AI Glossary

An AI system that can autonomously perform tasks by planning, using tools, and making decisions. Agents can break down complex goals into steps and execute them with minimal human intervention.

A set of rules and protocols that allows different software applications to communicate. COZHUB provides a REST API that lets you integrate AI models into your applications with simple HTTP requests.

A prompting technique that encourages the AI to show its reasoning process step by step. This often improves performance on complex reasoning tasks like math problems or logical puzzles.

The maximum number of tokens an LLM can process in a single interaction. A larger context window allows the model to consider more information when generating responses. GPT-4 has a 128K token context window, while Claude 3 supports up to 200K tokens.

A numerical representation of text (or other data) in a high-dimensional vector space. Similar texts have similar embeddings, making them useful for semantic search, clustering, and recommendation systems.

A technique where you provide a few examples in the prompt to help the AI model understand the desired task or output format. This is more effective than zero-shot (no examples) for complex or specific tasks.

The process of further training a pre-trained model on a specific dataset to adapt it for a particular task or domain. Fine-tuning allows models to learn specialized knowledge while retaining their general capabilities.

The ability of AI models to call external functions or APIs based on user requests. This enables AI to perform actions like searching the web, accessing databases, or controlling other software.

A family of large language models developed by OpenAI. GPT models are pre-trained on internet text and can be fine-tuned for specific tasks. GPT-4 is currently one of the most capable AI models available.

When an AI model generates false or fabricated information that sounds plausible but is not grounded in facts. Techniques like RAG and careful prompting can help reduce hallucinations.

The process of using a trained AI model to generate predictions or outputs from new inputs. This is what happens when you send a request to an AI API - the model performs inference on your input.

A type of AI model trained on massive amounts of text data to understand and generate human-like text. Examples include GPT-4, Claude, and Gemini. LLMs can perform tasks like answering questions, writing content, coding, and translation.

AI models that can process and generate multiple types of data, such as text, images, audio, and video. Examples include GPT-4V (text + images) and Gemini (text, images, audio, video).

The input text given to an AI model to guide its response. A well-crafted prompt can significantly improve the quality and relevance of AI outputs. Prompt engineering is the practice of designing effective prompts.

The practice of designing and optimizing prompts to get better results from AI models. Techniques include providing examples (few-shot learning), breaking down complex tasks, and specifying output formats.

A technique that combines information retrieval with text generation. RAG systems first retrieve relevant documents from a knowledge base, then use that information to generate more accurate and up-to-date responses.

A parameter that controls the randomness of AI outputs. Lower temperatures (e.g., 0.1) produce more deterministic, focused responses. Higher temperatures (e.g., 0.9) produce more creative, varied outputs.

The basic unit of text that LLMs process. A token can be a word, part of a word, or a character. For English, 1 token is approximately 4 characters or 0.75 words. API pricing is typically based on the number of tokens processed.

A neural network architecture that uses self-attention mechanisms to process sequential data. Introduced in the 2017 paper "Attention Is All You Need," transformers are the foundation of modern LLMs and have revolutionized natural language processing.