# K-FinHallu: A Hallucination Detection Benchmark for Multi-Turn RAG in Korean Finance

arXiv: https://arxiv.org/abs/2605.29523 · HTML: https://mlsanity.com/paper/2605.29523/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Hallucination detection in RAG, one of 15 directions in Retrieval-augmented generation.
- Builds on 19 of its 32 references that are in the map.
- First posted on arXiv 28 May 2026.

## Placed in

- Retrieval-augmented generation / [Hallucination detection in RAG](https://mlsanity.com/rag/hallucination-detection-in-rag/)

## Builds on (most cited first, 19 of 19)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Retrieval Augmentation Reduces Hallucination in Conversation](https://mlsanity.com/paper/2104.07567/) (2021)
- [CoQA: A Conversational Question Answering Challenge](https://mlsanity.com/paper/1808.07042/) (2018)
- [SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning](https://mlsanity.com/paper/2408.05517/) (2024)
- [RAGTruth: A Hallucination Corpus for Developing Trustworthy Retrieval-Augmented Language Models](https://mlsanity.com/paper/2401.00396/) (2023)
- [Fine-grained Hallucination Detection and Editing for Language Models](https://mlsanity.com/paper/2401.06855/) (2024)
- [FaithDial: A Faithful Benchmark for Information-Seeking Dialogue](https://mlsanity.com/paper/2204.10757/) (2022)
- [KMMLU: Measuring Massive Multitask Language Understanding in Korean](https://mlsanity.com/paper/2402.11548/) (2024)
- [MTRAG: A Multi-Turn Conversational Benchmark for Evaluating Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2501.03468/) (2025)
- [HAE-RAE Bench: Evaluation of Korean Knowledge in Language Models](https://mlsanity.com/paper/2309.02706/) (2023)
- [AlphaFin: Benchmarking Financial Analysis with Retrieval-Augmented Stock-Chain Framework](https://mlsanity.com/paper/2403.12582/) (2024)
- [EXAONE 4.0: Unified Large Language Models Integrating Non-reasoning and Reasoning Modes](https://mlsanity.com/paper/2507.11407/) (2025)
- [OmniEval: An Omnidirectional and Automatic RAG Evaluation Benchmark in Financial Domain](https://mlsanity.com/paper/2412.13018/) (2024)
- [Open Ko-LLM Leaderboard: Evaluating Large Language Models in Korean with Ko-H5 Benchmark](https://mlsanity.com/paper/2405.20574/) (2024)
- [Multi-Document Grounded Multi-Turn Synthetic Dialog Generation](https://mlsanity.com/paper/2409.11500/) (2024)

## Built on by (most cited first, 1 of 1)

- [Confidently Wrong: Detecting Hallucinations in Financial Question Answering from LLM Internal States](https://mlsanity.com/paper/2607.11414/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have advanced financial automation through Retrieval-Augmented Generation (RAG), yet hallucinations remain a critical barrier to deployment in high-stakes environments. Existing benchmarks focus on single-turn, English-centric tasks, leaving the multi-turn dynamics and linguistic-regulatory nuances of the Korean financial domain unaddressed. We introduce K-FinHallu, the first benchmark for hallucination detection in multi-turn Korean financial RAG. We construct multi-turn dialogues from authentic Korean financial documents and inject hallucinations under a proposed hierarchical taxonomy based on context answerability that explicitly accounts for justified abstention. Benchmarking frontier and open-source LLMs as hallucination detectors, we find that even the strongest models struggle with fine-grained financial diagnostics and refusal behavior. While fine-tuning an 8B model on our training split yields performance competitive with frontier LLMs, justified abstention remains the weakest axis across all evaluated models.