@inproceedings{samanta-etal-2026-simple,
title = "Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation {\&} The Evaluation Gap",
author = "Samanta, Kunal and
Shohan, Faisal Tareque and
Trabelsi, Amine and
Khoury, Richard",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.2006/",
pages = "43326--43345",
ISBN = "979-8-89176-390-6",
abstract = "Multi-party social dialogue remains underexplored in the literature,in part due to the difficulty and cost of evaluation. As a result,recent work on synthetic dialogue generation often relies on automatedmetrics and LLM-as-a-Judge frameworks, despite limited evidence thatsuch judges reflect human preferences in social settings. In this work,we introduce a lightweight and controllable multi-party dialoguegeneration framework (MPOD) as an experimental instrument forstudying generation and evaluation in social interaction. Using thisframework, we conduct human evaluations of open-domain multi-partydialogue simulation and directly compare human judgments againststate-of-the-art LLM judges. Across 319 pairwise comparisons, weobserve near-random agreement between humans and automated judges(Cohen{'}s $\kappa \approx 0.11$), driven by systematic behaviorsincluding extreme tie aversion and strong sensitivity toassistant-style verbosity. Crucially, human{--}human inter-annotatoragreement ($\kappa = 0.29$) is substantially higher than human{--}LLMagreement. To isolate themechanism underlying this misalignment, we introduce a controlled\textit{Transplant Ablation}, showing that LLM judges consistentlyprefer conversations containing a single proprietary, assistant-styleagent. Additional stress tests show that judges prefer GPT-styleconversations even when utterance order is randomly shuffled,indicating insensitivity to conversational structure and coherence.Our findings provide controlled evidence that currentinstruction-tuned LLM judges do not reliably reflect human preferences for naturalness, engagingness, and overall quality in multi-party social dialogue, calling into question their widespreaduse for validating synthetic conversational data."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="samanta-etal-2026-simple">
<titleInfo>
<title>Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation & The Evaluation Gap</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kunal</namePart>
<namePart type="family">Samanta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Faisal</namePart>
<namePart type="given">Tareque</namePart>
<namePart type="family">Shohan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amine</namePart>
<namePart type="family">Trabelsi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Richard</namePart>
<namePart type="family">Khoury</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>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</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">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>Multi-party social dialogue remains underexplored in the literature,in part due to the difficulty and cost of evaluation. As a result,recent work on synthetic dialogue generation often relies on automatedmetrics and LLM-as-a-Judge frameworks, despite limited evidence thatsuch judges reflect human preferences in social settings. In this work,we introduce a lightweight and controllable multi-party dialoguegeneration framework (MPOD) as an experimental instrument forstudying generation and evaluation in social interaction. Using thisframework, we conduct human evaluations of open-domain multi-partydialogue simulation and directly compare human judgments againststate-of-the-art LLM judges. Across 319 pairwise comparisons, weobserve near-random agreement between humans and automated judges(Cohen’s ąppa \approx 0.11), driven by systematic behaviorsincluding extreme tie aversion and strong sensitivity toassistant-style verbosity. Crucially, human–human inter-annotatoragreement (ąppa = 0.29) is substantially higher than human–LLMagreement. To isolate themechanism underlying this misalignment, we introduce a controlledTransplant Ablation, showing that LLM judges consistentlyprefer conversations containing a single proprietary, assistant-styleagent. Additional stress tests show that judges prefer GPT-styleconversations even when utterance order is randomly shuffled,indicating insensitivity to conversational structure and coherence.Our findings provide controlled evidence that currentinstruction-tuned LLM judges do not reliably reflect human preferences for naturalness, engagingness, and overall quality in multi-party social dialogue, calling into question their widespreaduse for validating synthetic conversational data.</abstract>
<identifier type="citekey">samanta-etal-2026-simple</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.2006/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>43326</start>
<end>43345</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation & The Evaluation Gap
%A Samanta, Kunal
%A Shohan, Faisal Tareque
%A Trabelsi, Amine
%A Khoury, Richard
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F samanta-etal-2026-simple
%X Multi-party social dialogue remains underexplored in the literature,in part due to the difficulty and cost of evaluation. As a result,recent work on synthetic dialogue generation often relies on automatedmetrics and LLM-as-a-Judge frameworks, despite limited evidence thatsuch judges reflect human preferences in social settings. In this work,we introduce a lightweight and controllable multi-party dialoguegeneration framework (MPOD) as an experimental instrument forstudying generation and evaluation in social interaction. Using thisframework, we conduct human evaluations of open-domain multi-partydialogue simulation and directly compare human judgments againststate-of-the-art LLM judges. Across 319 pairwise comparisons, weobserve near-random agreement between humans and automated judges(Cohen’s ąppa \approx 0.11), driven by systematic behaviorsincluding extreme tie aversion and strong sensitivity toassistant-style verbosity. Crucially, human–human inter-annotatoragreement (ąppa = 0.29) is substantially higher than human–LLMagreement. To isolate themechanism underlying this misalignment, we introduce a controlledTransplant Ablation, showing that LLM judges consistentlyprefer conversations containing a single proprietary, assistant-styleagent. Additional stress tests show that judges prefer GPT-styleconversations even when utterance order is randomly shuffled,indicating insensitivity to conversational structure and coherence.Our findings provide controlled evidence that currentinstruction-tuned LLM judges do not reliably reflect human preferences for naturalness, engagingness, and overall quality in multi-party social dialogue, calling into question their widespreaduse for validating synthetic conversational data.
%U https://aclanthology.org/2026.acl-long.2006/
%P 43326-43345
Markdown (Informal)
[Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation & The Evaluation Gap](https://aclanthology.org/2026.acl-long.2006/) (Samanta et al., ACL 2026)
ACL