SORT-AI: A Structural Safety and Reliability Framework for Advanced AI Systems with Retrieval-Augmented Generation as a Diagnostic Testbed

Abstract excerpt

Large language models and related generative AI systems increasingly operate in safety-critical and high-impact settings, where reliability, alignment, and robustness under distribution shift are central concerns. While retrieval-augmented generation (RAG) has emerged as a practical mechanism for grounding model outputs in external knowledge, it does not by itself provide guarantees against system-level failure mo...

Topics

Open a Topic to create a Post that cites this publication.

• Algorithms
• Artificial Intelligence
• Biological Evolution
• Computer Simulation
• Decision Support Systems, Clinical
• Electronic Health Records
• Genetic Variation
• Genetics, Medical

Identifiers and source

Literature Corpus work

f7115f6e-d065-5100-9aea-3a3acd0454b9

DOI

10.20944/preprints202512.1345.v1