HalluMix: A Task-Agnostic, Multi-Domain Benchmark for Real-World Hallucination Detection

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Emery, Deanna, Goitia, Michael, Vargus, Freddie, Neagu, Iulia
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910924723453952
author Emery, Deanna
Goitia, Michael
Vargus, Freddie
Neagu, Iulia
author_facet Emery, Deanna
Goitia, Michael
Vargus, Freddie
Neagu, Iulia
contents As large language models (LLMs) are increasingly deployed in high-stakes domains, detecting hallucinated content$\unicode{x2013}$text that is not grounded in supporting evidence$\unicode{x2013}$has become a critical challenge. Existing benchmarks for hallucination detection are often synthetically generated, narrowly focused on extractive question answering, and fail to capture the complexity of real-world scenarios involving multi-document contexts and full-sentence outputs. We introduce the HalluMix Benchmark, a diverse, task-agnostic dataset that includes examples from a range of domains and formats. Using this benchmark, we evaluate seven hallucination detection systems$\unicode{x2013}$both open and closed source$\unicode{x2013}$highlighting differences in performance across tasks, document lengths, and input representations. Our analysis highlights substantial performance disparities between short and long contexts, with critical implications for real-world Retrieval Augmented Generation (RAG) implementations. Quotient Detections achieves the best overall performance, with an accuracy of 0.82 and an F1 score of 0.84.
format Preprint
id arxiv_https___arxiv_org_abs_2505_00506
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle HalluMix: A Task-Agnostic, Multi-Domain Benchmark for Real-World Hallucination Detection
Emery, Deanna
Goitia, Michael
Vargus, Freddie
Neagu, Iulia
Computation and Language
Artificial Intelligence
As large language models (LLMs) are increasingly deployed in high-stakes domains, detecting hallucinated content$\unicode{x2013}$text that is not grounded in supporting evidence$\unicode{x2013}$has become a critical challenge. Existing benchmarks for hallucination detection are often synthetically generated, narrowly focused on extractive question answering, and fail to capture the complexity of real-world scenarios involving multi-document contexts and full-sentence outputs. We introduce the HalluMix Benchmark, a diverse, task-agnostic dataset that includes examples from a range of domains and formats. Using this benchmark, we evaluate seven hallucination detection systems$\unicode{x2013}$both open and closed source$\unicode{x2013}$highlighting differences in performance across tasks, document lengths, and input representations. Our analysis highlights substantial performance disparities between short and long contexts, with critical implications for real-world Retrieval Augmented Generation (RAG) implementations. Quotient Detections achieves the best overall performance, with an accuracy of 0.82 and an F1 score of 0.84.
title HalluMix: A Task-Agnostic, Multi-Domain Benchmark for Real-World Hallucination Detection
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2505.00506