AI Navigate

CREATE:LLMの連想的創造性のテスト

arXiv cs.CL / 2026/3/11

Ideas & Deep AnalysisModels & Research

要点

  • CREATEは、新たに導入されたベンチマークであり、大規模言語モデルの連想的創造性、特に概念間に新規かつ意味のある結びつきを形成する能力を評価することを目的としている。
  • ベンチマークでは、生成された概念経路の特異性と多様性に基づいてモデルを評価し、より多くの異なる意味深い結びつきを生成したモデルに高い評価を与える。
  • 最先端モデルの評価では、探索空間の広さと回答の多様性により課題が困難であることが明らかになり、高い計算リソースを投じても著しい性能向上は保証されない。
  • 現在のクリエイティブなプロンプト手法は多少の改善をもたらすが、ベンチマークはAIモデルの連想的創造性向上に向けた新たなアプローチの必要性を示している。
  • CREATEは、モデルの創造的連想推論能力向上を目指すさらなる研究開発のための貴重な実験場を提供する。

Computer Science > Computation and Language

arXiv:2603.09970 (cs)
[Submitted on 10 Mar 2026]

Title:CREATE: Testing LLMs for Associative Creativity

View a PDF of the paper titled CREATE: Testing LLMs for Associative Creativity, by Manya Wadhwa and 4 other authors
View PDF HTML (experimental)
Abstract:A key component of creativity is associative reasoning: the ability to draw novel yet meaningful connections between concepts. We introduce CREATE, a benchmark designed to evaluate models' capacity for creative associative reasoning. CREATE requires models to generate sets of paths connecting concepts in a model's parametric knowledge. Paths should have high specificity (distinctiveness and closeness of the concept connection) and high diversity (dissimilarity from other paths), and models are scored more highly if they produce a larger set of strong, diverse paths. This task shares demands of real creativity tasks like hypothesis generation, including an extremely large search space, but enables collection of a sizable benchmark with objective answer grading. Evaluation of frontier models shows that the strongest models achieve higher creative utility than others, with the high multiplicity of answers and complexity of the search making benchmark saturation difficult to achieve. Furthermore, our results illustrate that thinking models are not always more effective on our task, even with high token budgets. Recent approaches for creative prompting give some but limited additional improvement. CREATE provides a sandbox for developing new methods to improve models' capacity for associative creativity.
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2603.09970 [cs.CL]
  (or arXiv:2603.09970v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2603.09970
Focus to learn more
arXiv-issued DOI via DataCite

Submission history

From: Manya Wadhwa [view email]
[v1] Tue, 10 Mar 2026 17:58:44 UTC (1,764 KB)
Full-text links:

Access Paper:

Current browse context:
cs.CL
< prev   |   next >
Change to browse by:
cs

References & Citations

export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo
Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
Links to Code Toggle
Papers with Code (What is Papers with Code?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos

Demos

Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers

Recommenders and Search Tools

Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.