Free weekly email
Interview Question of the Week
One AI engineer interview question in your inbox each week. Answer it out loud first; the next email brings a model answer, the likely follow-up and the mistake to avoid, plus a new question.
Sample question
Why do models hallucinate, and why don't they just say "I don't know"?
What they’re checking: Whether you understand the mechanism and have practical mitigations instead of hoping a bigger model fixes it.
Say your answer out loud in under a minute before you open the model answer.
Show the model answer Hide the model answer
A model generates the most plausible continuation from patterns in its weights, and nothing in that process checks the result against the truth. Common facts are stored well; rare ones are blurry or missing, but a confident, fluent answer is still a high-probability continuation, so you get a real-looking citation for a paper that doesn't exist. As for "I don't know," two reasons. The model often can't tell what it doesn't know, because it has no reliable internal signal tied to correctness. And training and evaluation have tended to reward a guess over an abstention, the way a test with no penalty for wrong answers rewards always answering. In an application I reduce it by grounding answers in retrieved sources, telling the model it may say the answer isn't there, and checking important claims in code or with a human.
- Likely follow-up:
- How would you detect hallucinations in production?
- Avoid:
- Saying RAG or a newer model eliminates hallucinations.
More questions like this: 50 LLM Interview Questions and Answers →
You’re in
Check your inbox: this week’s question is on its way, and its model answer comes with next week’s email.
Get one question a week
Free. One email a week. Unsubscribe with one click.