The authors argue that time-bound AI competitions with hidden test data are the gold standard for generative AI evaluation, and that public static benchmarks should be considered contaminated once published.
State of M achine L earning C ompetitions in 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Position: AI Competitions Provide the Gold Standard for Empirical Rigor in GenAI Evaluation
The authors argue that time-bound AI competitions with hidden test data are the gold standard for generative AI evaluation, and that public static benchmarks should be considered contaminated once published.