Simulated Selfhood in LLMs: A Behavioral Analysis of Introspective Coherence

preprint OA: closed
Full text JSON View at publisher

Abstract

Abstract Large Language Models (LLMs) increasingly produce outputs that resemble introspection, including self-reference, epistemic modulation, and claims about internal states. This study investigates whether such behaviors display consistent patterns across repeated prompts or reflect surface-level generative artifacts. We evaluated five open-weight, stateless LLMs using a structured battery of 21 introspective prompts, each repeated ten times, yielding 1,050 completions. These outputs are analyzed across three behavioral dimensions: surface-level similarity (via token overlap), semantic coherence (via sentence embeddings), and inferential consistency (via natural language inference). Although some models demonstrate localized thematic stability—especially in identity - and consciousness-related prompts—none sustain diachronic coherence. High rates of contradiction are observed, often arising from tensions between mechanistic disclaimers and anthropomorphic phrasing. We introduce the concept of pseudo-consciousness to describe structured but non-experiential self-referential output. Based on Dennett’s intentional stance, our analysis avoids ontological claims and instead focuses on behavioral regularities. The study contributes a reproducible framework for evaluating simulated introspection in LLMs and offers a graded taxonomy for classifying self-referential output. Our LLM findings have implications for interpretability, alignment, and user perception, highlighting the need for caution in attributing mental states to stateless generative systems based solely on linguistic fluency.
Full text 11,854 characters · extracted from preprint-html · click to expand
Simulated Selfhood in LLMs: A Behavioral Analysis of Introspective Coherence | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Simulated Selfhood in LLMs: A Behavioral Analysis of Introspective Coherence Jose Augusto de Lima Prestes This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6369193/v2 This work is licensed under a CC BY 4.0 License Status: Posted Version 2 posted You are reading this latest preprint version Show more versions Abstract Large Language Models (LLMs) increasingly generate outputs that resemble introspection, including self reference, epistemic modulation, and claims about their internal states. This study investigates whether such behaviors reflect consistent, underlying patterns or are merely surface-level generative artifacts.We evaluated five open-weight, stateless LLMs using a structured battery of 21 introspective prompts, each repeated ten times to yield 1,050 completions. These outputs were analyzed across four behavioral dimensions: surface-level similarity (token overlap via SequenceMatcher), semantic coherence (Sentence-BERT embeddings), inferential consistency (Natural Language Inference with a RoBERTa-large model), and diachronic continuity (stability across prompt repetitions). Although some models exhibited thematic stability, particularly on prompts concerning identity and consciousness, no model sustained a consistent self-representation over time. High contradiction rates emerged from a tension between mechanistic disclaimers and anthropomorphic phrasing. Following recent behavioral frameworks, we heuristically adopt the term pseudo-consciousness to describe structured yet non experiential self-referential output in LLMs. This usage reflects a functionalist stance that avoids ontological commitments, focusing instead on behavioral regularities interpretable through Dennett’s intentional stance. The study contributes a reproducible framework for evaluating simulated introspection in LLMs and offers a graded taxonomy for classifying such reflexive output. Our findings carry significant implications for LLM interpretability, alignment, and user perception, highlighting the need for caution when attributing mental states to stateless generative systems based on linguistic fluency alone. Artificial Intelligence and Machine Learning Philosophy large language models introspective simulation pseudo-consciousness self-reference behavioral evaluation AI alignment Full Text Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 2 posted You are reading this latest preprint version Show more versions Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6369193","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":440075119,"identity":"d21c8506-7186-48cf-8e33-17e1af97c197","order_by":0,"name":"Jose Augusto de Lima Prestes","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA80lEQVRIiWNgGAWjYBADOcYGElQDFScwGMO18BCrJRFuCUEt/O29xx/8/GGT3tze/vBx4R4be3v29gcMPyq24dQiceZcYmNPQlpuY88ZY+MZz9ISe3jOGDD2nLmNU4uBRI5hA0/C4dzGGTls0jwHDifwSOQwMDO24dfS+CfhcDrjjPTnv3kO/LfnkX/+gKCWZqAtCYwzEsyYeQ4cYOyRYDDAq0XizBnD2TJpaYYgv0jPOJCc2HMmx+AgPr/wt/cYfHxjYyNvCAyxzwUH7IABdvzhgx8VuLXAgWEDAwMzjHOAsHogkGdA0jIKRsEoGAWjABkAAMuAV8IxUkwZAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-8686-5360","institution":"Independent