Saved in:
Bibliographic Details
Main Authors: Zhang, Dongxu, Gangal, Varun, Lattimer, Barrett Martin, Yang, Yi
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.05474
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911948267847680
author Zhang, Dongxu
Gangal, Varun
Lattimer, Barrett Martin
Yang, Yi
author_facet Zhang, Dongxu
Gangal, Varun
Lattimer, Barrett Martin
Yang, Yi
contents Detecting hallucinations in large language model (LLM) outputs is pivotal, yet traditional fine-tuning for this classification task is impeded by the expensive and quickly outdated annotation process, especially across numerous vertical domains and in the face of rapid LLM advancements. In this study, we introduce an approach that automatically generates both faithful and hallucinated outputs by rewriting system responses. Experimental findings demonstrate that a T5-base model, fine-tuned on our generated dataset, surpasses state-of-the-art zero-shot detectors and existing synthetic generation methods in both accuracy and latency, indicating efficacy of our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2407_05474
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing Hallucination Detection through Perturbation-Based Synthetic Data Generation in System Responses
Zhang, Dongxu
Gangal, Varun
Lattimer, Barrett Martin
Yang, Yi
Artificial Intelligence
Computation and Language
Detecting hallucinations in large language model (LLM) outputs is pivotal, yet traditional fine-tuning for this classification task is impeded by the expensive and quickly outdated annotation process, especially across numerous vertical domains and in the face of rapid LLM advancements. In this study, we introduce an approach that automatically generates both faithful and hallucinated outputs by rewriting system responses. Experimental findings demonstrate that a T5-base model, fine-tuned on our generated dataset, surpasses state-of-the-art zero-shot detectors and existing synthetic generation methods in both accuracy and latency, indicating efficacy of our approach.
title Enhancing Hallucination Detection through Perturbation-Based Synthetic Data Generation in System Responses
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2407.05474