@inproceedings{quirke-etal-2026-bergson,
title = "Bergson: An Open Source Library for Data Attribution",
author = "Quirke, Lucia and
Jaburi, Louis and
Johnston, David O. and
Li, William Z. and
Paulo, Gon{\c{c}}alo and
Martres, Guillaume and
Gupta, Girish and
Belrose, Nora and
Biderman, Stella",
editor = "Agrawal, Sweta and
Abad, Alberto",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-demos.56/",
pages = "582--592",
ISBN = "979-8-89176-461-3",
abstract = "Data attribution is a promising field in interpretability that aims to explain model behavior through the influence of its training data, with applications including debugging undesirable model behavior and training dataset curation. However, significant engineering effort is required to perform it at scale, and many cutting-edge techniques lack open-source tooling and support. Bergson is an open-source library that closes this gap, providing a host of techniques that scale to very large language models and pre-training datasets. The library natively supports on-disk gradient stores and multi-node distributed training, and provides quality-of-life tools for researchers. Finally, Bergson includes the first open-source implementations of three leading data attribution methods: MAGIC, SOURCE, and TrackStar. The library is available at https://github.com/EleutherAI/bergson."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="quirke-etal-2026-bergson">
<titleInfo>
<title>Bergson: An Open Source Library for Data Attribution</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lucia</namePart>
<namePart type="family">Quirke</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Louis</namePart>
<namePart type="family">Jaburi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="given">O</namePart>
<namePart type="family">Johnston</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">William</namePart>
<namePart type="given">Z</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gonçalo</namePart>
<namePart type="family">Paulo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guillaume</namePart>
<namePart type="family">Martres</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Girish</namePart>
<namePart type="family">Gupta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nora</namePart>
<namePart type="family">Belrose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stella</namePart>
<namePart type="family">Biderman</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sweta</namePart>
<namePart type="family">Agrawal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alberto</namePart>
<namePart type="family">Abad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-461-3</identifier>
</relatedItem>
<abstract>Data attribution is a promising field in interpretability that aims to explain model behavior through the influence of its training data, with applications including debugging undesirable model behavior and training dataset curation. However, significant engineering effort is required to perform it at scale, and many cutting-edge techniques lack open-source tooling and support. Bergson is an open-source library that closes this gap, providing a host of techniques that scale to very large language models and pre-training datasets. The library natively supports on-disk gradient stores and multi-node distributed training, and provides quality-of-life tools for researchers. Finally, Bergson includes the first open-source implementations of three leading data attribution methods: MAGIC, SOURCE, and TrackStar. The library is available at https://github.com/EleutherAI/bergson.</abstract>
<identifier type="citekey">quirke-etal-2026-bergson</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-demos.56/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>582</start>
<end>592</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Bergson: An Open Source Library for Data Attribution
%A Quirke, Lucia
%A Jaburi, Louis
%A Johnston, David O.
%A Li, William Z.
%A Paulo, Gonçalo
%A Martres, Guillaume
%A Gupta, Girish
%A Belrose, Nora
%A Biderman, Stella
%Y Agrawal, Sweta
%Y Abad, Alberto
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-461-3
%F quirke-etal-2026-bergson
%X Data attribution is a promising field in interpretability that aims to explain model behavior through the influence of its training data, with applications including debugging undesirable model behavior and training dataset curation. However, significant engineering effort is required to perform it at scale, and many cutting-edge techniques lack open-source tooling and support. Bergson is an open-source library that closes this gap, providing a host of techniques that scale to very large language models and pre-training datasets. The library natively supports on-disk gradient stores and multi-node distributed training, and provides quality-of-life tools for researchers. Finally, Bergson includes the first open-source implementations of three leading data attribution methods: MAGIC, SOURCE, and TrackStar. The library is available at https://github.com/EleutherAI/bergson.
%U https://aclanthology.org/2026.emnlp-demos.56/
%P 582-592
Markdown (Informal)
[Bergson: An Open Source Library for Data Attribution](https://aclanthology.org/2026.emnlp-demos.56/) (Quirke et al., EMNLP 2026)
ACL
- Lucia Quirke, Louis Jaburi, David O. Johnston, William Z. Li, Gonçalo Paulo, Guillaume Martres, Girish Gupta, Nora Belrose, and Stella Biderman. 2026. Bergson: An Open Source Library for Data Attribution. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 582–592, Budapest, Hungary. Association for Computational Linguistics.