Preference Model

Definitions

1. A model trained to predict or score human preferences among candidate outputs, often used in alignment and reinforcement learning pipelines. Preference models help turn qualitative human judgments into a signal that systems can optimize against.

   In plain English: A model that predicts which outputs humans are likely to prefer.

   Example: "The team trained a preference model to score draft responses before selecting the final version shown to users."

Related Terms

• Context Extension
• Data Mixture
• Dense Model
• Frontier Model
• Human Preference
• Internal Representation
• Large Context Model
• Likelihood
• Model Alignment
• Preference Data