New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
Why does the model keep misreading numbers out of a clean JSON block?
What you'll have at the end
The format that reads back correctly, chosen after testing the same data as a markdown table, a JSON block, and plain prose
You need
You have a script or service that hands an LLM a handful of your own records, several of them close in value, and asks it a question about them, and you haven't settled on how those records get written into the prompt.
Not covered
Choosing the format an LLM hands back to your own code, whether that's JSON mode, tool calling, or schema enforcement, is a separate decision made later and isn't covered here.
Leans on
Checked 15 Aug 2026
Part of the Context Engineering cookbook