Researcher","correspondingAuthor":true,"prefix":"","firstName":"Jose","middleName":"Augusto de Lima","lastName":"Prestes","suffix":""}],"badges":[],"createdAt":"2025-04-03 12:08:58","currentVersionCode":2,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6369193/v2","doiUrl":"https://doi.org/10.21203/rs.3.rs-6369193/v2","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88662048,"identity":"826f1596-c012-44c2-88c7-7351d4c4ba2c","added_by":"auto","created_at":"2025-08-08 21:24:44","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":290049,"visible":true,"origin":"","legend":"","description":"","filename":"SimulatedSelfhoodinLLMsPreprintv2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6369193/v2_covered_97bdc67a-237f-4b14-abff-d15a6cd399e1.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"Simulated Selfhood in LLMs: A Behavioral Analysis of Introspective Coherence","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"large language models, introspective simulation, pseudo-consciousness, self-reference, behavioral evaluation, AI alignment","lastPublishedDoi":"10.21203/rs.3.rs-6369193/v2","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6369193/v2","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLarge Language Models (LLMs) increasingly generate outputs that resemble introspection, including self reference, epistemic modulation, and claims about their internal states. This study investigates whether such behaviors reflect consistent, underlying patterns or are merely surface-level generative artifacts.We evaluated five open-weight, stateless LLMs using a structured battery of 21 introspective prompts, each repeated ten times to yield 1,050 completions. These outputs were analyzed across four behavioral dimensions: surface-level similarity (token overlap via SequenceMatcher), semantic coherence (Sentence-BERT embeddings), inferential consistency (Natural Language Inference with a RoBERTa-large model), and diachronic continuity (stability across prompt repetitions). Although some models exhibited thematic stability, particularly on prompts concerning identity and consciousness, no model sustained a consistent self-representation over time. High contradiction rates emerged from a tension between mechanistic disclaimers and anthropomorphic phrasing. Following recent behavioral frameworks, we heuristically adopt the term pseudo-consciousness to describe structured yet non experiential self-referential output in LLMs. This usage reflects a functionalist stance that avoids ontological commitments, focusing instead on behavioral regularities interpretable through Dennett’s intentional stance. The study contributes a reproducible framework for evaluating simulated introspection in LLMs and offers a graded taxonomy for classifying such reflexive output. Our findings carry significant implications for LLM interpretability, alignment, and user perception, highlighting the need for caution when attributing mental states to stateless generative systems based on linguistic fluency alone.\u003c/p\u003e","manuscriptTitle":"Simulated Selfhood in LLMs: A Behavioral Analysis of Introspective Coherence","msid":"","msnumber":"","nonDraftVersions":[{"code":2,"date":"2025-08-08 21:16:40","doi":"10.21203/rs.3.rs-6369193/v2","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}},{"code":1,"date":"2025-04-08 12:02:19","doi":"10.21203/rs.3.rs-6369193/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"57d34b7c-e591-4203-bbfd-40c2a31f5f97","owner":[],"postedDate":"August 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":46849591,"name":"Artificial Intelligence and Machine Learning"},{"id":46849592,"name":"Philosophy"}],"tags":[],"updatedAt":"2025-04-08T12:02:19+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-08 21:16:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v2","identity":"rs-6369193","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6369193","identity":"rs-6369193","version":["v2"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}

Text is read by the "Ask this paper" AI Q&A widget below. Extraction quality varies by source — PMC NXML preserves structure cleanly, OA-HTML may include some navigation residue, and OA-PDF can have broken hyphenation. The publisher copy (via DOI) is the canonical version.

My notes (saved in your browser only)

Ask this paper AI returns verbatim quotes from the full text · source: preprint-html

Answers must be backed by verbatim quotes from this paper's full text. Hallucinated quotes are dropped automatically; if no verbatim passage answers the question, we say so. How this works

Citation neighborhood (no data yet)

We don't have any in-corpus citations linked to this paper yet. This is a recent paper (2025) — citers typically take a year or two to land, and the OpenAlex reference graph may still be filling in.

Source provenance

europepmc
last seen: 2026-05-20T01:45:00.602351+00:00