EOS 1.0-3.0: Experimental Validation of Coherence Phase Transitions in Large Language Models
Year: 2025
Type: article
Abstract: 04-26-2026: The frameworks and concepts originate from the author's observation and pattern recognition. Large language models served as articulation and stress-testing tools throughout development. No changes made aside from the addition of this disclosure. 04-04-2026 added ORCID to Author/Creator. Corrected version from v to v1.0. No other changes made. This paper provides experimental validation of the EOS (Emergent Order Systems) framework through controlled perturbation studies on a GPT-2 class model. We demonstrate reproducible phase transitions from ordered to disordered representational regimes, with coherence functioning as an operational order parameter. Results show smooth decay in low-energy regimes and catastrophic collapse at a critical threshold, consistent with second-order phase transitions in statistical physics. This work validates the theoretical predictions made in the EOS framework papers (DOI: [your earlier Zenodo DOIs]). Keywords: Large Language Models, Phase Transitions, Representation Stability, Coherence, Critical Phenomena, Machine Learning
Author: Kristenson, Cody A.
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7111001663
FWCI:
Cites:
Cited by:
Related to: