CreativityPrism: A Holistic Evaluation Framework for Large Language Model Creativity

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hou, Zhaoyi Joey, Zhang, Bowei Alvin, Lu, Yining, Baghel, Bhiman Kumar, Brei, Anneliese, Lu, Ximing, Jiang, Meng, Brahman, Faeze, Chaturvedi, Snigdha, Chang, Haw-Shiuan, Khashabi, Daniel, Li, Xiang Lorraine
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910025257058304
author Hou, Zhaoyi Joey
Zhang, Bowei Alvin
Lu, Yining
Baghel, Bhiman Kumar
Brei, Anneliese
Lu, Ximing
Jiang, Meng
Brahman, Faeze
Chaturvedi, Snigdha
Chang, Haw-Shiuan
Khashabi, Daniel
Li, Xiang Lorraine
author_facet Hou, Zhaoyi Joey
Zhang, Bowei Alvin
Lu, Yining
Baghel, Bhiman Kumar
Brei, Anneliese
Lu, Ximing
Jiang, Meng
Brahman, Faeze
Chaturvedi, Snigdha
Chang, Haw-Shiuan
Khashabi, Daniel
Li, Xiang Lorraine
contents Creativity is often seen as a hallmark of human intelligence. While large language models (LLMs) are increasingly perceived as generating creative text, there is still no holistic and scalable framework to evaluate their creativity across diverse scenarios. Existing methods of LLM creativity evaluation either heavily rely on humans, limiting speed and scalability, or are fragmented across different domains and different definitions of creativity. To address this gap, we propose CREATIVITYPRISM, an evaluation analysis framework that consolidates eight tasks from three domains, divergent thinking, creative writing, and logical reasoning, into a taxonomy of creativity that emphasizes three dimensions: quality, novelty, and diversity of LLM generations. The framework is designed to be scalable with reliable automatic evaluation judges that have been validated against human annotations. We evaluate 17 state-of-the-art (SoTA) proprietary and open-sourced LLMs on CREATIVITYPRISM and find that while proprietary LLMs dominate creative writing and logical reasoning tasks by a 15% lead over open-sourced ones, they offer no significant advantage in divergent thinking, a domain much less explored in existing post-training regimes. Our analysis also shows that high performance in one creative dimension or domain rarely generalizes to others; specifically, novelty metrics often show weak or negative correlations with other metrics. This fragmentation confirms that a holistic, multi-dimensional framework like CREATIVITYPRISM is essential for meaningful assessment of LLM creativity.
format Preprint
id arxiv_https___arxiv_org_abs_2510_20091
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CreativityPrism: A Holistic Evaluation Framework for Large Language Model Creativity
Hou, Zhaoyi Joey
Zhang, Bowei Alvin
Lu, Yining
Baghel, Bhiman Kumar
Brei, Anneliese
Lu, Ximing
Jiang, Meng
Brahman, Faeze
Chaturvedi, Snigdha
Chang, Haw-Shiuan
Khashabi, Daniel
Li, Xiang Lorraine
Computation and Language
Artificial Intelligence
Creativity is often seen as a hallmark of human intelligence. While large language models (LLMs) are increasingly perceived as generating creative text, there is still no holistic and scalable framework to evaluate their creativity across diverse scenarios. Existing methods of LLM creativity evaluation either heavily rely on humans, limiting speed and scalability, or are fragmented across different domains and different definitions of creativity. To address this gap, we propose CREATIVITYPRISM, an evaluation analysis framework that consolidates eight tasks from three domains, divergent thinking, creative writing, and logical reasoning, into a taxonomy of creativity that emphasizes three dimensions: quality, novelty, and diversity of LLM generations. The framework is designed to be scalable with reliable automatic evaluation judges that have been validated against human annotations. We evaluate 17 state-of-the-art (SoTA) proprietary and open-sourced LLMs on CREATIVITYPRISM and find that while proprietary LLMs dominate creative writing and logical reasoning tasks by a 15% lead over open-sourced ones, they offer no significant advantage in divergent thinking, a domain much less explored in existing post-training regimes. Our analysis also shows that high performance in one creative dimension or domain rarely generalizes to others; specifically, novelty metrics often show weak or negative correlations with other metrics. This fragmentation confirms that a holistic, multi-dimensional framework like CREATIVITYPRISM is essential for meaningful assessment of LLM creativity.
title CreativityPrism: A Holistic Evaluation Framework for Large Language Model Creativity
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2510.20091