

17. Retrieval Cost

18. Context Tax

19. Quadratic Cost

20. Cost Forecast

A retrieved document arrives as raw chunks, and most of what they hold is not what the question needed. Compacting them before they enter the window means the agent reads a summary of the document instead of the document. Only the retrieval is compressed. The system prompt, the question, and the answer are what they were.

⬇ Compacted Retrieval worksheet (1 page)

---

Next:
14. The Compaction Threshold