Refusal Training

Definitions

1. Training or tuning methods intended to improve when and how an AI system refuses unsafe, disallowed, or unsupported requests. Refusal training is part of broader alignment and safety work.

   In plain English: Training an AI system to refuse unsafe or disallowed requests appropriately.

   Example: "They expanded refusal training after the model kept offering too much detail on prohibited dual-use chemistry questions."

Related Terms

• AI Sandbox
• Interpretability
• Jailbreak AI
• Next Token Prediction
• Noise Injection
• Objective Function
• Preference Data
• Red Team AI
• Reward Signal
• Safety Training
• Training Curriculum
• Training Distribution
• Training Recipe