@inproceedings{hoang-etal-2025-distill,
title = "Distill-{C}: Enhanced {NL}2{SQL} via Distilled Customization with {LLM}s",
author = "Hoang, Cong Duy Vu and
Tangari, Gioacchino and
Lanfranchi, Clemence and
Guo, Dalu and
Cayet, Paul and
Siu, Steve and
Dharmasiri, Don and
Li, Yuan-Fang and
Duong, Long and
Hilloulin, Damien and
Patra, Rhicheek and
Hong, Sungpack and
Chafi, Hassan",
editor = "Chen, Weizhu and
Yang, Yi and
Kachuee, Mohammad and
Fu, Xue-Yong",
booktitle = "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 3: Industry Track)",
month = apr,
year = "2025",
address = "Albuquerque, New Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.naacl-industry.64/",
doi = "10.18653/v1/2025.naacl-industry.64",
pages = "833--848",
ISBN = "979-8-89176-194-0",
abstract = "The growing adoption of large language models (LLMs) in business applications has amplified interest in Natural Language to SQL (NL2SQL) solutions, in which there is competing demand for high performance and efficiency. Domain- and customer-specific requirements further complicate the problem. To address this conundrum, we introduce Distill-C, a distilled customization framework tailored for NL2SQL tasks. Distill-C utilizes large teacher LLMs to produce high-quality synthetic data through a robust and scalable pipeline. Finetuning smaller and open-source LLMs on this synthesized data enables them to rival or outperform teacher models an order of magnitude larger. Evaluated on multiple challenging benchmarks, Distill-C achieves an average improvement of 36{\%} in execution accuracy compared to the base models from three distinct LLM families. Additionally, on three internal customer benchmarks, Distill-C demonstrates a 22.6{\%} performance improvement over the base models. Our results demonstrate that Distill-C is an effective, high-performing and generalizable approach for deploying lightweight yet powerful NL2SQL models, delivering exceptional accuracies while maintaining low computational cost."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hoang-etal-2025-distill">
<titleInfo>
<title>Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Cong</namePart>
<namePart type="given">Duy</namePart>
<namePart type="given">Vu</namePart>
<namePart type="family">Hoang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gioacchino</namePart>
<namePart type="family">Tangari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Clemence</namePart>
<namePart type="family">Lanfranchi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dalu</namePart>
<namePart type="family">Guo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="family">Cayet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steve</namePart>
<namePart type="family">Siu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Don</namePart>
<namePart type="family">Dharmasiri</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuan-Fang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Long</namePart>
<namePart type="family">Duong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Damien</namePart>
<namePart type="family">Hilloulin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rhicheek</namePart>
<namePart type="family">Patra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sungpack</namePart>
<namePart type="family">Hong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hassan</namePart>
<namePart type="family">Chafi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 3: Industry Track)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Weizhu</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="family">Kachuee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xue-Yong</namePart>
<namePart type="family">Fu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Albuquerque, New Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-194-0</identifier>
</relatedItem>
<abstract>The growing adoption of large language models (LLMs) in business applications has amplified interest in Natural Language to SQL (NL2SQL) solutions, in which there is competing demand for high performance and efficiency. Domain- and customer-specific requirements further complicate the problem. To address this conundrum, we introduce Distill-C, a distilled customization framework tailored for NL2SQL tasks. Distill-C utilizes large teacher LLMs to produce high-quality synthetic data through a robust and scalable pipeline. Finetuning smaller and open-source LLMs on this synthesized data enables them to rival or outperform teacher models an order of magnitude larger. Evaluated on multiple challenging benchmarks, Distill-C achieves an average improvement of 36% in execution accuracy compared to the base models from three distinct LLM families. Additionally, on three internal customer benchmarks, Distill-C demonstrates a 22.6% performance improvement over the base models. Our results demonstrate that Distill-C is an effective, high-performing and generalizable approach for deploying lightweight yet powerful NL2SQL models, delivering exceptional accuracies while maintaining low computational cost.</abstract>
<identifier type="citekey">hoang-etal-2025-distill</identifier>
<identifier type="doi">10.18653/v1/2025.naacl-industry.64</identifier>
<location>
<url>https://aclanthology.org/2025.naacl-industry.64/</url>
</location>
<part>
<date>2025-04</date>
<extent unit="page">
<start>833</start>
<end>848</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs
%A Hoang, Cong Duy Vu
%A Tangari, Gioacchino
%A Lanfranchi, Clemence
%A Guo, Dalu
%A Cayet, Paul
%A Siu, Steve
%A Dharmasiri, Don
%A Li, Yuan-Fang
%A Duong, Long
%A Hilloulin, Damien
%A Patra, Rhicheek
%A Hong, Sungpack
%A Chafi, Hassan
%Y Chen, Weizhu
%Y Yang, Yi
%Y Kachuee, Mohammad
%Y Fu, Xue-Yong
%S Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 3: Industry Track)
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-194-0
%F hoang-etal-2025-distill
%X The growing adoption of large language models (LLMs) in business applications has amplified interest in Natural Language to SQL (NL2SQL) solutions, in which there is competing demand for high performance and efficiency. Domain- and customer-specific requirements further complicate the problem. To address this conundrum, we introduce Distill-C, a distilled customization framework tailored for NL2SQL tasks. Distill-C utilizes large teacher LLMs to produce high-quality synthetic data through a robust and scalable pipeline. Finetuning smaller and open-source LLMs on this synthesized data enables them to rival or outperform teacher models an order of magnitude larger. Evaluated on multiple challenging benchmarks, Distill-C achieves an average improvement of 36% in execution accuracy compared to the base models from three distinct LLM families. Additionally, on three internal customer benchmarks, Distill-C demonstrates a 22.6% performance improvement over the base models. Our results demonstrate that Distill-C is an effective, high-performing and generalizable approach for deploying lightweight yet powerful NL2SQL models, delivering exceptional accuracies while maintaining low computational cost.
%R 10.18653/v1/2025.naacl-industry.64
%U https://aclanthology.org/2025.naacl-industry.64/
%U https://doi.org/10.18653/v1/2025.naacl-industry.64
%P 833-848
Markdown (Informal)
[Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs](https://aclanthology.org/2025.naacl-industry.64/) (Hoang et al., NAACL 2025)
ACL
- Cong Duy Vu Hoang, Gioacchino Tangari, Clemence Lanfranchi, Dalu Guo, Paul Cayet, Steve Siu, Don Dharmasiri, Yuan-Fang Li, Long Duong, Damien Hilloulin, Rhicheek Patra, Sungpack Hong, and Hassan Chafi. 2025. Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 3: Industry Track), pages 833–848, Albuquerque, New Mexico. Association for Computational Linguistics.