Struc-Bench: Are Large Language Models Really Good at Generating Complex Structured Data?

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Tang, Xiangru, Zong, Yiming, Phang, Jason, Zhao, Yilun, Zhou, Wangchunshu, Cohan, Arman, Gerstein, Mark
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866929303517659136
author Tang, Xiangru
Zong, Yiming
Phang, Jason
Zhao, Yilun
Zhou, Wangchunshu
Cohan, Arman
Gerstein, Mark
author_facet Tang, Xiangru
Zong, Yiming
Phang, Jason
Zhao, Yilun
Zhou, Wangchunshu
Cohan, Arman
Gerstein, Mark
contents Despite the remarkable capabilities of Large Language Models (LLMs) like GPT-4, producing complex, structured tabular data remains challenging. Our study assesses LLMs' proficiency in structuring tables and introduces a novel fine-tuning method, cognizant of data structures, to bolster their performance. We unveil Struc-Bench, a comprehensive benchmark featuring prominent LLMs (GPT-NeoX-20B, GPT-3.5, GPT-4, and Vicuna), which spans text tables, HTML, and LaTeX formats. Our proposed FormatCoT aids in crafting format-specific instructions from the intended outputs to populate this benchmark. Addressing the gap in task-centered evaluation, we propose two innovative metrics, P-Score (Prompting Score) and H-Score (Heuristical Score), to more accurately gauge LLM performance. Our experiments show that applying our structure-aware fine-tuning to LLaMA-7B leads to substantial performance gains, outshining its LLM counterparts across most measures. In-depth error analysis and creating an ability map across six dimensions -- coverage, formatting, reasoning, comprehension, pragmatics, and hallucination -- highlight areas for future enhancements and suggest forthcoming research trajectories. Our code and models can be found at https://github.com/gersteinlab/Struc-Bench.
format Preprint
id arxiv_https___arxiv_org_abs_2309_08963
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Struc-Bench: Are Large Language Models Really Good at Generating Complex Structured Data?
Tang, Xiangru
Zong, Yiming
Phang, Jason
Zhao, Yilun
Zhou, Wangchunshu
Cohan, Arman
Gerstein, Mark
Computation and Language
Despite the remarkable capabilities of Large Language Models (LLMs) like GPT-4, producing complex, structured tabular data remains challenging. Our study assesses LLMs' proficiency in structuring tables and introduces a novel fine-tuning method, cognizant of data structures, to bolster their performance. We unveil Struc-Bench, a comprehensive benchmark featuring prominent LLMs (GPT-NeoX-20B, GPT-3.5, GPT-4, and Vicuna), which spans text tables, HTML, and LaTeX formats. Our proposed FormatCoT aids in crafting format-specific instructions from the intended outputs to populate this benchmark. Addressing the gap in task-centered evaluation, we propose two innovative metrics, P-Score (Prompting Score) and H-Score (Heuristical Score), to more accurately gauge LLM performance. Our experiments show that applying our structure-aware fine-tuning to LLaMA-7B leads to substantial performance gains, outshining its LLM counterparts across most measures. In-depth error analysis and creating an ability map across six dimensions -- coverage, formatting, reasoning, comprehension, pragmatics, and hallucination -- highlight areas for future enhancements and suggest forthcoming research trajectories. Our code and models can be found at https://github.com/gersteinlab/Struc-Bench.
title Struc-Bench: Are Large Language Models Really Good at Generating Complex Structured Data?
topic Computation and Language
url https://arxiv.org/abs/2309.08963