Model Adaptation · Data artifact

Instruction Demonstration Dataset

Data artifactModel AdaptationModelsarc:InstructionDemonstrationDataset

A curated training dataset of instructions paired with high-quality responses demonstrating the desired style, tone and approach, used for supervised fine-tuning before preference optimization.

Responsibility. Supplies direct positive examples of ideal instruction-following behaviour to the SFT stage.

Also known as: SFT demonstration set, Demonstrations of ideal behavior

Direct neighbourhood (hover for relationship types)

Relationships

is read by dependency

Fine-Tuning Pipeline abstract Ch10.3

receives data from dynamic

Domain Expert Annotator Ch10.3

Design guidance

MAY be used alongside preference comparisons; demonstrations are particularly valuable early in training.

Quantitative guidance

As stated by the sources; verify before use.

InstructGPT used approximately 13,000 carefully crafted demonstrations (Ch10.3).

Sources

Ch10.3: T. Nguyen, "RLHF Methodology," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.3. ISBN: 9798244538229.