Bi'an: A Bilingual Benchmark and Model for Hallucination Detection in Retrieval-Augmented Generation

TL;DR

This paper introduces Bi'an, a bilingual benchmark and lightweight judge models for detecting hallucinations in Retrieval-Augmented Generation, improving evaluation and detection accuracy.

Contribution

It presents a novel bilingual benchmark dataset and fine-tuned lightweight judge models for better hallucination detection in RAG systems.

Findings

14B model outperforms larger baselines

Rivals state-of-the-art closed-source LLMs

Extensive evaluation on Bi'anBench

Abstract

Retrieval-Augmented Generation (RAG) effectively reduces hallucinations in Large Language Models (LLMs) but can still produce inconsistent or unsupported content. Although LLM-as-a-Judge is widely used for RAG hallucination detection due to its implementation simplicity, it faces two main challenges: the absence of comprehensive evaluation benchmarks and the lack of domain-optimized judge models. To bridge these gaps, we introduce \textbf{Bi'an}, a novel framework featuring a bilingual benchmark dataset and lightweight judge models. The dataset supports rigorous evaluation across multiple RAG scenarios, while the judge models are fine-tuned from compact open-source LLMs. Extensive experimental evaluations on Bi'anBench show our 14B model outperforms baseline models with over five times larger parameter scales and rivals state-of-the-art closed-source LLMs. We will release our data and models soon at https://github.com/OpenSPG/KAG.