What is Hallucination?
Also called: confabulation · AI making things up
A hallucination is confident, fluent output that is false — invented facts, citations, quotes, numbers or code APIs that don't exist.
Language models predict plausible text; they don't look things up unless a tool lets them. When the most plausible continuation is a citation, the model produces one whether or not it exists. That's a hallucination — and it looks exactly like a correct answer.
You can reduce it but not eliminate it: give the model the source material and tell it to use only that, ask it to mark uncertainty, require quotes with locations, and verify anything you'll rely on — especially references, statistics, legal and medical claims.
Answer using only the text below. If the answer isn't in the text, say "not in the source". Quote the sentence you relied on. Text: [paste]
- Paste sources; never ask the model to find them.
- Ask "which parts of this are you unsure about?" — it will usually tell you.
- Treat every number and citation as unverified until you check it.
Use it right now
Ask our brain anything on the homepage — it remembers the whole conversation — or write a brief in the Studio and see the prompt it compiles to.