Data & Machine Learning
NLP & LLMs
8 practice questions. Free questions open a full answer guide; the rest unlock with Pro.
- For a new feature that answers user questions from your company's internal documentation, how would you decide between prompting a base LLM, retrieval-augmented generation, and fine-tuning?
- Your team shipped a RAG-based support assistant and it sometimes states things that aren't in the retrieved documents. How would you reduce these hallucinations, and how would you measure faithfulness so you know whether a change actually helped?Go Pro
- How would you evaluate the quality of an LLM feature that summarises documents or answers questions, where there's no single correct output to compare against?Go Pro
- A product team wants an LLM assistant that answers questions over the company's internal documentation, and they're debating whether to fine-tune a model or build a retrieval-augmented (RAG) system. How do you reason through that choice, and when would you combine them?Go Pro
- How does retrieval-augmented generation (RAG) reduce hallucinations in an LLM application, and where can it still go wrong?Go Pro
- What do the temperature and top-p settings control when generating text from an LLM, and how would you set them for a task that needs reliable, consistent output?Go Pro
- What is a token in an LLM, and why does it matter that the model works in tokens rather than characters or words?Go Pro
- Why do modern language models split text into subword tokens instead of whole words or characters, and how does that choice affect a feature you'd build on top?Go Pro
Want questions matched to your role? Paste a job title, job description, or CV for a personalized set, or go Pro to unlock the full bank.