MétaCan
Menu
Back to cohort

A Reflective Architecture for LLM-Based Systems

2025· article· W4416184344 on OpenAlexaff
Parisa Salmani, Peter R. Lewis

Bibliographic record

Venuenot available
Typearticle
Language
FieldComputer Science
TopicExplainable Artificial Intelligence (XAI)
Canadian institutionsOntario Tech University
Fundersnot available
KeywordsReflection (computer programming)ArchitectureEvent (particle physics)Cognitive architectureReliability (semiconductor)TrustworthinessSystems architectureKey (lock)

Abstract

fetched live from OpenAlex

Modern artificial intelligence (AI) systems critically lack the ability to reflect on their own behavior, reasoning processes, and generated output. For example, despite recent advancements in large language models (LLMs), the appearance of reflection in such systems is a linguistic trick rather than a cognitive competence. This deficiency can cause significant challenges, particularly in contexts where safety and reliability are important, such as healthcare and security applications, and also in social situations where complex contexts drive expected behavior. Building on previous work in computational self-awareness and reflection in adaptive systems, in this paper, we propose a reflective agent architecture that incorporates formal models of social expectations and self-simulation mechanisms with LLMs This architecture enhances LLM-based systems to reflect on their decisions and outputs, evaluating alignment with expected behaviors. Furthermore, it enables agents to internally simulate potential actions and evaluate their consequences. Using the expectation event calculus (EEC), the system formally represents expectations, events, and derived outcomes, supporting systematic self-evaluation. Concurrently, self-simulation allows the agent to introspectively predict and analyze possible outcomes and refine its decision when necessary. Our results demonstrate enhanced alignment with human expectations, highlighting the architecture's promise of greater social sensitivity in complex scenarios that require robust and trustworthy AI interactions.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.001
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow), Scholarly communication
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Theoretical or conceptual · Consensus signal: none
GenreCandidate signal: Methods · Consensus signal: none
Teacher disagreement score0.902
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.001
Meta-epidemiology (narrow)0.0010.000
Meta-epidemiology (broad)0.0010.000
Bibliometrics0.0010.002
Science and technology studies0.0010.000
Scholarly communication0.0010.000
Open science0.0020.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.033
GPT teacher head0.328
Teacher spread0.295 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one teacher head, not a consensus.

Study designTheoretical or conceptual
Domainnot available
GenreMethods

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations1
Published2025
Admission routes1
Has abstractyes

Explore more

Same topicExplainable Artificial Intelligence (XAI)French-language works237,207