Preference Ranking
Domain: Evaluation
🎯 Motivation
Humans compare two AI outputs and pick the better one. Used for RLHF training and evaluation.
📋 Concrete Example
Which is better? A: ‘Python is a programming language.’ B: ‘Python is a high-level, interpreted programming language known for readability.’ Humans pick B (more informative). Rankings train the model.
🔗 Analogy
A taste test between two colas - people pick the one they prefer. Preference ranking has humans pick the better AI response.