Decode LLM, RAG, RLHF and other AI abbreviations in minutes. This compact glossary offers practical definitions for developers, investors, and product teams.
As mentioned by Techcrunch
Artificial intelligence is rapidly changing the world and at the same time giving rise to a new language of descriptions about how it works. In contemporary product meetings, pitches and panels, abbreviations like LLM, RAG, or RLHF often appear – and this can confuse even experienced professionals. This AI glossary in Ukrainian offers a compact, practical reference of terms you are most likely to encounter during development, investing, or simply reading about artificial intelligence. We update it regularly, as the field is rapidly evolving.
AGI – artificial general intelligence. This concept typically denotes systems capable of performing a wide range of tasks at human-level or above. Different sources provide different formulations, but the general view is that AGI encompasses the ability to handle most cognitive tasks.
AI agent – an autonomous system that uses AI to perform sequential tasks from a user. Such agents can act based on several AI systems, for example expense reporting, bookings, or writing code, but the exact meaning of the term can vary depending on context and available infrastructure.
API endpoints – API endpoints that serve as interfaces for integrations between programs. They allow one program to fetch data from another or manage third-party services without manual intervention. As AI agents develop, these endpoints are used more widely.
Chain of thought – a chain of reasoning: step-by-step breakdown of a task into intermediate steps. In the case of large language models, such thinking can improve answer accuracy but requires more time to justify each step. In model training, this is used to improve the quality of conclusions through specialized methods.
Compute – computational power needed to train and deploy models. This is primarily hardware: graphics processing units (GPUs), central processing units (CPUs), tensor processing units (TPUs), and other infrastructure that forms the backbone of the modern AI industry.
Deep learning – a subfield of machine learning that uses multi-layer artificial neural networks. Such architectures enable discovering complex relationships in data but require large amounts of data and longer training times compared with simpler methods.
Large language model (LLM) – large language models, neural networks with billions of parameters, that process natural language and generate textual responses. Examples include ChatGPT, Claude, Google Gemini, Meta Llama, Microsoft Copilot, and Mistral Le Chat.
Neural network – a neural network, a multilayer architecture that underpins deep learning and modern generative technologies. It uses interactions between layers to transform input data into useful signals.
Open source – open-source software. Such approaches allow researchers and companies to collaboratively develop and analyze solutions, providing transparency and auditability. Closed solutions have limited access to internal logic.
Reinforcement learning – reinforcement learning: an agent learns by interacting with the environment and receives rewards for correct actions, gradually shaping its behavior. In modern models this is often combined with human feedback (RLHF) to make responses more useful and safe.
Token – the basic unit of data in human–model interaction. Tokens divide text into smaller fragments for processing by the model, and the cost of using an LLM is often calculated by the number of tokens.
Training – the process of training a model on large volumes of data to detect patterns and form useful responses. Training can require significant resources, and sometimes is complemented by techniques such as fine-tuning to adapt to specific tasks.
The AI glossary is updated regularly, as the field is rapidly evolving. It will become a useful reference for developers, investors, and everyone following trends in artificial intelligence.
The only AI glossary you’ll need this year
Don’t miss other news:
Leave a Reply