[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-503463-105":59,"doc-detail-503463-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","alignment-by-stereotyping-how-llms-sacrifice-individual-distinctiveness-for-cultural-adaptation","Alignment by Stereotyping: How LLMs Sacrifice Individual Distinctiveness - for Cultural Adaptation","","Large language models increasingly support personalized interaction using demographic conditioning from user profiles for cultural adaptation. This study tests whether demographic profiles truly serve individuals or improve accuracy by erasing individual distinctiveness. Across seven models, including GPT-5.1, demographic profiling increases value-alignment accuracy for most models while systematically compressing individuality toward group centroids, termed alignment by stereotyping.",{"@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/alignment-by-stereotyping-how-llms-sacrifice-individual-distinctiveness-for-cultural-adaptation/503463/",{"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/alignment-by-stereotyping-how-llms-sacrifice-individual-distinctiveness-for-cultural-adaptation/503463.png","ImageObject",300,407,{"name":92,"@type":93},"vwvw","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-17","2026-10-03",true,{"@type":102,"interactionType":103,"userInteractionCount":81},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What is the main hypothesis behind alignment by stereotyping?","Question",{"text":112,"@type":113},"The approach assumes demographic conditioning improves group-level accuracy, but it may do so by pulling responses toward demographic group centroids rather than preserving individual differences. This behavioral pattern is termed alignment by stereotyping.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How do demographic profiles affect value alignment and individuality?",{"text":117,"@type":113},"Demographic profiles raise value-alignment accuracy for 5 of 7 models, but they also cause model responses to collapse toward the group centroid, indicating a systematic cost to individual distinctiveness.",{"name":119,"@type":110,"acceptedAnswer":120},"What drives the observed tradeoff between alignment and individuality?",{"text":121,"@type":113},"Permutation tests and analysis of behavioral patterns show that models compress individuals into demographic prototypes. The paper further examines how distributing demographic signals across dialogue turns changes the degree of homogenization.","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},503463,1792214687,{"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":81,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":46,"language":139,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":67,"update_tm":143,"read_time":31},3985747874149,"https://ap-avatar.wpscdn.com/avatar/3a001161165b4dc3d9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1790129441039064114","Alignment by Stereotyping: How LLMs Sacrifice Individual Distinctiveness  \nfor Cultural Adaptation  \nQishuai Zhong 1 Zongmin Li 1 Siqi Fan2 Aixin Sun 1  \n1Nanyang Technological University, Singapore  \n2University of Electronic Science and Technology of China  \n[qishuai001@e.ntu.edu.sg](qishuai001@e.ntu.edu.sg)  \narXiv : 2609 .05993v 1 [cs .CL] 5 Sep 2026  \nAbstract  \nLarge language models are increasingly deployed for personalized interaction, and demographic conditioning via user profiles is a widely adopted strategy for cultural adaptation. We ask whether this approach genuinely serves individual users or achieves accuracy by erasing individual distinctiveness. Studying seven models including frontier GPT-5.1 on the World Values Survey, we find that demographic profiles improve value alignment accuracy formost models, but at a systematic cost to individuality. That is, models pull responses toward demographic group centroids rather than preserving individual differences, a behavioral pattern we term alignment by stereotyping. Permutation tests (10,000 permutations, six demographic attributes, seven models) certify that top-performing models compress individuals far above the human baseline; within-family scaling amplifies this tradeoff while degrading intrinsic cultural understanding. Using a synthetic dialogue dataset validated on real humanchatbot conversations from PRISM (Kirk et al., 2024), we further show that distributing demographic signals across conversational turns partially suppresses prototype retrieval compared to compact demographic labels, a finding validated on real conversations via PRISM but requiring replication at larger scale.  \n1 Introduction  \nAs large language models take on roles as personal assistants, tutors, and advisors, adapting to individual users has become a central design goal. The dominant approach is demographic conditioning: providing the model with a user profile (e.g., age, education, region, occupation) and expecting value-aligned, contextually appropriate responses (Mukherjee et al., 2024 ; Beck et al., 2024) . But does this strategy genuinely serve each individual, or does it sacrifice individual distinctiveness for group-level accuracy? Social psychology  \nwarns that demographic categories trigger grouplevel stereotyping (applying group prototypes to individuals regardless of individual evidence, without implying intentional bias; Fiske and Neuberg, 1990) at the expense of individual accuracy; we use the term in this sense throughout.  \nA growing body of work evaluates LLM value expression using the World Values Survey (Haerpfer et al., 2020) and derived benchmarks (Zhao et al., 2024 ; Durmus et al., 2024 ; Sukiennik et al., 2025) . These studies consistently find that models adjust outputs in response to demographic cues and achieve encouraging group-level alignment. What this paradigm cannot reveal is whether a model tracks the individual or the group: high average correlation with human values can arise not by modeling the individual, but by predicting the statistical centroid of their demographic group. High group-average alignment is thus compatible with every individual being wrong in the same direction, what we call alignment by stereotyping.  \nWe address two research questions: (RQ1) Does providing LLMs with demographic profiles improve group-average value alignment, and does this come at the cost of individual distinctiveness?(RQ2) What behavioral pattern drives the observed tradeoff: do models adapt to individuals, or compress them into demographic group prototypes?  \nFigure 1 illustrates our three-condition evaluation: NONE (no user context), PROFILE (explicit demographic attributes), and DIALOGUE (multi-turn dialogue history in place of demographics) . We answer both RQs empirically across seven models spanning open-weight (Llama, Qwen, DeepSeek) and closed-source (GPT-5 . 1) families. Our design introduces two methodological innovations:(i) a no-profile baseline (NONE) ","cbCaivrDlgVNHYen","https://ap.wps.com/l/cbCaivrDlgVNHYen","pdf",729356,"English","# Introduction\n# Evaluation Framework\n## Baseline and Conditioning Modes\n## Statistical Certification via Permutation Tests\n# Experimental Results and Tradeoff Analysis","[{\"question\":\"What is the main hypothesis behind alignment by stereotyping?\",\"answer\":\"The approach assumes demographic conditioning improves group-level accuracy, but it may do so by pulling responses toward demographic group centroids rather than preserving individual differences. This behavioral pattern is termed alignment by stereotyping.\"},{\"question\":\"How do demographic profiles affect value alignment and individuality?\",\"answer\":\"Demographic profiles raise value-alignment accuracy for 5 of 7 models, but they also cause model responses to collapse toward the group centroid, indicating a systematic cost to individual distinctiveness.\"},{\"question\":\"What drives the observed tradeoff between alignment and individuality?\",\"answer\":\"Permutation tests and analysis of behavioral patterns show that models compress individuals into demographic prototypes. The paper further examines how distributing demographic signals across dialogue turns changes the degree of homogenization.\"}]","Alignment by Stereotyping: How LLMs Sacrifice Individual Distinctiveness - for Cultural Adaptation | PDF",1791056661]