@inproceedings{schlesinger-etal-2026-claro,
title = "{CLARO}: Controlled Attribute-Driven Reasoning Optimization for Efficient Chain-of-Thought",
author = "Schlesinger, Oded and
Kim, Young Kyung and
Di Martino, J. Matias and
Sapiro, Guillermo",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.1335/",
doi = "10.18653/v1/2026.findings-acl.1335",
pages = "26779--26799",
ISBN = "979-8-89176-395-1",
abstract = "Large language models exhibit strong reasoning capabilities but often require significant computational resources due to verbose, unstructured Chain-of-Thought outputs. Recent approaches guide reasoning length through token penalties or truncation, risking the omission of necessary steps. We posit that conciseness should be an emergent property of structured thought, rather than a result of artificially forced brevity. To this end, we first demonstrate that Attribute-Guided Prompting, a lightweight zero-shot strategy, improves reasoning performance while reducing inference cost. Building on this foundation, we introduce Controlled Attribute-Driven Reasoning Optimization (CLARO), a reinforcement learning framework designed to internalize these benefits. CLARO guides models to embed high-quality structural attributes, such as readability, math density, syntactic compression, and low redundancy, within a user-defined token budget. The proposed method outperforms state-of-the-art baselines across diverse benchmarks, yielding accuracy gains of up to 63.6{\%}, demonstrating that guiding generated output language structure enhances reasoning. Overall, our findings establish that optimizing the thought process structure refines reasoning efficacy, with computational efficiency emerging as a derivative benefit of a clearer thought process. Code and models are available at \url{https://github.com/odedsc/CLARO}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="schlesinger-etal-2026-claro">
<titleInfo>
<title>CLARO: Controlled Attribute-Driven Reasoning Optimization for Efficient Chain-of-Thought</title>
</titleInfo>
<name type="personal">
<namePart type="given">Oded</namePart>
<namePart type="family">Schlesinger</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Young</namePart>
<namePart type="given">Kyung</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">J</namePart>
<namePart type="given">Matias</namePart>
<namePart type="family">Di Martino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guillermo</namePart>
<namePart type="family">Sapiro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Large language models exhibit strong reasoning capabilities but often require significant computational resources due to verbose, unstructured Chain-of-Thought outputs. Recent approaches guide reasoning length through token penalties or truncation, risking the omission of necessary steps. We posit that conciseness should be an emergent property of structured thought, rather than a result of artificially forced brevity. To this end, we first demonstrate that Attribute-Guided Prompting, a lightweight zero-shot strategy, improves reasoning performance while reducing inference cost. Building on this foundation, we introduce Controlled Attribute-Driven Reasoning Optimization (CLARO), a reinforcement learning framework designed to internalize these benefits. CLARO guides models to embed high-quality structural attributes, such as readability, math density, syntactic compression, and low redundancy, within a user-defined token budget. The proposed method outperforms state-of-the-art baselines across diverse benchmarks, yielding accuracy gains of up to 63.6%, demonstrating that guiding generated output language structure enhances reasoning. Overall, our findings establish that optimizing the thought process structure refines reasoning efficacy, with computational efficiency emerging as a derivative benefit of a clearer thought process. Code and models are available at https://github.com/odedsc/CLARO.</abstract>
<identifier type="citekey">schlesinger-etal-2026-claro</identifier>
<identifier type="doi">10.18653/v1/2026.findings-acl.1335</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.1335/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>26779</start>
<end>26799</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CLARO: Controlled Attribute-Driven Reasoning Optimization for Efficient Chain-of-Thought
%A Schlesinger, Oded
%A Kim, Young Kyung
%A Di Martino, J. Matias
%A Sapiro, Guillermo
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F schlesinger-etal-2026-claro
%X Large language models exhibit strong reasoning capabilities but often require significant computational resources due to verbose, unstructured Chain-of-Thought outputs. Recent approaches guide reasoning length through token penalties or truncation, risking the omission of necessary steps. We posit that conciseness should be an emergent property of structured thought, rather than a result of artificially forced brevity. To this end, we first demonstrate that Attribute-Guided Prompting, a lightweight zero-shot strategy, improves reasoning performance while reducing inference cost. Building on this foundation, we introduce Controlled Attribute-Driven Reasoning Optimization (CLARO), a reinforcement learning framework designed to internalize these benefits. CLARO guides models to embed high-quality structural attributes, such as readability, math density, syntactic compression, and low redundancy, within a user-defined token budget. The proposed method outperforms state-of-the-art baselines across diverse benchmarks, yielding accuracy gains of up to 63.6%, demonstrating that guiding generated output language structure enhances reasoning. Overall, our findings establish that optimizing the thought process structure refines reasoning efficacy, with computational efficiency emerging as a derivative benefit of a clearer thought process. Code and models are available at https://github.com/odedsc/CLARO.
%R 10.18653/v1/2026.findings-acl.1335
%U https://aclanthology.org/2026.findings-acl.1335/
%U https://doi.org/10.18653/v1/2026.findings-acl.1335
%P 26779-26799
Markdown (Informal)
[CLARO: Controlled Attribute-Driven Reasoning Optimization for Efficient Chain-of-Thought](https://aclanthology.org/2026.findings-acl.1335/) (Schlesinger et al., Findings 2026)
ACL