Understanding generative AI output with embedding models
Science Advances · 28.11.2025 · 11 (48)
DOI: 10.1126/sciadv.adx4082 ↗Kurz gefasst
Fachartikel in der Zeitschrift Science Advances, erschienen am 28.11.2025. Verfasst von Max Vargas und 4 weiteren Autor:innen. Beteiligte Einrichtung: Pacific Northwest National Laboratory, Richland, WA 99354, USA. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 101 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: U.S. Department of Energy.
Abstract (Originalsprache)
Constructing high-quality features is critical to any quantitative data analysis. While feature engineering was historically addressed by carefully handcrafting data representations on the basis of domain expertise, deep neural networks (DNNs) now offer a radically different approach. DNNs implicitly engineer features by transforming their input data into hidden feature vectors called embeddings. For embedding vectors produced by foundation models—which are trained to be useful across many contexts—we demonstrate that simple and well-studied dimensionality-reduction techniques such as principal components analysis uncover inherent heterogeneity in input data concordant with human-understandable explanations. Of the many applications for this framework, we find empirical evidence that there is intrinsic separability between real samples and those generated by artificial intelligence.
Autor:innen und Institutionen
- Pacific Northwest National Laboratory, Richland, WA 99354, USA.
- Department of Electrical and Computer Engineering, Rutgers University, New Brunswick, NJ 08854, USA.
- Department of Mathematics, University of Washington, Seattle, WA 98195, USA.
ORCID-Profile: Max Vargas, Reilly Cannon, Andrew Engel, Anand D Sarwate, Tony Chiang
Förderung
- U.S. Department of Energy
Zitieren
APA
Vargas, M., Cannon, R., Engel, A., Sarwate, A. D., & Chiang, T. (2025). Understanding generative AI output with embedding models. Science Advances, 11(48). https://doi.org/10.1126/sciadv.adx4082
Vancouver
Vargas M, Cannon R, Engel A, Sarwate AD, Chiang T. Understanding generative AI output with embedding models. Science Advances. 2025;11(48). doi:10.1126/sciadv.adx4082
BibTeX
@article{Vargas2025,
author = {Vargas, Max and Cannon, Reilly and Engel, Andrew and Sarwate, Anand D and Chiang, Tony},
title = {Understanding generative AI output with embedding models},
journal = {Science Advances},
year = {2025},
volume = {11},
number = {48},
doi = {10.1126/sciadv.adx4082}
}
EUVIMED