@inproceedings{ma-etal-2026-capabilities,
title = "Capabilities and Evaluation Biases of Large Language Models in Classical {C}hinese Poetry Generation: A Case Study on Tang Poetry",
author = "Ma, Bolei and
Yao, Yina and
Haensch, Anna-Carolina",
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.836/",
pages = "16933--16957",
ISBN = "979-8-89176-395-1",
abstract = "Large Language Models (LLMs) are increasingly applied to creative domains, yet their performance in classical Chinese poetry generation and evaluation remains poorly understood. We propose a three-step evaluation framework that combines computational metrics, LLM-as-a-judge assessment, and human expert validation. Using this framework, we evaluate six state-of-the-art LLMs across multiple dimensions of poetic quality, including themes, emotions, imagery, form, and style, in the context of Tang poetry (唐诗) generation. Our analysis reveals a critical ``echo chamber'' effect: LLMs systematically overrate machine-generated poems that mimic statistical patterns yet fail strict prosodic rules, diverging significantly from human expert judgments. These findings underscore the limitations of using LLMs as standalone evaluators for culturally complex tasks, highlighting the necessity of hybrid human-model validation frameworks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ma-etal-2026-capabilities">
<titleInfo>
<title>Capabilities and Evaluation Biases of Large Language Models in Classical Chinese Poetry Generation: A Case Study on Tang Poetry</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bolei</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yina</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anna-Carolina</namePart>
<namePart type="family">Haensch</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 (LLMs) are increasingly applied to creative domains, yet their performance in classical Chinese poetry generation and evaluation remains poorly understood. We propose a three-step evaluation framework that combines computational metrics, LLM-as-a-judge assessment, and human expert validation. Using this framework, we evaluate six state-of-the-art LLMs across multiple dimensions of poetic quality, including themes, emotions, imagery, form, and style, in the context of Tang poetry (唐诗) generation. Our analysis reveals a critical “echo chamber” effect: LLMs systematically overrate machine-generated poems that mimic statistical patterns yet fail strict prosodic rules, diverging significantly from human expert judgments. These findings underscore the limitations of using LLMs as standalone evaluators for culturally complex tasks, highlighting the necessity of hybrid human-model validation frameworks.</abstract>
<identifier type="citekey">ma-etal-2026-capabilities</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.836/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>16933</start>
<end>16957</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Capabilities and Evaluation Biases of Large Language Models in Classical Chinese Poetry Generation: A Case Study on Tang Poetry
%A Ma, Bolei
%A Yao, Yina
%A Haensch, Anna-Carolina
%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 ma-etal-2026-capabilities
%X Large Language Models (LLMs) are increasingly applied to creative domains, yet their performance in classical Chinese poetry generation and evaluation remains poorly understood. We propose a three-step evaluation framework that combines computational metrics, LLM-as-a-judge assessment, and human expert validation. Using this framework, we evaluate six state-of-the-art LLMs across multiple dimensions of poetic quality, including themes, emotions, imagery, form, and style, in the context of Tang poetry (唐诗) generation. Our analysis reveals a critical “echo chamber” effect: LLMs systematically overrate machine-generated poems that mimic statistical patterns yet fail strict prosodic rules, diverging significantly from human expert judgments. These findings underscore the limitations of using LLMs as standalone evaluators for culturally complex tasks, highlighting the necessity of hybrid human-model validation frameworks.
%U https://aclanthology.org/2026.findings-acl.836/
%P 16933-16957
Markdown (Informal)
[Capabilities and Evaluation Biases of Large Language Models in Classical Chinese Poetry Generation: A Case Study on Tang Poetry](https://aclanthology.org/2026.findings-acl.836/) (Ma et al., Findings 2026)
ACL