Saved in:
Bibliographic Details
Main Authors: Cao, Qian, Wang, Xiting, Yuan, Yuzhuo, Liu, Yahui, Luo, Fang, Song, Ruihua
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.19236
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917231861956608
author Cao, Qian
Wang, Xiting
Yuan, Yuzhuo
Liu, Yahui
Luo, Fang
Song, Ruihua
author_facet Cao, Qian
Wang, Xiting
Yuan, Yuzhuo
Liu, Yahui
Luo, Fang
Song, Ruihua
contents Creativity evaluation remains a challenging frontier for large language models (LLMs). Current evaluations heavily rely on inefficient and costly human judgments, hindering progress in enhancing machine creativity. While automated methods exist, ranging from psychological testing to heuristic- or prompting-based approaches, they often lack generalizability or alignment with human judgment. To address these issues, we propose a novel pairwise-comparison framework for assessing textual creativity that leverages shared contextual instructions to improve evaluation consistency. We introduce CreataSet, a large-scale dataset with 100K+ human-level and 1M+ synthetic creative instruction-response pairs spanning diverse open-domain tasks. Through training on CreataSet, we develop an LLM-based evaluator named CrEval. CrEval demonstrates remarkable superiority over existing methods in alignment with human judgments. Experimental results underscore the indispensable significance of integrating both human and synthetic data to train highly robust evaluators, and showcase the practical utility of CrEval in boosting the creativity of LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2505_19236
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Evaluating Text Creativity across Diverse Domains: A Dataset and Large Language Model Evaluator
Cao, Qian
Wang, Xiting
Yuan, Yuzhuo
Liu, Yahui
Luo, Fang
Song, Ruihua
Computation and Language
Creativity evaluation remains a challenging frontier for large language models (LLMs). Current evaluations heavily rely on inefficient and costly human judgments, hindering progress in enhancing machine creativity. While automated methods exist, ranging from psychological testing to heuristic- or prompting-based approaches, they often lack generalizability or alignment with human judgment. To address these issues, we propose a novel pairwise-comparison framework for assessing textual creativity that leverages shared contextual instructions to improve evaluation consistency. We introduce CreataSet, a large-scale dataset with 100K+ human-level and 1M+ synthetic creative instruction-response pairs spanning diverse open-domain tasks. Through training on CreataSet, we develop an LLM-based evaluator named CrEval. CrEval demonstrates remarkable superiority over existing methods in alignment with human judgments. Experimental results underscore the indispensable significance of integrating both human and synthetic data to train highly robust evaluators, and showcase the practical utility of CrEval in boosting the creativity of LLMs.
title Evaluating Text Creativity across Diverse Domains: A Dataset and Large Language Model Evaluator
topic Computation and Language
url https://arxiv.org/abs/2505.19236