Model Adaptation · Model asset

Perplexity Reference Model

Model assetModel AdaptationModelsarc:PerplexityReferenceModel

A generative language model trained on relatively clean text whose average per-token log-likelihood is used to score how typical a document is.

Responsibility. Provides per-document perplexity scores to the perplexity filter.

Also known as: Reference language model

Direct neighbourhood (hover for relationship types)

Relationships

deployed on structural

Perplexity Filter Ch7.5

Design guidance

SHOULD be domain-calibrated, because a generic reference model inflates perplexity for specialized terminology.

Classification

Technologies: GPT-2

Sources

Ch7.5: T. Nguyen, "NeMo Curator, Riva Speech AI & Multimodal Integration," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 7.5. ISBN: 9798244538229.