Levels of analysis for large language models
Clinical Snapshot
PICO Framework
| P — Population | Large language models and artificial intelligence systems |
| I — Intervention | Application of cognitive science methods using Marr's levels of analysis framework |
| C — Comparator | Traditional approaches to understanding AI systems |
| O — Outcomes | Framework for understanding behaviour and internal organization of large language models |
Bottom Line
This theoretical paper proposes applying David Marr's cognitive science framework to understand large language models through three levels of analysis: computational, algorithmic, and implementational. While not directly clinical, this work is relevant to healthcare professionals as AI systems become increasingly integrated into medical practice. The framework could inform how we evaluate and understand AI diagnostic tools, clinical decision support systems, and other healthcare AI applications. However, the work remains theoretical without empirical validation. For Australian healthcare, this contributes to the broader understanding needed for TGA evaluation of AI medical devices and development of appropriate governance frameworks for AI in clinical settings.
Key Findings
Effect Size: Not applicable - theoretical work
Primary Outcome: Theoretical framework for understanding large language models using cognitive science methods
Nnt Or Sensitivity: Framework applicability across multiple AI research domains
Clinical Application
High feasibility for research applications, limited direct clinical implementation Relevant to Australian AI research initiatives and TGA considerations for AI in healthcare, though not directly applicable to clinical practice guidelines AI researchers, cognitive scientists, and healthcare professionals working with AI systems
Abstract
Modern artificial intelligence systems, such as large language models, are increasingly powerful but also increasingly hard to understand. Recognizing this problem as analogous to the historical difficulties in understanding the human mind, we argue that methods developed in cognitive science can be useful for understanding large language models. We propose a framework for applying these methods based on the levels of analysis that David Marr proposed for studying information processing systems. By revisiting established cognitive science techniques relevant to each level and illustrating their potential to yield insights into the behaviour and internal organization of large language models, we aim to provide a toolkit for making sense of these new kinds of minds. This article is part of the theme issue 'World models in natural and artificial intelligence'.
References
- 1.Ku, A., Campbell, D., Bai, X., Geng, J., Liu, R., Marjieh, R., McCoy, R. T., Nam, A., Sucholutsky, I., Zhang, L., Zhu, J. Q., & Griffiths, T. (2026). Levels of analysis for large language models. Philosophical Transactions of the Royal Society A, 384(2274). https://doi.org/10.1098/rsta.2025.0012
Related Research
The Journal of rheumatology
Examining the Role of Wearables in Inflammatory Arthritis Care: A Narrative Literature Review
3 Aug 2026
Acta obstetricia et gynecologica Scandinavica
ChatGPT in urogynecology: Comparing large language model responses to human experts
3 Aug 2026
Sleep medicine
Concordance of wearable device sleep metrics with patient-reported sleep quality: A systematic review
3 Aug 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service