@inproceedings{long-etal-2026-latent,
title = "Are Latent Reasoning Tokens Unnecessary for Model Performance?",
author = "Long, Austin and
Paterson, Colin and
Xu, Jing",
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-1.70/",
pages = "1223--1232",
ISBN = "979-8-89176-476-7",
abstract = "Training large language models to perform latent reasoning may improve performance, but the lack of legible reasoning poses challenges for model interpretability and safety assurance. This challenge has motivated interpretability research that aims to characterise the functional role of latent reasoning tokens. Some studies indicate that latent reasoning tokens are effective, whereas other work suggests that these tokens are under-utilised. In this paper, we consider the tension between existing findings by reproducing prior work which shows that some latent reasoning tokens are not necessary. Two interesting observations emerge from our experiments. First, we find that latent reasoning models often generate redundant tokens due to the fixed token budgets. Second, we observe that explicit reasoning models fine-tuned on gold reasoning traces may copy the result of the most recent reasoning step when generating an answer."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="long-etal-2026-latent">
<titleInfo>
<title>Are Latent Reasoning Tokens Unnecessary for Model Performance?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Austin</namePart>
<namePart type="family">Long</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Colin</namePart>
<namePart type="family">Paterson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Xu</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>Training large language models to perform latent reasoning may improve performance, but the lack of legible reasoning poses challenges for model interpretability and safety assurance. This challenge has motivated interpretability research that aims to characterise the functional role of latent reasoning tokens. Some studies indicate that latent reasoning tokens are effective, whereas other work suggests that these tokens are under-utilised. In this paper, we consider the tension between existing findings by reproducing prior work which shows that some latent reasoning tokens are not necessary. Two interesting observations emerge from our experiments. First, we find that latent reasoning models often generate redundant tokens due to the fixed token budgets. Second, we observe that explicit reasoning models fine-tuned on gold reasoning traces may copy the result of the most recent reasoning step when generating an answer.</abstract>
<identifier type="citekey">long-etal-2026-latent</identifier>
<location>
<url>https://aclanthology.org/2026.blackboxnlp-1.70/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1223</start>
<end>1232</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Are Latent Reasoning Tokens Unnecessary for Model Performance?
%A Long, Austin
%A Paterson, Colin
%A Xu, Jing
%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 long-etal-2026-latent
%X Training large language models to perform latent reasoning may improve performance, but the lack of legible reasoning poses challenges for model interpretability and safety assurance. This challenge has motivated interpretability research that aims to characterise the functional role of latent reasoning tokens. Some studies indicate that latent reasoning tokens are effective, whereas other work suggests that these tokens are under-utilised. In this paper, we consider the tension between existing findings by reproducing prior work which shows that some latent reasoning tokens are not necessary. Two interesting observations emerge from our experiments. First, we find that latent reasoning models often generate redundant tokens due to the fixed token budgets. Second, we observe that explicit reasoning models fine-tuned on gold reasoning traces may copy the result of the most recent reasoning step when generating an answer.
%U https://aclanthology.org/2026.blackboxnlp-1.70/
%P 1223-1232
Markdown (Informal)
[Are Latent Reasoning Tokens Unnecessary for Model Performance?](https://aclanthology.org/2026.blackboxnlp-1.70/) (Long et al., BlackboxNLP 2026)
ACL