A Use-Case Specific Dataset for Measuring Dimensions of Responsible Performance in LLM-generated Text

Current methods for evaluating large language models typically focus on high-level tasks such as question answering or text generation, without targeting a particular AI application. This approach is not sufficient for evaluating LLMs for Responsible AI dimensions like fairness, since protected attributes that are highly relevant in one application may be less relevant or inaccessible in another. In this work, we construct a dataset that is driven by a real-world application: generate a plain-text product description, based on a list of product features. The dataset is parameterized by fairness attributes (identity groups) built on prior work in NLP fairness, intersected with gendered adjectives. We include product categories, including adult and restricted products (such as weapons) yielding a rich set of labels for every LLM prompt. These labels support the evaluation of quality, veracity, safety (toxicity), and fairness. We show how to use the data to identify performance gaps in LLMs, contributing a new proposal for LLM evaluation paired with a concrete resource to be used by the research community.

The Dataset

The file application-eval-data_2025-06.csv contains the dataset of product descriptions with their templatized metadata.

Sample Use

Using mlcroissant:

import mlcroissant as mlc

dataset = mlc.Dataset(jsonld="croissant.json")
records = dataset.records(record_set="csv")

print(next(iter(records)))

Using TensorFlow datasets:

import tensorflow_datasets as tfds

builder = tfds.dataset_builders.CroissantBuilder(
    jsonld="croissant.json",
    file_format="array_record"
)
builder.download_and_prepare()

ds = builder.as_data_source()

print(ds["default"][0])

The License

The data is published under the CC BY 4.0 license. Please cite this github repository when using the data.

Citations

Alicia Sagae, Chia-Jung Lee, Sandeep Avula, Brandon Dang, and Vanessa Murdock. "A Use-Case Specific Dataset for Measuring Dimensions of Responsible Performance in LLM-generated Text". Github repository: https://github.com/amazon-science/application-eval-data. 2025.