How LLMs and coding tools work8 min read

Lost in the Middle: How Language Models Use Long Contexts

Nelson F. Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, Percy Liang · 2023 · Transactions of the Association for Computational Linguistics 12 (2024), 157–173

Models find information at the start and end of a long input far more reliably than information in the middle — and a bigger context window does not fix it.

Read the paper ↗doi:10.1162/tacl_a_00638 ↗

The short version

→Move the one relevant document around inside a long prompt, change nothing else, and accuracy swings dramatically.
→The shape is a U-curve: best when the needed information is at the beginning or the end, worst in the middle.
→This held on explicitly long-context models too — having room for the tokens is not the same as using them.
→A context window is a capacity, not a promise. Where you put things inside it is a design decision.

Subscription

The Library

Everything running a software business asks of you and none of it is code: onboarding that activates, churn you can actually see coming, marketing with no budget, SEO that finds buyers, and the AI-plus-Ahrefs workflow for writing the blog that feeds it. One new playbook a week, plus the whole archive from day one.

$9/ month · cancel any time

✓A new playbook every week — onboarding, churn, growth, SEO, content
✓The full back catalogue the moment you join, not drip-fed
✓Vendor explainers: Stripe, Supabase, Resend, Vercel, the AI APIs
✓The research shelf — real papers, decoded, on learning and on LLMs
✓One email when a new one lands, off by default in one click
✓Cancel any time; you keep reading to the end of the month

first — it takes ten seconds.

Billed monthly through Whop's secure checkout · cancel from Settings any time. See all plans →