John Le

Augmenting Large Language Models with External Data Sources: A Systematic Review of Methodologies, Performance Metrics, and Information Fidelity

This review focuses specifically on Retrieval-Augmented Generation and fine-tuning, identifying them as the two dominant paradigms for integrating external knowledge: RAG for retrieval-based context injection and fine-tuning for parametric knowledge adaptation.