The model can only read what fits the window. Feed it the chunks that answer this:

User query

How do we verify the Stripe webhook signature?

12 chunks were retrieved. Pick the set with the most relevance that still fits the 1,238-token window.

Why it matters · in an interview

Context packing is the RAG question every AI-engineering interview asks — “what goes in the window?” The answer is relevance per token, not raw relevance.

Relevance packed

Context window

0 / 1,238

Tap sources to pack. The trick is relevance per token — watch the efficiency bars, not just the relevance number.

You don't need a perfect score — get as close to the optimum as you can. Each play teaches the pattern.

Context Budget — practice the decision, scored against the optimum

What does the Context Budget puzzle teach?

Packing the most relevant chunks into a fixed token window is knapsack — greedy by relevance loses.

Why do interviews ask about this?

Context packing is the RAG question every AI-engineering interview asks — “what goes in the window?” The answer is relevance per token, not raw relevance.

How is it scored?

The engine computes the provably-optimal answer for the exact instance you played (dynamic programming, shortest-path, or exhaustive search — never a heuristic) and scores your attempt as a percentage of it. No login needed to see your score.

← Arena

Practice0:00

Context Budget · D4

The model can only read what fits the window. Feed it the chunks that answer this:

User query

How do we verify the Stripe webhook signature?

12 chunks were retrieved. Pick the set with the most relevance that still fits the 1,238-token window.

Why it matters · in an interview

Context packing is the RAG question every AI-engineering interview asks — “what goes in the window?” The answer is relevance per token, not raw relevance.

Relevance packed

Context window

0 / 1,238

Tap sources to pack. The trick is relevance per token — watch the efficiency bars, not just the relevance number.

You don't need a perfect score — get as close to the optimum as you can. Each play teaches the pattern.