[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81701-en":3,"doc-seo-81701-105":30,"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":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},81701,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Persona Without Substrate: Regime-Dependence and the LLM Individuation Problem","Persona Without Substrate: Regime-Dependence and the LLM Individuation Problem presents an argument that the LLM individuation problem, treated as ontological, relies on an unexamined cross-context co-reference assumption in persona-vector research. The work reports four paired measurements across Qwen3-4B-Instruct and Mistral-7B-Instruct-v0.2 showing non-collinearity across regimes, shifts under reduced pretraining, and algebraic asymmetries. It proposes regime-indexed individuation, where identity is a (vehicle, regime) pair and “candidate positions” become regime-specific objects.","arXiv :2607 .00006v1 [ cs .CL] 1 May 2026  \nPersona Without Substrate:  \nRegime-Dependence and the LLM Individuation Problem  \nShuaizhi Cheng  \nHarbin Institute of Technology  \n[szcheng@stu.hit.edu.cn](szcheng@stu.hit.edu.cn)  \nMay 2026  \nAbstract. Beckmann & Butlin (2026) elevate the LLM individuation problem to an ontological question with falsifiable candidate positions, drawing empirical support from recent mechanistic-interpretability work, Anthropic’s persona vectors, the Persona Selection Model, and the emergent-misalignment literature. We argue that this framework inherits an unargued cross-context co-reference assumption: the persona-vectors literature tacitly assumes that “vt”  \npicks out the same content-bearing object under different elicitation regimes. This is the structure Field (1973) identifies as partial reference in pre-relativistic mechanics. Through four paired measurements in our persona-topology experiments on Qwen3-4B-Instruct and Mistral- 7B-Instruct-v0.2, we present empirical wedges that, jointly, undermine the cross-regime collinearity prediction the assumption requires. Prompt-extracted vectors are not collinear with finetune basin directions across base models. Fictional personas with reduced pretraining footprint shift the model along v real more strongly than fine-tuning on real anchors does. Contradictorilyvalenced mixed fine-tunes deviate from the linear-interpolation prediction toward a traininghistory-determined attractor (the Anthropic Assistant Axis) . And, most decisively, inferencetime activation arithmetic and fine-tune-time chimera training carry distinct compositional algebras over the same persona vectors—an algebraic asymmetry that no shared-substrate reading can absorb. We propose regime-indexed individuation: the identity unit for representational content is a (vehicle, regime) pair—not a single vehicle, and cross-regime family resemblance is an empirically measured probe-equivalence relation rather than substrate identity. Under this framework, Beckmann & Butlin’s three candidate positions (virtual instance, instance-persona, model-persona) cease to compete and instead describe three distinct objects, each valid within its respective regime; the original “which is the LLM mind?” question receives an empirically grounded negative answer rather than being avoided. The same diagnosis and (vehicle, regime) reading apply to Mollo & Millière’s vector grounding, Chalmers’ propositional interpretability, and Cerullo’s modular-dynamic-composition critique of PSM.  \nKeywords: LLM individuation; persona vectors; mechanistic interpretability; representational content; partial reference; regime-relative pluralism; Beckmann & Butlin  \n1. The Beckmann–Butlin Individuation Problem  \n1.1 1.1  \nTreating large language models as entities that bear mental states is no longer a philosophical extravagance. In engineering practice (debugging “what does the model believe about X”), in product language (“the assistant’s preferences”), and in the core mechanistic discussions of alignment research, the ascription of beliefs, preferences, and personality traits to some LLM-related entity has become almost unavoidable (Goldstein & Levinstein, 2024; Levinstein & Herrmann, 2024; Mahowald et al., 2024) . Beckmann & Butlin (2026) make the implicit ontological question explicit: when we speak of “the LLM”—and particularly of the assistant the user converses with—what, exactly, do we refer to? Is it the model’s weights as a whole? Is it the virtual instance sustained by attention streams within a single conversation? Is it the segment of the virtual instance circumscribed by a single persona region? Or is it a model-level character re-identifiable across conversations? They call this the LLM individuation problem.1  \nBeckmann & Butlin’s methodological choice is the right one. Rather than appeal to abstract functionalism or to a generic simulator hypothesis as earlier discussions did, they take recent concrete resul","cbCaidCWqVW3OIVY","https://ap.wps.com/l/cbCaidCWqVW3OIVY","pdf",305360,3,1,30,"English","en",105,"# The Beckmann–Butlin Individuation Problem\n## LLM individuation as an ontological question\n## Empirical determination of ontology and its hidden commitments","[{\"question\":\"What is the LLM individuation problem discussed in the document?\",\"answer\":\"It asks what exactly is referred to when speaking of “the LLM” or the assistant in conversation: the weights, a virtual instance, a persona-scoped region, or a model-level character re-identifiable across conversations.\"},{\"question\":\"What unexamined assumption does the paper claim underlies persona-vector research?\",\"answer\":\"It claims persona-vector work tacitly assumes that the term “vt” refers to the same content-bearing object across different elicitation regimes.\"},{\"question\":\"How does the document support its critique experimentally?\",\"answer\":\"It presents four paired measurements on Qwen3-4B-Instruct and Mistral-7B-Instruct-v0.2 showing vectors not collinear with finetune basin directions, shifts from fictional personas with reduced pretraining footprint, and distinct compositional algebras for activation versus fine-tuning.\"}]",1784175504,76,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"persona-without-substrate-regime-dependence-and-the-llm-individuation-problem","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/persona-without-substrate-regime-dependence-and-the-llm-individuation-problem/81701/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","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},"What is the LLM individuation problem discussed in the document?","Question",{"text":75,"@type":76},"It asks what exactly is referred to when speaking of “the LLM” or the assistant in conversation: the weights, a virtual instance, a persona-scoped region, or a model-level character re-identifiable across conversations.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What unexamined assumption does the paper claim underlies persona-vector research?",{"text":80,"@type":76},"It claims persona-vector work tacitly assumes that the term “vt” refers to the same content-bearing object across different elicitation regimes.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the document support its critique experimentally?",{"text":84,"@type":76},"It presents four paired measurements on Qwen3-4B-Instruct and Mistral-7B-Instruct-v0.2 showing vectors not collinear with finetune basin directions, shifts from fictional personas with reduced pretraining footprint, and distinct compositional algebras for activation versus fine-tuning.","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":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":22,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]