# Citation Generation

> Citation generation is having a model attach citations to the sources that support each claim in its answer, so readers can verify the answer and unsupported claims are easier to spot.

- Identifier: PTL-0139
- Category: Retrieval & Tool Use
- Canonical URL: https://protologue.com/t/citation-generation/
- Also known as: attributed generation, source attribution, grounded citations
- Introduced: 2023

## Description

Citations are usually drawn from retrieved documents or search results, and some APIs return them as structured spans pointing into the source. Evaluations have found that models often cite sources that do not fully support the sentence they are attached to, so citation quality, both recall and precision, needs to be measured rather than assumed.

## Related terms

- [Retrieval-Augmented Generation](https://protologue.com/t/retrieval-augmented-generation/)
- [Web Search Tool](https://protologue.com/t/web-search-tool/)
- [Hallucination](https://protologue.com/t/hallucination/)

## Sources

- Gao et al. (2023). Enabling Large Language Models to Generate Text with Citations. https://arxiv.org/abs/2305.14627

## Cite this entry

Protologue. (2026). Citation Generation. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0139). https://protologue.com/t/citation-generation/

License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)