QABattle

AI & LLM Evals / Quiz

Precision, Recall & F1 for QA Classifiers

Your hallucination detector is itself a classifier. If you can't read a confusion matrix, you can't tell whether your evaluator works.

Difficulty
Easy
Format
Quiz
Points
100
Estimate
9 min

// MISSION BRIEF

Your Mission

Half of AI-eval work is evaluating classifiers: a safety filter, a hallucination detector, a spam gate. Judging them requires precision, recall, F1, and a confusion matrix. This quiz drills the fundamentals every AI tester must own.

Scoring: equal weight per question.

// FIRST CONTACT

Battle teaser

A safety classifier flags 100 items as harmful; 80 truly are. Its precision is:

  1. A80% of all harmful items in the dataset
  2. B80% (true positives / all flagged positives)
  3. CThe false-negative rate
  4. DUndefined without latency
Answers, scoring, hints, and the full battle stay sealed.

// SKILL TAGS

ai-evalsprecision-recallf1confusion-matrixclassification