Back to Academy
Relevance 7/10Data and MetricsAdvanced6 min read
Benchmark Contamination
Benchmark contamination means evaluation data was seen during training or tuning.
Why it matters for annotators
It can produce misleadingly high evaluation scores.
Visual mental model
Benchmark data audit -> exposure check -> decontaminate.
Examples (bad vs good)
Scenario: Real annotation scenario involving Benchmark Contamination
Bad: Labeling quickly without applying project rubric.
Good: Applying rubric criteria, documenting rationale, and escalating uncertainty.
Common mistakes
- Skipping guideline details for edge cases.
- Applying inconsistent criteria across similar samples.
- Avoiding escalation even when uncertain.
Submission checklist
- Read the latest guideline update before each batch.
- Apply rubric dimensions explicitly in each decision.
- Escalate ambiguous items with concise rationale.