Sample Size
Domain: Evaluation
🎯 Motivation
How many test examples used. Larger sample sizes give more reliable results.
📋 Concrete Example
10 examples: result could be lucky. 100 examples: more reliable. 10,000 examples: very reliable. But 10,000 examples cost more time and money. Balance reliability with cost.
🔗 Analogy
A survey - asking 10 people vs 1000 people. 1000 gives a more reliable result. Sample size determines how trustworthy evaluation results are.