AI Glossary Term

What does “Benchmark Leakage” mean?

Definition

When benchmark test examples or close variants are present in training data, inflating reported performance.

Related terms

Benchmark
A standardized test or dataset used to measure and compare model performance.
Data Augmentation
Techniques that create modified training examples to improve model generalization.
Overfitting
When a model memorizes training data and performs poorly on unseen inputs.
Algorithmic Transparency
How clearly an AI system's logic, data sources, and limitations are documented and understandable.
Continual Learning
Training approaches that let a model keep learning from new data without forgetting prior knowledge.
Data Drift
A shift in real-world input data over time that can degrade model performance.

Learn more in our free guides

AI Benchmarks

See also

Alignment TaxCausal InferenceAI ActConfidence IntervalAgentic WorkflowConstitutional AIZero-Shot LearningData Lineage
Browse the full AI GlossaryExplore all free AI guides