StrucText-Eval: Evaluating Large Language Model's Reasoning Ability in Structure-Rich Text

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gu, Zhouhong, Ye, Haoning, Chen, Xingzhou, Zhou, Zeyang, Feng, Hongwei, Xiao, Yanghua
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916447118163968
author Gu, Zhouhong
Ye, Haoning
Chen, Xingzhou
Zhou, Zeyang
Feng, Hongwei
Xiao, Yanghua
author_facet Gu, Zhouhong
Ye, Haoning
Chen, Xingzhou
Zhou, Zeyang
Feng, Hongwei
Xiao, Yanghua
contents The effective utilization of structured data, integral to corporate data strategies, has been challenged by the rise of large language models (LLMs) capable of processing unstructured information. This shift prompts the question: can LLMs interpret structured data directly in its unstructured form? We propose an automatic evaluation data generation method for assessing LLMs' reasoning capabilities on structure-rich text to explore this. Our approach supports 8 structured languages and 29 tasks, generating data with adjustable complexity through controllable nesting and structural width. We introduce StrucText-Eval, a benchmark containing 5,800 pre-generated and annotated samples designed to evaluate how well LLMs understand and reason through structured text. StrucText-Eval is divided into two suites: a regular Test suite (3,712 samples) and a Test-Hard suite (2,088 samples), the latter emphasizing the gap between human and model performance on more complex tasks. Experimental results show that while open-source LLMs achieve a maximum accuracy of 74.9\% on the standard dataset, their performance drops significantly to 45.8\% on the harder dataset. In contrast, human participants reach an accuracy of 92.6\% on StrucText-Eval-Hard, highlighting LLMs' current limitations in handling intricate structural information. The benchmark and generation codes are open sourced in \url{https://github.com/MikeGu721/StrucText-Eval}
format Preprint
id arxiv_https___arxiv_org_abs_2406_10621
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle StrucText-Eval: Evaluating Large Language Model's Reasoning Ability in Structure-Rich Text
Gu, Zhouhong
Ye, Haoning
Chen, Xingzhou
Zhou, Zeyang
Feng, Hongwei
Xiao, Yanghua
Computation and Language
Artificial Intelligence
The effective utilization of structured data, integral to corporate data strategies, has been challenged by the rise of large language models (LLMs) capable of processing unstructured information. This shift prompts the question: can LLMs interpret structured data directly in its unstructured form? We propose an automatic evaluation data generation method for assessing LLMs' reasoning capabilities on structure-rich text to explore this. Our approach supports 8 structured languages and 29 tasks, generating data with adjustable complexity through controllable nesting and structural width. We introduce StrucText-Eval, a benchmark containing 5,800 pre-generated and annotated samples designed to evaluate how well LLMs understand and reason through structured text. StrucText-Eval is divided into two suites: a regular Test suite (3,712 samples) and a Test-Hard suite (2,088 samples), the latter emphasizing the gap between human and model performance on more complex tasks. Experimental results show that while open-source LLMs achieve a maximum accuracy of 74.9\% on the standard dataset, their performance drops significantly to 45.8\% on the harder dataset. In contrast, human participants reach an accuracy of 92.6\% on StrucText-Eval-Hard, highlighting LLMs' current limitations in handling intricate structural information. The benchmark and generation codes are open sourced in \url{https://github.com/MikeGu721/StrucText-Eval}
title StrucText-Eval: Evaluating Large Language Model's Reasoning Ability in Structure-Rich Text
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2406.10621