@inproceedings{durandard-etal-2026-geometry,
title = "The Geometry of Political Abstraction in Large Language Models",
author = "Durandard, No{\'e} and
Dhawan, Saurabh and
Poibeau, Thierry",
editor = "Arad, Dana and
Belinkov, Yonatan and
Chen, Hanjie and
Feldhus, Nils and
Mueller, Aaron and
Mohebbi, Hosein and
Sarti, Gabriele and
Tutek, Martin",
booktitle = "Proceedings of the 9th {B}lackbox{NLP} Workshop: Analyzing and Interpreting Neural Networks for {NLP}",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.blackboxnlp-1.43/",
pages = "739--760",
ISBN = "979-8-89176-476-7",
abstract = "Large language models encode high-level concepts with heterogeneous domain-specific manifestations, but it remains unclear whether their internal representations are organized in a form that supports cross-domain generalization. We investigate this question by adapting representational geometry methods from systems neuroscience and using political ideology as a naturalistic test case for a candidate latent variable spanning diverse policy domains. Across three open-weight LLMs, we analyze hidden states elicited by matched left- and right-leaning statements from ten policy domains. Within each domain, while political polarity is linearly decodable, lexical features achieve comparable performance. By contrast, cross-domain generalization substantially exceeds the lexical baseline and peaks in intermediate layers. This peak coincides with a marked reduction in participation ratio, suggesting that domain-specific political distinctions, instead of being maintained as independent axes, become aligned within a shared low-dimensional subspace. These findings suggest that LLMs transiently organize heterogeneous inputs into a common representational geometry that supports cross-domain abstraction, offering insight into how models structure complex social knowledge."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="durandard-etal-2026-geometry">
<titleInfo>
<title>The Geometry of Political Abstraction in Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Noé</namePart>
<namePart type="family">Durandard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Saurabh</namePart>
<namePart type="family">Dhawan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thierry</namePart>
<namePart type="family">Poibeau</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 9th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dana</namePart>
<namePart type="family">Arad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<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">Hanjie</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nils</namePart>
<namePart type="family">Feldhus</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">Hosein</namePart>
<namePart type="family">Mohebbi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gabriele</namePart>
<namePart type="family">Sarti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Martin</namePart>
<namePart type="family">Tutek</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">979-8-89176-476-7</identifier>
</relatedItem>
<abstract>Large language models encode high-level concepts with heterogeneous domain-specific manifestations, but it remains unclear whether their internal representations are organized in a form that supports cross-domain generalization. We investigate this question by adapting representational geometry methods from systems neuroscience and using political ideology as a naturalistic test case for a candidate latent variable spanning diverse policy domains. Across three open-weight LLMs, we analyze hidden states elicited by matched left- and right-leaning statements from ten policy domains. Within each domain, while political polarity is linearly decodable, lexical features achieve comparable performance. By contrast, cross-domain generalization substantially exceeds the lexical baseline and peaks in intermediate layers. This peak coincides with a marked reduction in participation ratio, suggesting that domain-specific political distinctions, instead of being maintained as independent axes, become aligned within a shared low-dimensional subspace. These findings suggest that LLMs transiently organize heterogeneous inputs into a common representational geometry that supports cross-domain abstraction, offering insight into how models structure complex social knowledge.</abstract>
<identifier type="citekey">durandard-etal-2026-geometry</identifier>
<location>
<url>https://aclanthology.org/2026.blackboxnlp-1.43/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>739</start>
<end>760</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T The Geometry of Political Abstraction in Large Language Models
%A Durandard, Noé
%A Dhawan, Saurabh
%A Poibeau, Thierry
%Y Arad, Dana
%Y Belinkov, Yonatan
%Y Chen, Hanjie
%Y Feldhus, Nils
%Y Mueller, Aaron
%Y Mohebbi, Hosein
%Y Sarti, Gabriele
%Y Tutek, Martin
%S Proceedings of the 9th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-476-7
%F durandard-etal-2026-geometry
%X Large language models encode high-level concepts with heterogeneous domain-specific manifestations, but it remains unclear whether their internal representations are organized in a form that supports cross-domain generalization. We investigate this question by adapting representational geometry methods from systems neuroscience and using political ideology as a naturalistic test case for a candidate latent variable spanning diverse policy domains. Across three open-weight LLMs, we analyze hidden states elicited by matched left- and right-leaning statements from ten policy domains. Within each domain, while political polarity is linearly decodable, lexical features achieve comparable performance. By contrast, cross-domain generalization substantially exceeds the lexical baseline and peaks in intermediate layers. This peak coincides with a marked reduction in participation ratio, suggesting that domain-specific political distinctions, instead of being maintained as independent axes, become aligned within a shared low-dimensional subspace. These findings suggest that LLMs transiently organize heterogeneous inputs into a common representational geometry that supports cross-domain abstraction, offering insight into how models structure complex social knowledge.
%U https://aclanthology.org/2026.blackboxnlp-1.43/
%P 739-760
Markdown (Informal)
[The Geometry of Political Abstraction in Large Language Models](https://aclanthology.org/2026.blackboxnlp-1.43/) (Durandard et al., BlackboxNLP 2026)
ACL