AI News HubLIVE
原文2 min read

UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs

UnpredictaBench is a new benchmark that tests the ability of large language models (LLMs) to capture true underlying distributions. As LLMs are increasingly used as substitutes for other entities (e.g., for humans in economic simulations), many models tend to collapse towards a single plausible answer, failing to capture the unpredictability of real systems. The benchmark includes 448 problems involving canonical statistical distributions, distributions induced by stochastic programs, and natural-language scenarios describing random processes. It uses the KS@N metric, based on the Kolmogorov-Smirnov test, to quantify how well model outputs approximate target distributions. Experiments show a wide spread in performance, with KS@100 scores ranging from near 0 to over 20%, and no model exceeding 40%. Adding reasoning can slightly improve scores but does not solve the issue. UnpredictaBench demonstrates that even simple distributional simulation remains challenging, marking a necessary first step toward using LLMs as stand-ins for complex systems.

SourcearXiv Computational LinguisticsAuthor: Amirhossein Abaskohi, Amirhossein Dabiriaghdam, Liang Luo, Ellie Dingqiao Wen, Lele Wang, Giuseppe Carenini, Peter West

[2606.06622] UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs

[Submitted on 4 Jun 2026]

Title:UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs

View a PDF of the paper titled UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs, by Amirhossein Abaskohi and 6 other authors

View PDF HTML (experimental)

Abstract:We introduce UnpredictaBench, an evaluation that tests the ability of large language models (LLMs) to capture true underlying distributions. As LLMs are increasingly used as substitutes for other entities (e.g., for humans in economic simulations), the tendency of many models to collapse towards a single plausible answer means a failure to capture the unpredictability of real systems. Recent work on improving output diversity is insufficient for this setting: simulation requires samples that are calibrated to a target distribution, not merely varied outputs. UnpredictaBench isolates a simplified but fundamental version of this problem: sampling outcomes from individual target distributions, including canonical statistical distributions, distributions induced by stochastic programs, and natural-language scenarios that describe random processes. We introduce 448 such problems together with KS@N, a general-purpose evaluation metric that quantifies how well a model outputs approximate black-box target distributions via the Kolmogorov-Smirnov statistical test. This is the rate at which we fail to reject model samples of size N against ground-truth samples, with larger N indicating greater difficulty. Tested across open and proprietary models, we find a large spread in distributional capabilities. For instance, when models generate samples of size 100 (KS@100, our standard metric), scores range from near 0 to over 20%. No model is able to achieve over 40% at KS@100, showing significant headroom in distributional sampling as a capability. Although adding reasoning can somewhat increase scores, we find no immediate solution for this issue. UnpredictaBench shows that even simple distributional simulation remains challenging, making it a necessary first step toward using LLMs as stand-ins for complex systems.

Subjects:

Computation and Language (cs.CL)

Cite as: arXiv:2606.06622 [cs.CL]

(or arXiv:2606.06622v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2606.06622

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Amirhossein Abaskohi [view email] [v1] Thu, 4 Jun 2026 18:18:02 UTC (4,191 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs, by Amirhossein Abaskohi and 6 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CL

new | recent | 2026-06

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

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?)

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?)

Author

Venue

Institution

Topic

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.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)