Advanced NLP Interview Questions, issue 14, Dec 20, 2025

The Tokenization Brittleness Trap

AI Engineer interview at Meta, and the interviewer asks:

We deployed a Llama-3 based app. We removed a single whitespace in the prompt template, and our benchmark accuracy tanked by 12%. Why is the model so brittle to a simple format change, and why didn’t instruction tuning prevent this?

Don’t say: The model is confused by the bad grammar.

Why deleting a single whitespace can cause a 12% accuracy drop and why instruction tuning doesn’t actually make LLMs robust.

The full answer, with the mechanism and the arithmetic, is for paid subscribers on Substack.

Read it on Substack

Get the next one

Free on Substack. Unsubscribe in one click.