[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-154613-105":59,"doc-detail-154613-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","creative-convergence-or-imitation-genre-specific-homogeneity-in-llm-generated-chinese-literature","Creative Convergence or Imitation? Genre-Specific Homogeneity in LLM-Generated Chinese Literature","","Large Language Models can generate fluent, grammatically coherent narratives at scale, yet their stories often become structurally homogenized, relying on repetitive event combinations and stereotypical resolutions. This paper proposes a theoretical framework that operationalizes Proppian narratology by defining 34 narrative functions tailored to modern Chinese web literature. A human-annotated corpus and experiments show that the dominant causes of singular narrative logic and severe homogenization are failures to correctly comprehend narrative-function meanings and adherence to rigid generation paradigms.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/creative-convergence-or-imitation-genre-specific-homogeneity-in-llm-generated-chinese-literature/154613/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/creative-convergence-or-imitation-genre-specific-homogeneity-in-llm-generated-chinese-literature/154613.png","ImageObject",300,407,{"name":92,"@type":93},"Kyle","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-06","2026-08-28",true,{"@type":102,"interactionType":103,"userInteractionCount":39},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What structural problem do the authors identify in LLM-generated Chinese literature?","Question",{"text":112,"@type":113},"The authors identify structural homogenization, where stories follow rigid plot templates, producing repetitive event patterns and stereotypical resolutions despite local sentence fluency.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How does the proposed framework use Proppian narratology to analyze LLM narratives?",{"text":117,"@type":113},"The framework operationalizes Proppian narratology by defining 34 narrative functions adapted for modern Chinese web narrative structures, then uses them to uncover underlying narrative logic in generated texts.",{"name":119,"@type":110,"acceptedAnswer":120},"Why do the authors argue the homogenization occurs in current LLMs?",{"text":121,"@type":113},"Experiments indicate LLMs cannot accurately comprehend the meanings of narrative functions, causing the models to adhere to rigid narrative generation paradigms rather than coherent, causally linked plot development.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},154613,1787895489,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":39,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":129,"read_time":144},3985741905716,"https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d","Creative Convergence or Imitation? Genre-Specific Homogeneity in LLM-Generated Chinese Literature  \nYuanchi Ma, Kaize Shi, Hui He, Zhihua Zhang, Zhongxiang Lei, Ziliang Qiu, Renfen Hu, Jiamou Liu  \narXiv :2603 . 14430v1 [ cs .CL] 15 Mar 2026  \nAbstract  \nLarge Language Models (LLMs) have demonstrated remarkable capabilities in narrative generation. However, they often produce structurally homogenized stories, frequently following repetitive arrangements and combinations of plot events along with stereotypical resolutions. In this paper, we propose a novel theoretical framework for analysis by incorporating Proppian narratology and narrative functions.  \nThis framework is used to analyze the composition of narrative texts generated by LLMs to uncover their underlying narrative logic. Taking Chinese web literature as our research focus, we extend Propp’s narrative theory, defining  \n34 narrative functions suited to modern web narrative structures. We further construct a human-annotated corpus to support the analysis of narrative structures within LLM-generated text. Experiments reveal that the primary reasons for the singular narrative logic and severe homogenization in generated texts are that current LLMs are unable to correctly comprehend the meanings of narrative functions and instead adhere to rigid narrative generation paradigms.  \n1 Introduction  \nThe intersection of AI and creative writing has witnessed rapid evolution, with Large Language Models (LLMs) now demonstrating the ability to produce fluent, grammatically coherent stories at scale (Brown et al., 2020 ; Guo et al., 2025) . This generative capability is particularly disruptive for high-volume literary markets, such as Chinese Online Literature (web fiction) (Marco et al., 2024b) . As of 2024, this industry generated approximately USD 6.9 billion in revenue and served nearly a billion readers globally 1. Beyond its market scale, this genre offers a unique technical vantage point: it is characterized by highly serialized formats and distinct, trope-driven architectures such as the “cul-  \n1[https://www.reuters.com/markets/emerging/chinas-online-literature](https://www.reuters.com/markets/emerging/chinas-online-literature)expands-overseas-readership-cultural-export-push-2025-06-17/  \ntivation arcs” in Xianxia or specific role dynamics in Urban romance (Jin and Liu, 2022 ; Zhang, 2023) . These structural constraints make the genre an ideal, rigorous testbed for evaluating whether AI models can master complex narrative patterns beyond mere sentence-level fluency.  \nDespite these advancements and the evident structural richness inherent in genres like Chinese web fiction, a critical limitation persists: LLMgenerated narratives are frequently criticized for being “homogenous” or “formulaic”(Marco et al., 2024b) . While the text can be fluent and grammatically correct on a sentence-by-sentence basis, the overarching stories often feel repetitious and lack genuine engagement over extended contexts. Existing diagnostic approaches have primarily focused on surface-level linguistic features, such as lexical diversity or n-gram repetition (Xu et al., 2025) . However, we contend that these metrics fail to capture the fundamental issue. A compelling story transcends its individual words; it is intrinsically a dynamic sequence of causally linked events and character interactions. The perceived ‘boredom’and lack of novelty in AI-generated fiction thus derive not from word choice, but from a deeper“structural collapse”, which manifests as a rigid adherence to simplistic plot templates that hinder meaningful narrative development and lead to predictable, uninspired outcomes.  \nDiagnosing this structural collapse proves challenging given the limitations of current quantitative frameworks in NLP for evaluating ‘plot logic’or narrative coherence. Standard metrics like BLEU or perplexity, while effective for surfacelevel linguistic evaluation, are ill-equipped to assess whether a prota","cbCaiohNXh6AIFsJ","https://ap.wps.com/l/cbCaiohNXh6AIFsJ","pdf",2698386,17,"English","# Introduction\n## Motivation: AI and narrative generation in online literature\n## Problem: structural collapse and homogenous story logic\n## Limitation of existing surface-level metrics\n## Proposed framework: Proppian narrative functions for modern plots\n## Contributions: function taxonomy and annotated corpus\n## Empirical evaluation approach","[{\"question\":\"What structural problem do the authors identify in LLM-generated Chinese literature?\",\"answer\":\"The authors identify structural homogenization, where stories follow rigid plot templates, producing repetitive event patterns and stereotypical resolutions despite local sentence fluency.\"},{\"question\":\"How does the proposed framework use Proppian narratology to analyze LLM narratives?\",\"answer\":\"The framework operationalizes Proppian narratology by defining 34 narrative functions adapted for modern Chinese web narrative structures, then uses them to uncover underlying narrative logic in generated texts.\"},{\"question\":\"Why do the authors argue the homogenization occurs in current LLMs?\",\"answer\":\"Experiments indicate LLMs cannot accurately comprehend the meanings of narrative functions, causing the models to adhere to rigid narrative generation paradigms rather than coherent, causally linked plot development.\"}]","Creative Convergence or Imitation? Genre-Specific Homogeneity in LLM-Generated Chinese Literature | PDF",43]