@inproceedings{wang-etal-2026-regulated,
title = "Regulated Reflection for Efficient Reasoning in Large reasoning language models",
author = "Wang, Mingqu and
Xu, Xinjie and
Zou, Yinan and
Ye, Qi",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.431/",
pages = "9068--9083",
ISBN = "979-8-89176-462-0",
abstract = "Large reasoning language models (LRLMs) achieve strong performance on complex reasoning tasks by generating long chain-of-thought trajectories, but such long-form reasoning also increases token consumption and latency. Existing inference-time methods that do not fine-tune the target LRLM optimize long reasoning through budgeting, early exit, uncertainty control, or reflection suppression, but they often treat reflection as a local signal rather than as a factor shaping the reasoning trajectory. In this paper, we identify unregulated reflection as a source of inefficient reasoning and characterize two failure modes: Trajectory Drift, where intermediate reflection opens weakly related branches that restart, reorient, or detour from the current path, and Verification Saturation, where answer-adjacent reflection repeatedly checks or restates an already-emerged provisional answer. To address these issues, we propose Regulated Reflection Decoding (RRD), an inference-time framework that requires no fine-tuning or parameter updates to the target LRLM and treats reflection events as online control points. RRD combines Trajectory Drift Suppression (TDS), which guides intermediate reasoning toward validated continuation branches, and Redundant Verification Suppression (RVS), which suppresses redundant verification when the provisional answer is sufficiently reliable. Experiments on multiple reasoning benchmarks show that RRD reduces reasoning length while preserving or improving accuracy, yielding a stronger accuracy-efficiency trade-off than direct inference and representative inference-time baselines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-regulated">
<titleInfo>
<title>Regulated Reflection for Efficient Reasoning in Large reasoning language models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mingqu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinjie</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yinan</namePart>
<namePart type="family">Zou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qi</namePart>
<namePart type="family">Ye</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>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Large reasoning language models (LRLMs) achieve strong performance on complex reasoning tasks by generating long chain-of-thought trajectories, but such long-form reasoning also increases token consumption and latency. Existing inference-time methods that do not fine-tune the target LRLM optimize long reasoning through budgeting, early exit, uncertainty control, or reflection suppression, but they often treat reflection as a local signal rather than as a factor shaping the reasoning trajectory. In this paper, we identify unregulated reflection as a source of inefficient reasoning and characterize two failure modes: Trajectory Drift, where intermediate reflection opens weakly related branches that restart, reorient, or detour from the current path, and Verification Saturation, where answer-adjacent reflection repeatedly checks or restates an already-emerged provisional answer. To address these issues, we propose Regulated Reflection Decoding (RRD), an inference-time framework that requires no fine-tuning or parameter updates to the target LRLM and treats reflection events as online control points. RRD combines Trajectory Drift Suppression (TDS), which guides intermediate reasoning toward validated continuation branches, and Redundant Verification Suppression (RVS), which suppresses redundant verification when the provisional answer is sufficiently reliable. Experiments on multiple reasoning benchmarks show that RRD reduces reasoning length while preserving or improving accuracy, yielding a stronger accuracy-efficiency trade-off than direct inference and representative inference-time baselines.</abstract>
<identifier type="citekey">wang-etal-2026-regulated</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.431/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>9068</start>
<end>9083</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Regulated Reflection for Efficient Reasoning in Large reasoning language models
%A Wang, Mingqu
%A Xu, Xinjie
%A Zou, Yinan
%A Ye, Qi
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F wang-etal-2026-regulated
%X Large reasoning language models (LRLMs) achieve strong performance on complex reasoning tasks by generating long chain-of-thought trajectories, but such long-form reasoning also increases token consumption and latency. Existing inference-time methods that do not fine-tune the target LRLM optimize long reasoning through budgeting, early exit, uncertainty control, or reflection suppression, but they often treat reflection as a local signal rather than as a factor shaping the reasoning trajectory. In this paper, we identify unregulated reflection as a source of inefficient reasoning and characterize two failure modes: Trajectory Drift, where intermediate reflection opens weakly related branches that restart, reorient, or detour from the current path, and Verification Saturation, where answer-adjacent reflection repeatedly checks or restates an already-emerged provisional answer. To address these issues, we propose Regulated Reflection Decoding (RRD), an inference-time framework that requires no fine-tuning or parameter updates to the target LRLM and treats reflection events as online control points. RRD combines Trajectory Drift Suppression (TDS), which guides intermediate reasoning toward validated continuation branches, and Redundant Verification Suppression (RVS), which suppresses redundant verification when the provisional answer is sufficiently reliable. Experiments on multiple reasoning benchmarks show that RRD reduces reasoning length while preserving or improving accuracy, yielding a stronger accuracy-efficiency trade-off than direct inference and representative inference-time baselines.
%U https://aclanthology.org/2026.findings-emnlp.431/
%P 9068-9083
Markdown (Informal)
[Regulated Reflection for Efficient Reasoning in Large reasoning language models](https://aclanthology.org/2026.findings-emnlp.431/) (Wang et al., Findings 2026)
ACL