AI Glossary (용어모음)

Generative models

Unsupervised Learning

Supervised Learning

Corpus of Data

Language Model

Token

Tokenization

Token Limits

ChatGPT

Prompt

Prompt Engineering

Response Generation

Training Data

Fine-Tuning

Context

Response Length

Bias

Human Parity

Gen AI APIs

Knowledge Cut-off

hallucination

User Feedback

Open-ended Prompt

Closed-ended Prompt

Zero-shot Learning

Few-shot Learning

Question Refinement

The process of improving or clarifying a question or query to elicit more precise and relevant information or responses. This can involve rephrasing a question, providing additional context, or specifying the desired level of detail.

Text Completion

Using a Generative AI model to fill in the missing parts of a text prompt, sentence, or paragraph.

Text Expansion

Using a Generative AI model to generate additional content based on an initial prompt, often used for content generation or writing assistance.

Text Summarization

Instructing a Generative AI model to provide a concise summary of a longer piece of text, such as an article or document.

Text Translation

Using a Generative AI model to translate text from one language to another by providing a source language text as input.

Text Sentiment Analysis

Using a Generative AI model to analyze and determine the sentiment (positive, negative, neutral) of a given text.

Text Classification

Instructing a Generative AI model to categorize a given text into predefined categories or labels.

Conditional Generation

Providing a Generative AI model with conditional instructions to generate text that adheres to specific criteria, such as writing in a certain style or tone.

Transfer Learning

The concept of using knowledge gained from one task or domain to improve performance on another task or domain, a key feature of a Generative AI model's training.

Chunking

Chunking is a process of breaking down text input into smaller, more manageable groups or "chunks." Structuring a prompt or input text into smaller, more focused segments to can help with token limits.

Flipped Interaction

Flipped interaction is a concept where the traditional roles of user and AI model are reversed. Instead of the user providing a prompt and the AI generating a response, the AI takes on a more proactive role, initiating the interaction by providing information, suggestions, or questions.

Audience Persona

An audience persona is a fictional representation of a target audience or user group. It includes details about the audience's demographics, preferences, behaviors, and needs this prompt technique can guide the model to generate content that is more relevant and appealing to a specific group.

Made with SlashPage