@inproceedings{ouyang-etal-2026-svote,
title = "{SV}ote-2{S}: Input-Side Evidence Voting for Two-Stage Self-Consistency in Multi-Hop {QA}",
author = "Ouyang, Guo and
Wang, Ge and
Zhu, Tao and
Zou, Yi and
Yin, Xuehao and
Li, Guoliang and
Ni, Fuchuan",
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.1741/",
doi = "10.18653/2026.findings-emnlp.1741",
pages = "37229--37249",
ISBN = "979-8-89176-462-0",
abstract = "Self-consistency (SC) improves chain-of-thought reasoning by aggregating multiple sampled answers, yet all existing variants operate in the output space{---}they refine answer selection but leave the noisy input context unchanged. We observe that when multiple reasoning traces independently cite the same support items, those items are likely key evidence for the question. Building on this input-side consensus signal, we propose SVote-2S, a two-stage self-consistency framework for multi-hop question answering. In Stage 1, SVote-2S performs confidence-weighted voting over support items cited across N sampled traces and selects a compact evidence set. In Stage 2, the model re-reasons over a filtered context containing only the selected evidence paragraphs (about 15{\%} of the original length on average on our benchmarks), enabling more focused reasoning. The final answer is determined by confidence-weighted fusion of samples from both stages. The fusion serves as a robustness safeguard, while the primary gain comes from evidence voting and re-reasoning. Across three benchmarks (HotpotQA, MuSiQue, 2WikiMultihopQA) and three LLMs, SVote-2S improves over SC by 5.36 EM on average and consistently outperforms the strongest output-side baseline (CISC) across all nine settings, without additional training or external retrieval."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ouyang-etal-2026-svote">
<titleInfo>
<title>SVote-2S: Input-Side Evidence Voting for Two-Stage Self-Consistency in Multi-Hop QA</title>
</titleInfo>
<name type="personal">
<namePart type="given">Guo</namePart>
<namePart type="family">Ouyang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ge</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tao</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Zou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuehao</namePart>
<namePart type="family">Yin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guoliang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fuchuan</namePart>
<namePart type="family">Ni</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>Self-consistency (SC) improves chain-of-thought reasoning by aggregating multiple sampled answers, yet all existing variants operate in the output space—they refine answer selection but leave the noisy input context unchanged. We observe that when multiple reasoning traces independently cite the same support items, those items are likely key evidence for the question. Building on this input-side consensus signal, we propose SVote-2S, a two-stage self-consistency framework for multi-hop question answering. In Stage 1, SVote-2S performs confidence-weighted voting over support items cited across N sampled traces and selects a compact evidence set. In Stage 2, the model re-reasons over a filtered context containing only the selected evidence paragraphs (about 15% of the original length on average on our benchmarks), enabling more focused reasoning. The final answer is determined by confidence-weighted fusion of samples from both stages. The fusion serves as a robustness safeguard, while the primary gain comes from evidence voting and re-reasoning. Across three benchmarks (HotpotQA, MuSiQue, 2WikiMultihopQA) and three LLMs, SVote-2S improves over SC by 5.36 EM on average and consistently outperforms the strongest output-side baseline (CISC) across all nine settings, without additional training or external retrieval.</abstract>
<identifier type="citekey">ouyang-etal-2026-svote</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1741</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1741/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>37229</start>
<end>37249</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SVote-2S: Input-Side Evidence Voting for Two-Stage Self-Consistency in Multi-Hop QA
%A Ouyang, Guo
%A Wang, Ge
%A Zhu, Tao
%A Zou, Yi
%A Yin, Xuehao
%A Li, Guoliang
%A Ni, Fuchuan
%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 ouyang-etal-2026-svote
%X Self-consistency (SC) improves chain-of-thought reasoning by aggregating multiple sampled answers, yet all existing variants operate in the output space—they refine answer selection but leave the noisy input context unchanged. We observe that when multiple reasoning traces independently cite the same support items, those items are likely key evidence for the question. Building on this input-side consensus signal, we propose SVote-2S, a two-stage self-consistency framework for multi-hop question answering. In Stage 1, SVote-2S performs confidence-weighted voting over support items cited across N sampled traces and selects a compact evidence set. In Stage 2, the model re-reasons over a filtered context containing only the selected evidence paragraphs (about 15% of the original length on average on our benchmarks), enabling more focused reasoning. The final answer is determined by confidence-weighted fusion of samples from both stages. The fusion serves as a robustness safeguard, while the primary gain comes from evidence voting and re-reasoning. Across three benchmarks (HotpotQA, MuSiQue, 2WikiMultihopQA) and three LLMs, SVote-2S improves over SC by 5.36 EM on average and consistently outperforms the strongest output-side baseline (CISC) across all nine settings, without additional training or external retrieval.
%R 10.18653/2026.findings-emnlp.1741
%U https://aclanthology.org/2026.findings-emnlp.1741/
%U https://doi.org/10.18653/2026.findings-emnlp.1741
%P 37229-37249
Markdown (Informal)
[SVote-2S: Input-Side Evidence Voting for Two-Stage Self-Consistency in Multi-Hop QA](https://aclanthology.org/2026.findings-emnlp.1741/) (Ouyang et al., Findings 2026)
ACL