[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-293655-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-293655-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","beyond-dialogue-a-profile-dialogue-alignment-framework-towards-general-role-playing-language-model","BEYOND DIALOGUE - A Profile-Dialogue Alignment Framework Towards General Role-Playing Language Model","","Large language models have reshaped role-playing, but role-playing training still suffers from two major problems. First, prompting dialogue training with a predefined role profile for specific scenarios can create bias and contradictions between profile traits and generated dialogue. Second, models often imitate roles using only the overall profile, neglecting sentence-level profile-dialogue alignment. To address this, the BEYOND DIALOGUE framework introduces beyond-dialogue tasks and an automated low-cost prompting mechanism to generate reasoning data, enabling fine-grained alignment and improved adherence to role profiles.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/beyond-dialogue-a-profile-dialogue-alignment-framework-towards-general-role-playing-language-model/293655/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/beyond-dialogue-a-profile-dialogue-alignment-framework-towards-general-role-playing-language-model/293655.png","ImageObject",300,407,{"name":42,"@type":43},"Jasmine","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-11-12","2026-09-17",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",8,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What are the two key issues in current role-playing training?","Question",{"text":63,"@type":64},"The training can introduce bias when a predefined role profile conflicts with scenario dialogues. It can also lack sentence-level fine-grained alignment between profile traits and dialogue, leading to vague mapping.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"How does the BEYOND DIALOGUE framework address these issues?",{"text":68,"@type":64},"It introduces beyond-dialogue tasks to align dialogue with role profile traits per scenario. It also uses a prompting mechanism to generate reasoning data for sentence-level alignment.",{"name":70,"@type":61,"acceptedAnswer":71},"What evidence shows the benefit of the proposed approach?",{"text":72,"@type":64},"The document reports that experimental results improve adherence to role profiles and outperform many general and specialized role-playing baselines, with provided code and data.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},293655,1792041857,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,114,119,122,126],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":55,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":112,"slug":113},30,"research-report",{"id":115,"doc_module":4,"doc_module_name":25,"category_name":116,"show_sort_weight":117,"slug":118},9,"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":25,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":25,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":97,"slug":129},19,"General","general",{"code":4,"msg":82,"data":131},{"doc_id":79,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":55,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":139,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":145,"preview_page_count":4},2336478487870,"https://ap-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45","BEYOND DIALOGUE: A Profile-Dialogue Alignment Framework Towards  \nGeneral Role-Playing Language Model  \nYeyong Yu1 , Runsheng Yu3 , Haojie Wei2 , Zhanqiu Zhang2 , Quan Qian*1  \n1 School of Computer Engineering & Science, Shanghai University  \n2LIGHTSPEED  \n3The Hong Kong University of Science and Technology  \nCorrespondence: [qqian@shu.edu.cn](qqian@shu.edu.cn)  \nAbstract  \nThe rapid advancement of large language models (LLMs) has revolutionized role-playing, enabling the development of general role-playing models. However, current role-playing training has two significant issues: (I) Using a predefined role profile to prompt dialogue training for specific scenarios usually leads to biases and even conflicts between the dialogue and the profile, resulting in training biases. (II) Models learn to imitate the role based solely on the profile, neglecting profile-dialogue alignment at the sentence level. To overcome the aforementioned hurdles, we propose a novel framework BEYOND DIALOGUE, which introduces “beyond dialogue” tasks to align dialogue with profile traits for each scenario, eliminating biases during training. Furthermore, the framework achieves a sentence-level fine-grained alignment between profile and dialogue through an innovative prompting mechanism that generates reasoning data for training. Moreover, the aforementioned methods are fully automated and low-cost. Experimental results demonstrate our model excels in adhering to role profiles, outperforming most proprietary general and specialized role-playing baselines. The code and data are provided in [https://](https://)[ ](https://)[github.com/yuyouyu32/BeyondDialogue](github.com/yuyouyu32/BeyondDialogue).  \n1 Introduction  \nThe rapid advancement of large language models (LLMs) has demonstrated their significant potential in user interactions (Achiam et al., 2023 ; Yang et al., 2024 ; Dubey et al., 2024) . A particularly promising area is the development of roleplaying LLM agents, which are capable of simulating both real and fictional roles to deliver immersive and interactive experiences (Chen et al., 2024b,a) . Applications like Character AI 1 , Replika 2 , and Baichuan-NPC (Yang et al., 2023) exemplify the growing impact of these agents. However, the  \n1 [https://character.ai](https://character.ai)  \nRole-Playing Training Task  \nTraining dialogues do not directly reflect 'verbose' and'bossy' speaking styles, potentially introducing bias.  \nFigure 1: An example of the bias issue between predefined role profile and dialogues in role-playing training.  \nproprietary nature of these platforms restricts users’ability to customize specific role models. Thus, the development of an open-source, general roleplaying model is imperative.  \nAt the current stage of role-playing dialogue training tasks, role profiles are typically manually compiled (Shao et al., 2023) or generated through LLM summaries (Zhou et al., 2024), including information such as personality and speaking style as depicted throughout a whole novel or script (Zhou et al., 2024 ; Li et al., 2023) . However, a bias exists between the training dialogues extracted from individual scenarios and the predefined role profiles. As an illustration in Figure 1, Hermione’s speaking style is predefined into four categories, which are manually extracted from the entire content of the novel. However, the training dialogues in Figure 1 predominantly reflect only two styles (“bookish–  \n2 [https://replika.com](https://replika.com)  \n11992  \nProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages 11992–12022 July 27-August 1, 2025 ©2025 Association for Computational Linguistics  \nblue” and “encouraging–green”) . The bias arises because a single scenario often fails to fully capture the predefined profile, resulting in misalignment even contradiction between the prompted profile and the training dialogues. Such biases are common: from the HPD dataset published","cbCaitv3JKpxxxSp","https://ap.wps.com/l/cbCaitv3JKpxxxSp","pdf",8548251,31,"English","# Introduction\n## Role-Playing Training Task\n## Bias Between Profile and Dialogues\n## Inadequate Fine-Grained Alignment","[{\"question\":\"What are the two key issues in current role-playing training?\",\"answer\":\"The training can introduce bias when a predefined role profile conflicts with scenario dialogues. It can also lack sentence-level fine-grained alignment between profile traits and dialogue, leading to vague mapping.\"},{\"question\":\"How does the BEYOND DIALOGUE framework address these issues?\",\"answer\":\"It introduces beyond-dialogue tasks to align dialogue with role profile traits per scenario. It also uses a prompting mechanism to generate reasoning data for sentence-level alignment.\"},{\"question\":\"What evidence shows the benefit of the proposed approach?\",\"answer\":\"The document reports that experimental results improve adherence to role profiles and outperform many general and specialized role-playing baselines, with provided code and data.\"}]","BEYOND DIALOGUE - A Profile-Dialogue Alignment Framework Towards General Role-Playing Language Model | PDF",1789667092,78]