Saved in:
Bibliographic Details
Main Authors: Bhamidipati, Patanjali, Malladi, Advaith, Shrivastava, Manish, Mamidi, Radhika
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.12244
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914719038701568
author Bhamidipati, Patanjali
Malladi, Advaith
Shrivastava, Manish
Mamidi, Radhika
author_facet Bhamidipati, Patanjali
Malladi, Advaith
Shrivastava, Manish
Mamidi, Radhika
contents In recent studies, the extensive utilization of large language models has underscored the importance of robust evaluation methodologies for assessing text generation quality and relevance to specific tasks. This has revealed a prevalent issue known as hallucination, an emergent condition in the model where generated text lacks faithfulness to the source and deviates from the evaluation criteria. In this study, we formally define hallucination and propose a framework for its quantitative detection in a zero-shot setting, leveraging our definition and the assumption that model outputs entail task and sample specific inputs. In detecting hallucinations, our solution achieves an accuracy of 0.78 in a model-aware setting and 0.61 in a model-agnostic setting. Notably, our solution maintains computational efficiency, requiring far less computational resources than other SOTA approaches, aligning with the trend towards lightweight and compressed models.
format Preprint
id arxiv_https___arxiv_org_abs_2403_12244
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Zero-Shot Multi-task Hallucination Detection
Bhamidipati, Patanjali
Malladi, Advaith
Shrivastava, Manish
Mamidi, Radhika
Computation and Language
In recent studies, the extensive utilization of large language models has underscored the importance of robust evaluation methodologies for assessing text generation quality and relevance to specific tasks. This has revealed a prevalent issue known as hallucination, an emergent condition in the model where generated text lacks faithfulness to the source and deviates from the evaluation criteria. In this study, we formally define hallucination and propose a framework for its quantitative detection in a zero-shot setting, leveraging our definition and the assumption that model outputs entail task and sample specific inputs. In detecting hallucinations, our solution achieves an accuracy of 0.78 in a model-aware setting and 0.61 in a model-agnostic setting. Notably, our solution maintains computational efficiency, requiring far less computational resources than other SOTA approaches, aligning with the trend towards lightweight and compressed models.
title Zero-Shot Multi-task Hallucination Detection
topic Computation and Language
url https://arxiv.org/abs/2403.12244