@inproceedings{teixeira-de-lima-etal-2025-know,
title = "Know Your {RAG}: Dataset Taxonomy and Generation Strategies for Evaluating {RAG} Systems",
author = "Teixeira de Lima, Rafael and
Gupta, Shubham and
Berrospi Ramis, Cesar and
Mishra, Lokesh and
Dolfi, Michele and
Staar, Peter and
Vagenas, Panagiotis",
editor = "Rambow, Owen and
Wanner, Leo and
Apidianaki, Marianna and
Al-Khalifa, Hend and
Eugenio, Barbara Di and
Schockaert, Steven and
Darwish, Kareem and
Agarwal, Apoorv",
booktitle = "Proceedings of the 31st International Conference on Computational Linguistics: Industry Track",
month = jan,
year = "2025",
address = "Abu Dhabi, UAE",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.coling-industry.4/",
pages = "39--57",
abstract = "Retrieval Augmented Generation (RAG) systems are a widespread application of Large Language Models (LLMs) in the industry. While many tools exist empowering developers to build their own systems, measuring their performance locally, with datasets reflective of the system`s use cases, is a technological challenge. Solutions to this problem range from non-specific and cheap (most public datasets) to specific and costly (generating data from local documents). In this paper, we show that using public question and answer (Q{\&}A) datasets to assess retrieval performance can lead to non-optimal systems design, and that common tools for RAG dataset generation can lead to unbalanced data. We propose solutions to these issues based on the characterization of RAG datasets through labels and through label-targeted data generation. Finally, we show that fine-tuned small LLMs can efficiently generate Q{\&}A datasets. We believe that these observations are invaluable to the know-your-data step of RAG systems development."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="teixeira-de-lima-etal-2025-know">
<titleInfo>
<title>Know Your RAG: Dataset Taxonomy and Generation Strategies for Evaluating RAG Systems</title>
</titleInfo>
<name type="personal">
<namePart type="given">Rafael</namePart>
<namePart type="family">Teixeira de Lima</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shubham</namePart>
<namePart type="family">Gupta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cesar</namePart>
<namePart type="family">Berrospi Ramis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lokesh</namePart>
<namePart type="family">Mishra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Michele</namePart>
<namePart type="family">Dolfi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">Staar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Panagiotis</namePart>
<namePart type="family">Vagenas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-01</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 31st International Conference on Computational Linguistics: Industry Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Owen</namePart>
<namePart type="family">Rambow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leo</namePart>
<namePart type="family">Wanner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marianna</namePart>
<namePart type="family">Apidianaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hend</namePart>
<namePart type="family">Al-Khalifa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Barbara</namePart>
<namePart type="given">Di</namePart>
<namePart type="family">Eugenio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Schockaert</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kareem</namePart>
<namePart type="family">Darwish</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Apoorv</namePart>
<namePart type="family">Agarwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, UAE</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Retrieval Augmented Generation (RAG) systems are a widespread application of Large Language Models (LLMs) in the industry. While many tools exist empowering developers to build their own systems, measuring their performance locally, with datasets reflective of the system‘s use cases, is a technological challenge. Solutions to this problem range from non-specific and cheap (most public datasets) to specific and costly (generating data from local documents). In this paper, we show that using public question and answer (Q&A) datasets to assess retrieval performance can lead to non-optimal systems design, and that common tools for RAG dataset generation can lead to unbalanced data. We propose solutions to these issues based on the characterization of RAG datasets through labels and through label-targeted data generation. Finally, we show that fine-tuned small LLMs can efficiently generate Q&A datasets. We believe that these observations are invaluable to the know-your-data step of RAG systems development.</abstract>
<identifier type="citekey">teixeira-de-lima-etal-2025-know</identifier>
<location>
<url>https://aclanthology.org/2025.coling-industry.4/</url>
</location>
<part>
<date>2025-01</date>
<extent unit="page">
<start>39</start>
<end>57</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Know Your RAG: Dataset Taxonomy and Generation Strategies for Evaluating RAG Systems
%A Teixeira de Lima, Rafael
%A Gupta, Shubham
%A Berrospi Ramis, Cesar
%A Mishra, Lokesh
%A Dolfi, Michele
%A Staar, Peter
%A Vagenas, Panagiotis
%Y Rambow, Owen
%Y Wanner, Leo
%Y Apidianaki, Marianna
%Y Al-Khalifa, Hend
%Y Eugenio, Barbara Di
%Y Schockaert, Steven
%Y Darwish, Kareem
%Y Agarwal, Apoorv
%S Proceedings of the 31st International Conference on Computational Linguistics: Industry Track
%D 2025
%8 January
%I Association for Computational Linguistics
%C Abu Dhabi, UAE
%F teixeira-de-lima-etal-2025-know
%X Retrieval Augmented Generation (RAG) systems are a widespread application of Large Language Models (LLMs) in the industry. While many tools exist empowering developers to build their own systems, measuring their performance locally, with datasets reflective of the system‘s use cases, is a technological challenge. Solutions to this problem range from non-specific and cheap (most public datasets) to specific and costly (generating data from local documents). In this paper, we show that using public question and answer (Q&A) datasets to assess retrieval performance can lead to non-optimal systems design, and that common tools for RAG dataset generation can lead to unbalanced data. We propose solutions to these issues based on the characterization of RAG datasets through labels and through label-targeted data generation. Finally, we show that fine-tuned small LLMs can efficiently generate Q&A datasets. We believe that these observations are invaluable to the know-your-data step of RAG systems development.
%U https://aclanthology.org/2025.coling-industry.4/
%P 39-57
Markdown (Informal)
[Know Your RAG: Dataset Taxonomy and Generation Strategies for Evaluating RAG Systems](https://aclanthology.org/2025.coling-industry.4/) (Teixeira de Lima et al., COLING 2025)
ACL