AI Fundamentals

Reward Model

A model trained to predict human preference over pairs of LLM outputs, used to provide the training signal in RLHF. Quality of the reward model bounds the quality of the RLHF-tuned policy — a limitation that DPO and Constitutional AI partly address.

Related terms

RLHF
Reinforcement Learning from Human Feedback — a training method where human preferences over pairs of outputs are used to train a reward model, which in turn shapes the base LLM via RL (typically PPO or DPO). Central to aligning modern chat assistants.
Alignment
The general problem of making an AI system behave according to its designers' or operators' intent, especially as capability grows. Practical alignment work spans RLHF, Constitutional AI, red-teaming, evaluation, and interpretability research.

Framework dimensions

Free tools

LLM Model Comparison

Whitepapers

LLM Cost Optimization Patterns

Next steps

Take the assessment— See where you stand

Where this fits

Reward Model is part of the AI Fundamentals vocabulary used in the Generative AI Maturity Framework. See the full glossary for the complete set of 149 defined terms, or take the free maturity assessment to see where your organisation stands.