@inproceedings{lieberum-etal-2024-gemma,
title = "Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2",
author = "Lieberum, Tom and
Rajamanoharan, Senthooran and
Conmy, Arthur and
Smith, Lewis and
Sonnerat, Nicolas and
Varma, Vikrant and
Kramar, Janos and
Dragan, Anca and
Shah, Rohin and
Nanda, Neel",
editor = "Belinkov, Yonatan and
Kim, Najoung and
Jumelet, Jaap and
Mohebbi, Hosein and
Mueller, Aaron and
Chen, Hanjie",
booktitle = "Proceedings of the 7th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP",
month = nov,
year = "2024",
address = "Miami, Florida, US",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.blackboxnlp-1.19",
pages = "278--300",
abstract = "Sparse autoencoders (SAEs) are an unsupervised method for learning a sparse decomposition of a neural network{'}s latent representations into seemingly interpretable features.Despite recent excitement about their potential, research applications outside of industry are limited by the high cost of training a comprehensive suite of SAEs.In this work, we introduce Gemma Scope, an open suite of JumpReLU SAEs trained on all layers and sub-layers of Gemma 2 2B and 9B and select layers of Gemma 2 27B base models.We primarily train SAEs on the Gemma 2 pre-trained models, but additionally release SAEs trained on instruction-tuned Gemma 2 9B for comparison.We evaluate the quality of each SAE on standard metrics and release these results.We hope that by releasing these SAE weights, we can help make more ambitious safety and interpretability research easier for the community. Weights and a tutorial can be found at \url{https://huggingface.co/google/gemma-scope} and an interactive demo can be found at \url{https://neuronpedia.org/gemma-scope}.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lieberum-etal-2024-gemma">
<titleInfo>
<title>Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2</title>
</titleInfo>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Lieberum</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Senthooran</namePart>
<namePart type="family">Rajamanoharan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arthur</namePart>
<namePart type="family">Conmy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lewis</namePart>
<namePart type="family">Smith</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nicolas</namePart>
<namePart type="family">Sonnerat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vikrant</namePart>
<namePart type="family">Varma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Janos</namePart>
<namePart type="family">Kramar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anca</namePart>
<namePart type="family">Dragan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rohin</namePart>
<namePart type="family">Shah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Neel</namePart>
<namePart type="family">Nanda</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 7th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yonatan</namePart>
<namePart type="family">Belinkov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Najoung</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaap</namePart>
<namePart type="family">Jumelet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hosein</namePart>
<namePart type="family">Mohebbi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aaron</namePart>
<namePart type="family">Mueller</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hanjie</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Miami, Florida, US</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Sparse autoencoders (SAEs) are an unsupervised method for learning a sparse decomposition of a neural network’s latent representations into seemingly interpretable features.Despite recent excitement about their potential, research applications outside of industry are limited by the high cost of training a comprehensive suite of SAEs.In this work, we introduce Gemma Scope, an open suite of JumpReLU SAEs trained on all layers and sub-layers of Gemma 2 2B and 9B and select layers of Gemma 2 27B base models.We primarily train SAEs on the Gemma 2 pre-trained models, but additionally release SAEs trained on instruction-tuned Gemma 2 9B for comparison.We evaluate the quality of each SAE on standard metrics and release these results.We hope that by releasing these SAE weights, we can help make more ambitious safety and interpretability research easier for the community. Weights and a tutorial can be found at https://huggingface.co/google/gemma-scope and an interactive demo can be found at https://neuronpedia.org/gemma-scope.</abstract>
<identifier type="citekey">lieberum-etal-2024-gemma</identifier>
<location>
<url>https://aclanthology.org/2024.blackboxnlp-1.19</url>
</location>
<part>
<date>2024-11</date>
<extent unit="page">
<start>278</start>
<end>300</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
%A Lieberum, Tom
%A Rajamanoharan, Senthooran
%A Conmy, Arthur
%A Smith, Lewis
%A Sonnerat, Nicolas
%A Varma, Vikrant
%A Kramar, Janos
%A Dragan, Anca
%A Shah, Rohin
%A Nanda, Neel
%Y Belinkov, Yonatan
%Y Kim, Najoung
%Y Jumelet, Jaap
%Y Mohebbi, Hosein
%Y Mueller, Aaron
%Y Chen, Hanjie
%S Proceedings of the 7th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP
%D 2024
%8 November
%I Association for Computational Linguistics
%C Miami, Florida, US
%F lieberum-etal-2024-gemma
%X Sparse autoencoders (SAEs) are an unsupervised method for learning a sparse decomposition of a neural network’s latent representations into seemingly interpretable features.Despite recent excitement about their potential, research applications outside of industry are limited by the high cost of training a comprehensive suite of SAEs.In this work, we introduce Gemma Scope, an open suite of JumpReLU SAEs trained on all layers and sub-layers of Gemma 2 2B and 9B and select layers of Gemma 2 27B base models.We primarily train SAEs on the Gemma 2 pre-trained models, but additionally release SAEs trained on instruction-tuned Gemma 2 9B for comparison.We evaluate the quality of each SAE on standard metrics and release these results.We hope that by releasing these SAE weights, we can help make more ambitious safety and interpretability research easier for the community. Weights and a tutorial can be found at https://huggingface.co/google/gemma-scope and an interactive demo can be found at https://neuronpedia.org/gemma-scope.
%U https://aclanthology.org/2024.blackboxnlp-1.19
%P 278-300
Markdown (Informal)
[Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://aclanthology.org/2024.blackboxnlp-1.19) (Lieberum et al., BlackboxNLP 2024)
ACL
- Tom Lieberum, Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith, Nicolas Sonnerat, Vikrant Varma, Janos Kramar, Anca Dragan, Rohin Shah, and Neel Nanda. 2024. Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2. In Proceedings of the 7th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP, pages 278–300, Miami, Florida, US. Association for Computational Linguistics.