AI Groundedness Tutorial: Citations, Attribution

TL;DRGroundedness means tying answers to sources — citations and attribution reduce hallucination.

An answer is grounded when it's supported by provided, citable sources rather than the model's memory. RAG enables this: feed documents and ask the model to answer only from them and cite which passage. You can then verify claims against sources. Groundedness + citations turn an LLM from a confident guesser into an auditable assistant — essential for research, legal, medical, and enterprise use.

Key points

Common mistakes

Try it: Explain how citations make an LLM answer auditable.

Example code

<!doctype html><html><head><meta charset="utf-8"></head>
<body style="background:#06040d;color:#e6e0ff;font-family:monospace;padding:20px"><pre>prompt: "answer ONLY from these docs, cite the passage"
answer: "X is true [doc 2, p3]"
→ verify against the source</pre></body></html>
Open the interactive lesson →
Hallucinations · Why Models Invent Red Teaming · Finding Failures on Purpose