Less Is More? When Dataset Context Hurts LLM-Generated Dataset Descriptions
2026-06-01 • Databases
Databases
AI summaryⓘ
The authors studied how well large language models (LLMs) can create descriptions for datasets, which help people find and reuse data. They tested using different levels of information: just the dataset title, the title plus schema (like table structure), and adding actual data samples. They found that just having the schema often made descriptions worse, while adding data samples helped somewhat but did not improve overall quality. Their research gives advice on how to better use LLMs for generating dataset descriptions.
large language modelsdataset metadatadataset schemadata reusenatural language descriptiondata publishingablation studysemantic analysisEuropean data portaldata quality
Authors
Lisa-Yao Gan, Arunav Das, Johanna Walker, Klaus Diepold, Elena Simperl
Abstract
Dataset search and reuse are strongly constrained by the quality of metadata such as natural language descriptions, which are often sparse or inconsistent. Although large language models (LLMs) can generate such descriptions automatically, little empirical guidance exists on what makes a good dataset description and what dataset context LLMs actually need. We study these questions through a literature-grounded framework of dataset description quality and a large-scale ablation study using 252 datasets (1,336 CSV files) from the European data portal data.europa.eu. We generate descriptions with LLMs in a baseline scenario and two ablation scenarios: (1) using only dataset titles, (2) titles and schema, and (3) titles, schema and representative data, and evaluate them with an LLM-as-a- judge framework and a semantic descriptive attribute analysis grounded in our quality dimensions. Our results reveal a consis- tent schema penalty: table-schemas alone often degrade narrative quality, while representative data partially restores grounding without improving overall human-facing quality. We further show that different LLMs exhibit stable descriptive personas. These findings provide practical guidance for LLM-supported data publishing workflows.