@inproceedings{zou-etal-2026-development,
title = "Development of Linear Truth Encodings in Language Models: A Replication Study",
author = "Zou, Mingchuan and
Salhan, Suchir and
Buttery, Paula",
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-1.72/",
pages = "1248--1263",
ISBN = "979-8-89176-476-7",
abstract = "Probing studies consistently find that language models (LMs) encode the truth of a statement along a low-dimensional linear direction of the residual stream. The Truth Co-occurrence Hypothesis (TCH) (Ravfogel et al., 2025) provides a mechanistic interpretation: tracking a latent truth variable is loss-reducing because true statements co-occur with true statements in natural text, evaluated using toy models and base LLMs. We replicate and extend the toy setting. In the tested fully trainable single-layer transformers, removing truth co-occurrence while preserving fact frequency eliminates linear truth decodability, supporting the necessity of co-occurrence in this setting. However, the same manipulation leaves the original toy model largely unchanged, revealing that probe-based evidence can arise from architectural artefacts rather than a functional truth-tracking mechanism. In pretrained Llama-3.1 models (8B and 70B), we reproduce the reported behavioural effect that false context reduces confidence in true facts, but show that nearly half of the original effect is explained by surface confounds. After controlling for these confounds, the effect remains robust across six factual relations. Under our probing protocol, cross-relation performance is closer to within-relation performance at 70B than at 8B. Our work characterises the emergence conditions for linear truth representations, and highlights the limitations of probe-based interpretability studies. Reproducibility Code is available here: https://github.com/Samzmc/truth-encoding-blackboxnlp26"
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zou-etal-2026-development">
<titleInfo>
<title>Development of Linear Truth Encodings in Language Models: A Replication Study</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mingchuan</namePart>
<namePart type="family">Zou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Suchir</namePart>
<namePart type="family">Salhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paula</namePart>
<namePart type="family">Buttery</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>Probing studies consistently find that language models (LMs) encode the truth of a statement along a low-dimensional linear direction of the residual stream. The Truth Co-occurrence Hypothesis (TCH) (Ravfogel et al., 2025) provides a mechanistic interpretation: tracking a latent truth variable is loss-reducing because true statements co-occur with true statements in natural text, evaluated using toy models and base LLMs. We replicate and extend the toy setting. In the tested fully trainable single-layer transformers, removing truth co-occurrence while preserving fact frequency eliminates linear truth decodability, supporting the necessity of co-occurrence in this setting. However, the same manipulation leaves the original toy model largely unchanged, revealing that probe-based evidence can arise from architectural artefacts rather than a functional truth-tracking mechanism. In pretrained Llama-3.1 models (8B and 70B), we reproduce the reported behavioural effect that false context reduces confidence in true facts, but show that nearly half of the original effect is explained by surface confounds. After controlling for these confounds, the effect remains robust across six factual relations. Under our probing protocol, cross-relation performance is closer to within-relation performance at 70B than at 8B. Our work characterises the emergence conditions for linear truth representations, and highlights the limitations of probe-based interpretability studies. Reproducibility Code is available here: https://github.com/Samzmc/truth-encoding-blackboxnlp26</abstract>
<identifier type="citekey">zou-etal-2026-development</identifier>
<location>
<url>https://aclanthology.org/2026.blackboxnlp-1.72/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1248</start>
<end>1263</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Development of Linear Truth Encodings in Language Models: A Replication Study
%A Zou, Mingchuan
%A Salhan, Suchir
%A Buttery, Paula
%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 zou-etal-2026-development
%X Probing studies consistently find that language models (LMs) encode the truth of a statement along a low-dimensional linear direction of the residual stream. The Truth Co-occurrence Hypothesis (TCH) (Ravfogel et al., 2025) provides a mechanistic interpretation: tracking a latent truth variable is loss-reducing because true statements co-occur with true statements in natural text, evaluated using toy models and base LLMs. We replicate and extend the toy setting. In the tested fully trainable single-layer transformers, removing truth co-occurrence while preserving fact frequency eliminates linear truth decodability, supporting the necessity of co-occurrence in this setting. However, the same manipulation leaves the original toy model largely unchanged, revealing that probe-based evidence can arise from architectural artefacts rather than a functional truth-tracking mechanism. In pretrained Llama-3.1 models (8B and 70B), we reproduce the reported behavioural effect that false context reduces confidence in true facts, but show that nearly half of the original effect is explained by surface confounds. After controlling for these confounds, the effect remains robust across six factual relations. Under our probing protocol, cross-relation performance is closer to within-relation performance at 70B than at 8B. Our work characterises the emergence conditions for linear truth representations, and highlights the limitations of probe-based interpretability studies. Reproducibility Code is available here: https://github.com/Samzmc/truth-encoding-blackboxnlp26
%U https://aclanthology.org/2026.blackboxnlp-1.72/
%P 1248-1263
Markdown (Informal)
[Development of Linear Truth Encodings in Language Models: A Replication Study](https://aclanthology.org/2026.blackboxnlp-1.72/) (Zou et al., BlackboxNLP 2026)
ACL