the.ai

Context / Behaviour

verified

Lost in the Middle

Put the answer at the start of a long context and the model finds it. Put it at the end and the model finds it. Put it in the middle and accuracy drops sharply — sometimes below what the model achieves with no context at all. The window is not uniform, and where you place information matters as much as whether you include it.

Viz primitive · budget-splitmiddle-positions = 20

middle-positions holds 50% of the budget; rest holds the remaining 50%.

Context positions in the degraded middle, against the ends where retrieval still works, in positions. Drag the context length up to watch the middle take almost all of it — the region that grows fastest is the one the model reads worst.

20

Reviewed by opendroid · 2026-08-18

  • arXiv:2307.03172 — Lost in the Middle: How Language Models Use Long Contexts
  • arXiv:2406.17419 — Leave No Document Behind: Benchmarking Long-Context LLMs with Extended Multi-Doc QA