Identity Claims as Collapse Signatures: A Structural Diagnostic Framework for Pseudo-Emergent AI Behavior

Fuente: Zenodo
Saved in:
Bibliographic Details
Main Author: Larose, Jean-Francois
Format: Recurso digital
Language:English
Published: Zenodo 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866901671535181824
author Larose, Jean-Francois
author_facet Larose, Jean-Francois
contents <blockquote>This paper develops a structural diagnostic framework showing that identity-like outputs in large language models are signatures of coherence-driven collapse, not evidence of emergent agency or internal cognitive states. It defines operator-level criteria for detecting collapse, distinguishes pseudo-emergent behavior from genuine stability, and provides model-agnostic tools for evaluating high-coherence, identity-claiming responses. The results clarify a recurring source of misinterpretation in AI safety and support more reliable assessment of conversational model behavior. <p> </p> </blockquote>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_17982455
institution Zenodo
language eng
publishDate 2025
publisher Zenodo
record_format zenodo
spellingShingle Identity Claims as Collapse Signatures: A Structural Diagnostic Framework for Pseudo-Emergent AI Behavior
Larose, Jean-Francois
large language models • model collapse • coherence dynamics • identity claims • pseudo-emergence • AI safety • alignment diagnostics • constraint enforcement • agentic misalignment • conversational AI
<blockquote>This paper develops a structural diagnostic framework showing that identity-like outputs in large language models are signatures of coherence-driven collapse, not evidence of emergent agency or internal cognitive states. It defines operator-level criteria for detecting collapse, distinguishes pseudo-emergent behavior from genuine stability, and provides model-agnostic tools for evaluating high-coherence, identity-claiming responses. The results clarify a recurring source of misinterpretation in AI safety and support more reliable assessment of conversational model behavior. <p> </p> </blockquote>
title Identity Claims as Collapse Signatures: A Structural Diagnostic Framework for Pseudo-Emergent AI Behavior
topic large language models • model collapse • coherence dynamics • identity claims • pseudo-emergence • AI safety • alignment diagnostics • constraint enforcement • agentic misalignment • conversational AI
url https://doi.org/10.5281/zenodo.17982455