[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85031-en":3,"doc-seo-85031-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85031,1374391975076,"Riley","https://ap-avatar.wpscdn.com/avatar/14000253ca4ec9f6853?x-image-process=image/resize,m_fixed,w_180,h_180&k=1783305029341752051",8,"Research & Report","Persona Cartography: Charting Language Model Personality Traits in Weight Space","Large language models exhibit recurring behavioural patterns—personas—that shape generalisation and safety, yet reliable tools for decomposing, measuring, and controlling them remain limited. This work treats personas as coordinates in a behavioural-trait space using the OCEAN framework (Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism). Low-rank adapters amplify or suppress traits, evaluated via calibrated LLM judges, trait-specific benchmarks, and capability tests. Results show monotonic, near-additive trait control and preserved competence, with safety-relevant behaviour shifting along induced trait axes. An unsupervised psychometric pipeline further recovers interpretable factors from rollouts, connecting personality measurement, model editing, and safety.","Persona Cartography: Charting Language Model Personality Traits in Weight Space  \nLuke Baines 1  \n[luke.sid.baines@gmail.com](luke.sid.baines@gmail.com)  \nMariia Koroliuk 1 [maria.koroliuk@gmail.com](maria.koroliuk@gmail.com)  \nAnton Gonzalvez Hawthorne 1 [antonhawthorne2@gmail.com](antonhawthorne2@gmail.com)  \nIrakli Shalibashvili 1 [iraklishali@gmail.com](iraklishali@gmail.com)  \nClément Dumas2 [dumasclement2002@gmail.com](dumasclement2002@gmail.com)  \nKonstantinos Voudouris3  \n[konstantinos.voudouris@dsit.gov.uk](konstantinos.voudouris@dsit.gov.uk)  \nDavid Demitri Africa3 [david.demitri.africa@gmail.com](david.demitri.africa@gmail.com)  \narXiv :2607 .079 16v 1 [ cs .AI] 8 Jul 2026  \nAbstract  \nLarge language models exhibit recurring behavioural patterns—personas—that shape generalisation and safety, but we lack reliable tools for decomposing, measuring, and controlling them. Our central insight is to treat personas as positions ina space of behavioural traits, using the OCEAN framework to describe model personas in terms of Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. We train low-rank adapters to amplify or suppress individual traits, and evaluate their effects using an LLM-judge calibrated against a humanvalidated panel, trait-specific multiple-choice benchmarks, and standard capability evaluations. Across six models from three families (4B-32B), we find that each adapter moves its target trait largely monotonically with scale, combines approximately additively with other adapters to construct mixed personas, and preserves performance on capability benchmarks at moderate scales. We further show that the induced trait axes affect safety-relevant behaviour in downstream evaluations:  \nfor example, moving along neuroticism and agreeableness axes affects frustration and sycophancy respectively. We also introduce an unsupervised psychometric pipeline that recovers four interpretable behavioural factors (tone, initiative, didacticism, epistemic caution) from model rollouts. Persona control can then be considered in terms of learning, scaling, and composing traits in weight space, providing a bridge between personality measurement, model editing, and safety.  \n persona-cartography  \n persona-cartography/monorepo  \n1 Introduction  \nLarge language models (LLMs) do not only differ in what they know or whether they answer correctly. They also differ in how they behave, such as how much they defer, what stances they take on open questions, and so on [Huang et al., 2025, Zhang et al., 2025, Slama et al., 2026] . These recurring patterns in how models behave can be explained in terms of the persona that models learn during training, with these personas being key to explaining how and why models generalise the way they do [Marks et al., 2026] .  \n1LASR Labs; equal contribution. 2ENS Paris-Saclay and MATS 3UK AI Security Institute  \nPreprint.  \nComposable LoRA adapters for shaping LLM personas along OCEAN trait axes  \nFigure 1: Overview of the experimental setup and methodology. (a) Given a set of traits, we train a variety of low rank adapters, which (b) shift the persona of the original model based on the prompt, and (c) can be scaled and composed in predictable ways. (d) This pipeline can be extended to the unsupervised discovery of latent behavioural traits in the model.  \nExisting approaches to persona control fall between two imperfect extremes. On the one hand, prompting and activation-space interventions [Sun et al., 2024, Cao et al., 2024, Chen et al., 2025, Feng et al., 2026] can alter behaviour at inference time with high flexibility, but may be brittle to context and require continued intervention. On the other hand, full pre-and post-training can instill broad behavioural defaults [Li et al., 2026, Christiano et al., 2017, Ouyang et al., 2022, Tice et al., 2026, Maiya et al., 2025], but it is expensive and inflexible: a model may need different behavioural modes in different deployment contexts.","cbCaioCKidgMSqYT","https://ap.wps.com/l/cbCaioCKidgMSqYT","pdf",4924191,1,85,"English","en",105,"# Abstract\n# Introduction\n## Trait-space view of LLM personas\n## Composable LoRA approach\n# Methodology Overview\n## Experimental setup and scalability\n## Unsupervised trait discovery pipeline\n# Contributions and Findings","[{\"question\":\"How does the paper represent LLM personas?\",\"answer\":\"It models personas as positions in a behavioural-trait space, using the OCEAN framework to parameterize persona traits such as Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism.\"},{\"question\":\"What technique is used to control OCEAN traits in the models?\",\"answer\":\"The paper trains low-rank adapter modules (LoRAs) to amplify or suppress individual OCEAN traits, enabling trait scaling and composition in weight space.\"},{\"question\":\"How are the effects of trait adapters evaluated?\",\"answer\":\"Effects are measured using calibrated LLM judges aligned with a human-validated panel, trait-specific multiple-choice benchmarks, and standard capability evaluations to verify performance preservation.\"}]",1784200507,214,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"persona-cartography-charting-language-model-personality-traits-in-weight-space","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/persona-cartography-charting-language-model-personality-traits-in-weight-space/85031/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"How does the paper represent LLM personas?","Question",{"text":75,"@type":76},"It models personas as positions in a behavioural-trait space, using the OCEAN framework to parameterize persona traits such as Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What technique is used to control OCEAN traits in the models?",{"text":80,"@type":76},"The paper trains low-rank adapter modules (LoRAs) to amplify or suppress individual OCEAN traits, enabling trait scaling and composition in weight space.",{"name":82,"@type":73,"acceptedAnswer":83},"How are the effects of trait adapters evaluated?",{"text":84,"@type":76},"Effects are measured using calibrated LLM judges aligned with a human-validated panel, trait-specific multiple-choice benchmarks, and standard capability evaluations to verify performance preservation.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]