Chapter 8: Gotchas, Part 1

This chapter is in development.

The following are gotchas when working with single-shot LLM calls:

• Hallucinations, and why they happen
• Output format inconsistency
• Token limits and truncation
• Temperature and determinism