Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.getcited.in/llms.txt

Use this file to discover all available pages before exploring further.

Hallucination is when a large language model generates information that is factually incorrect or entirely fabricated — not grounded in real sources or training data. An LLM might claim a brand offers a feature it does not, cite a publication that does not exist, or attribute a statistic to a source that never published it.

​
Why it matters

Hallucinated brand mentions are a GEO-specific risk that traditional SEO does not have. If an AI falsely claims your product has a safety issue, that hallucination shapes customer perception — and there is no traditional mechanism (like a DMCA takedown or search result removal) to address it. Conversely, hallucinated positive mentions are unreliable — they may appear in one run and disappear in the next, contributing to non-deterministic measurement noise.

​
How it applies in practice

Cited’s sentiment classification can sometimes detect hallucinated negative claims when the sentiment is strongly negative and the claim does not match known brand information. However, detecting hallucination systematically requires comparing AI claims against verified facts — which is beyond the scope of automated monitoring. Brands that notice suspicious claims in their dashboard data should review the raw AI responses to verify accuracy.
  • Grounding — the process that prevents hallucination
  • Sentiment — where hallucinated claims surface in measurement
  • Non-determinism — hallucinated content tends to be inconsistent across runs