easyAI Engineering

What are hallucinations and what causes them?

Q: What are hallucinations and what causes them?

**TL;DR.** Hallucinations are **fluent but false** outputs — invented citations, nonexistent APIs, confident wrong facts. Root cause: the model optimizes **plausibility, not truth**; when the answer is missing from weights and context, the most likely-sounding continuation may be fabricated. training data: no answer context: no answer <rect class='d-box-accent' x='208' y='48' width='110' height

303 views

Understand the problem

Why models state falsehoods fluently, and the failure modes every AI engineer must anticipate.

hallucinationgroundingreliability

Attempt it yourself

Sketch your approach before reading the solution — that's what interviews test.

Stuck? AI Nudge Available

Get a conceptual hint to guide your logic without spoiling the final implementation.

Study the solution

The solution is waiting

Give it an honest attempt first — then compare your thinking with the full walkthrough.

Read the code

The abstain-or-cite pattern

const system = [
  "Answer ONLY from the provided context.",
  "Cite the doc id for every claim, like [doc-3].",
  "If the context does not contain the answer, reply exactly:",
  '"I could not find this in the provided documents."',
].join("\n");

// downstream: reject answers whose citations do not resolve
const ok = extractCitations(reply).every((id) => contextIds.has(id));

Join the discussion

Discussion (0)

No responses yet. Be the first to share what you think.

What are evals and why does every LLM feature need them?

What is the difference between zero-shot, one-shot and few-shot prompting?