GenAI - Terminology

The terminology shared in this page is organized in alphabetical order.

Fine-Tuning

Fine-tuning is the process of retraining an already trained LLM on a new dataset to enhance the LLM's skills for a specific task.

Grounding

Grounding is the process of sharing additional data—such as domain-specific information—with the LLM. The main purpose is to limit hallucinations from the model and improve its performance

Hallucination

We refer to an unexpected output from an LLM (e.g., wrong format, wrong content) as a hallucination. A factual error from an LLM will also be qualified as a hallucination. It is more commonly defined as an "error" in other software development fields.

Large Language Model (LLM)

Large Language Models (LLMs) are specialized algorithms trained on very large datasets with unique skills to perform tasks such as content generation, content summarization, translation, and more.

Prompt Engineering

Prompt engineering is the process of structuring text that can be interpreted and understood by a generative AI model (Source: Wikipedia)

Retrieval-Augmented Generation (RAG)

Retrieval-Augmented Generation (RAG) is the process of grounding a model with knowledge (e.g., data), improving context relevance, and reducing hallucinations.