---
language: en
title: "Citation faithfulness"
description: "Citation faithfulness checks whether a source supports what an answer claims. Learn to check conditions, numbers, and limits before using the information."
canonical: "https://tropk.ai/en/resources/glossary/citation-faithfulness"
last_modified: "2026-10-12T16:13:00.958Z"
---

# Citation faithfulness

Reviewed on 2026-10-06.

## What it means

[Citation](/en/resources/glossary/ai-citation) faithfulness is the correspondence between a claim in an answer and the content of the source it cites. Checking this faithfulness requires reading the reference and verifying that it supports the attributed meaning, including numbers, conditions, and limits. A valid URL or the name of a respected institution alone does not demonstrate this correspondence.

## Why it matters for marketing

For marketing, appearing as a source can result in an inaccurate representation of an offer. An answer may attribute a guarantee to a page that it does not provide, or turn a limited condition into a general benefit. Assessing faithfulness makes it possible to investigate the specific information that reached the public, rather than simply counting links pointing to the company’s domain.

## Example in practice

Imagine a store whose policy allows free exchanges only for defective items. A hypothetical answer cites this page to claim that any exchange is free. The URL is correct, but the condition has been removed. By comparing the two texts, the team records the distortion and reviews the clarity of the policy and frequently asked questions.

## How to apply or verify

Match each claim to the cited reference and capture the passage that supports it. Compare the subject, value, timeframe, scope, and exceptions. Classify the support as complete, partial, or absent, and keep the rationale. If the page changed after the answer was given, record the dates and look for the relevant version before attributing the issue to the model’s output.

## Important distinctions

Source faithfulness and factual truth are distinct checks: an answer can faithfully reproduce an outdated source. The study of verifiability uses citation precision to measure how well references support sentences, with specific rules for combinations of sources. [Citation recall](/en/resources/glossary/citation-recall), by contrast, measures coverage of the claims in an answer, including when references are missing.

## Sources and references

- [Evaluating Verifiability in Generative Search Engines](<https://arxiv.org/html/2304.09848v2>) — Liu, Zhang e Liang, arXiv

## Continue a leitura

- [AI hallucination](</en/resources/glossary/ai-hallucination>)
- [Primary source](</en/resources/glossary/primary-source>)
- [Citation recall](</en/resources/glossary/citation-recall>)

## About the author

[Vinicius Pacheco](</en/team/vinicius-pacheco>)

Responsible for Tropk’s technical content, focusing on crawling, indexing, structured data, and agent access to content.

---

[All pages in Markdown](<https://tropk.ai/sitemap.md>)