Loading...
Loading...
Browse 3 real-world technical and behavioral interview questions about Groundedness. Review scenarios, edge cases, and architectural best practices.
Walk three gates in order for that one request — was the supporting passage retrieved, did it survive into the rendered prompt, does the answer contradict it. Each gate blames a different stage, which is why the diagnosis depends on having logged both the retrieved chunk ids and the prompt as sent.
Report two rates separately: ungroundedness, meaning claims the provided sources do not support, and factual error, meaning claims wrong about the world. Measure per claim, publish the abstention rate beside them, and name the query set that produced the numbers. Use this hallucination answer to show the decision, trade-off, and evidence rather than a memorised definition.
Recall at k is the primary one, since a passage that never reaches the prompt cannot be used; precision at k measures wasted window; MRR or NDCG measure ordering. Read them beside the groundedness of the answer, because high recall with bad answers points at generation or context ordering.