Test contamination
Benchmark content appearing in a model’s training or accessible context.
What Test contamination does not show
High scores then may measure memorization rather than general capability.
Terms this definition uses
Related terms in Measurement and evaluation
How a measurement earns trust: evaluation sets, error rates, calibration, reproducibility, and the biases that make a number look better than it is.
AI answer reading · name confusion · hedged answer · Benchmark · Evaluation set · Golden dataset · Ground truth · Human evaluation · LLM-as-judge · Pairwise evaluation · Pointwise evaluation · Rubric · Inter-rater agreement · Accuracy · F1 score · False positive rate · False negative · True positive · True negative · Calibration · Confidence score · Coverage · Abstention · Reproducibility · Repeatability · Variance · Confidence interval · Sample size · Sampling bias · Selection bias · Survivorship bias · Data leakage
Questions about Test contamination
What is Test contamination?
Benchmark content appearing in a model’s training or accessible context.
What does Test contamination not show or guarantee?
High scores then may measure memorization rather than general capability.
Which area of the glossary does Test contamination belong to?
Measurement and evaluation: How a measurement earns trust: evaluation sets, error rates, calibration, reproducibility, and the biases that make a number look better than it is.
See it in the full glossary · 772 terms across 19 areas.
Free browser extension
See this on any site you visit, in one click.
CrawlCheck for Chrome grades every site you visit A–F and shows, from your own browser, what search engines and AI assistants are actually served. Free, no account.