Skip to content
AI News HubLIVE
Original source2 min read

Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards

Summary

A new arXiv preprint from Stanford University and City University of Macau separates two concerns about benchmark contamination: inflated absolute scores versus reordered leaderboard rankings. Using paraphrase-controlled comparisons across 47 public models and 74 deliberately contaminated fine-tuned models on ARC, GSM8K, HellaSwag and MMLU, the authors find contamination is largely uniform—it lifts scores but barely changes rankings, with a 0.997 rank correlation between standard and paraphrase-controlled leaderboards. They recommend reporting paraphrase-controlled rankings with confidence intervals.

SourcearXiv Computational LinguisticsAuthor: Xingyao Xiao (Stanford University), Yihong Cheng (City University of Macau)
Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 5 Jul 2026]

Title:Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards

View a PDF of the paper titled Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards, by Xingyao Xiao (Stanford University) and Yihong Cheng (City University of Macau)

View PDF

Abstract:Benchmark contamination, the leakage of test items into training data, is widely described as a threat to the reliability of large language model (LLM) leaderboards. We argue that this concern conflates two distinct questions: whether contamination inflates absolute scores, and whether it reorders the ranking of models. We recast contamination as a violation of anchor-item invariance and measure it through the differential functioning of original versus semantically equivalent paraphrased items, a within-item contrast that holds the measured skill fixed and isolates memorization from capability. Using per-instance responses from 47 publicly released models and 74 models finetuned with a known dose of contamination, across four benchmarks (ARC, GSM8K, HellaSwag, MMLU), we first calibrate the measure against ground truth: it recovers injected contamination dose-responsively (a corrected effect of +0.187 accuracy points for test-set leakage) and never flags a negative-control model trained only on the legitimate training split (-0.012). We then quantify leaderboard impact: the rank correlation between a standard leaderboard and a paraphrase-controlled leaderboard is 0.997, and a sensitivity analysis shows that the observed differential contamination is far below the level needed to move rankings, with only 3 of 188 model-by-benchmark cases showing differential contamination corroborated across two references. Contamination among these public models is therefore largely uniform: it inflates absolute scores without reordering the leaderboard, and ranking distortion requires the rare case of differential contamination. We provide a calibrated invariance audit, released as a reference implementation, and recommend that leaderboards report paraphrase-controlled rankings alongside confidence intervals.

Comments: 21 pages, 4 figures, 3 tables. Code and data: this https URL

Subjects:

Computation and Language (cs.CL); Applications (stat.AP); Methodology (stat.ME)

Cite as: arXiv:2609.02899 [cs.CL]

(or arXiv:2609.02899v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2609.02899

arXiv-issued DOI via DataCite

Submission history

From: Xingyao Xiao [view email] [v1] Sun, 5 Jul 2026 00:55:18 UTC (91 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards, by Xingyao Xiao (Stanford University) and Yihong Cheng (City University of Macau)

View PDF

TeX Source

view license

Current browse context:

cs.CL

new | recent | 2026-09

Change to browse by:

cs stat stat.AP stat.ME

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?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • The paper distinguishes contamination inflating absolute scores from contamination reordering model rankings.
  • A paraphrase-based invariance measure reliably detects injected contamination dose-responsively and does not flag a negative-control model.
  • Across 47 public models and four benchmarks, only 3 of 188 model-by-benchmark cases show corroborated differential contamination.
  • The authors recommend leaderboards report paraphrase-controlled rankings with confidence intervals.

Highlights and analysis are generated automatically and may contain errors. Check the original source.

More recent stories on this topic · Models