Results for "instruction design"
System Prompt
Intermediate
A high-priority instruction layer setting overarching behavior constraints for a chat model.
SFT
Intermediate
Fine-tuning on (prompt, response) pairs to align a model with instruction-following behaviors.
Safety Filter
Intermediate
Automated detection/prevention of disallowed outputs (toxicity, self-harm, illegal instruction, etc.).
Zero-Shot Prompting
Intro
Task instruction without examples.
Human-in-the-Loop
Intermediate
System design where humans validate or guide model outputs, especially for high-stakes decisions.
Automation Bias
Intermediate
Tendency to trust automated suggestions even when incorrect; mitigated by UI design, training, and checks.
Emergent Competition
Advanced
Competition arises without explicit design.
Cognitive Architecture
Frontier
System-level design for general intelligence.
Natural Language Instruction
Frontier
Controlling robots via language.
Mechanism Design
Advanced
Designing systems where rational agents behave as desired.
Market Design
Advanced
Designing efficient marketplaces.