Finetune-RAG: Fine-Tuning Language Models to Resist Hallucination in Retrieval-Augmented Generation (2505.10792v2)

Abstract: Retrieval-Augmented Generation (RAG) has emerged as a powerful framework to improve factuality in LLMs by grounding their outputs in retrieved documents. However, ensuring perfect retrieval of relevant information remains challenging, and when irrelevant content is passed downstream to an LLM, it can lead to hallucinations. In this work, we propose Finetune-RAG, a simple and effective fine-tuning approach that features the first-of-its-kind RAG training dataset constructed to mimic real-world imperfections. Experimental results show that Finetune-RAG improves factual accuracy by 21.2% over the base model. We also propose Bench-RAG, an LLM-as-a-judge evaluation pipeline that stress tests models under realistic imperfect retrieval scenarios. Our codebase and dataset are fully open sourced for community use.