Context-Adaptive Synthesis and Compression for Enhanced Retrieval-Augmented Generation in Complex Domains

Peiran Zhou, Junnan Zhu, Yichen Shen +1

Large Language Models (LLMs) excel in language tasks but are prone to hallucinations and outdated knowledge. Retrieval-Augmented Generation (RAG) mitigates these by grounding LLMs…