Towards a Robust Retrieval-Based Summarization System

arXiv:2403.19889v1 Announce Type: cross
Abstract: This paper describes an investigation of the robustness of large language models (LLMs) for retrieval augmented generation (RAG)-based summarization tasks. While LLMs provide summarization capabilities, their performance in complex, real-world scenarios remains under-explored. Our first contribution is LogicSumm, an innovative evaluation framework incorporating realistic scenarios to assess LLM robustness during RAG-based summarization. Based on limitations identified by LogiSumm, we then developed SummRAG, a comprehensive system to create training dialogues and fine-tune a model to enhance …

abstract arxiv capabilities cs.ai cs.cl cs.ir cs.lg evaluation framework investigation language language models large language large language models llm llms paper performance rag retrieval retrieval augmented generation robust robustness summarization tasks type world