Plain-English definitions for models, training, interfaces, safety, and common AI product concepts.
A hypothetical AI system that possesses human-level intelligence across all cognitive tasks. Unlike narrow AI (designed for specific tasks), AGI would be able to learn and perform any intellectual task a human can.
Hypothetical AI that can understand and learn any intellectual task that a human can.
An AI system that can autonomously perform tasks, make decisions, and interact with its environment to achieve goals. Agents can use tools, browse the web, write code, and chain multiple actions together.
A set of protocols and tools that allows different software applications to communicate. AI APIs let developers integrate AI capabilities into their applications without building models from scratch.
The simulation of human intelligence processes by computer systems, including learning, reasoning, and self-correction. AI encompasses machine learning, natural language processing, computer vision, and robotics.
Systematic errors or unfair outcomes in AI systems that arise from biased training data, flawed algorithms, or problematic design choices. AI bias can perpetuate discrimination based on race, gender, or other factors.
A field of AI that trains computers to interpret and understand visual information from the world, including images and videos. It powers facial recognition, object detection, and image generation.
The maximum amount of text (measured in tokens) that a language model can process in a single interaction. Larger context windows allow models to handle longer documents and conversations.
A subset of machine learning that uses artificial neural networks with multiple layers (deep neural networks) to model complex patterns in data. It powers image recognition, speech processing, and language understanding.
A type of generative model that creates data by gradually removing noise from a random signal. Used in image generators like Stable Diffusion, DALL-E, and Midjourney to create photorealistic images from text.
A generative AI model that creates images by gradually removing noise from random data.
A numerical representation of data (text, images, etc.) in a high-dimensional vector space. Embeddings capture semantic meaning, allowing AI to understand relationships and similarities between concepts.
A technique where an AI model learns to perform a task from only a few examples provided in the prompt. This allows models to adapt to new tasks without retraining.
The process of taking a pre-trained model and further training it on a specific dataset to adapt it for a particular task or domain. This is more efficient than training from scratch.
The process of further training a pre-trained model on a specific dataset for a particular task.
AI systems that can create new content including text, images, audio, video, and code. Examples include ChatGPT (text), Midjourney (images), Suno (music), and GitHub Copilot (code).
A family of large language models developed by OpenAI. GPT models are pre-trained on large text corpora and can be fine-tuned for specific tasks. GPT-4 is the latest publicly available version.
When an AI model generates information that sounds plausible but is factually incorrect or fabricated. This is a common challenge with LLMs that can confidently produce false statements.
When an AI model generates information that sounds plausible but is factually incorrect.
The process of using a trained AI model to make predictions or generate outputs on new, unseen data. Inference speed and cost are key factors in deploying AI applications.
A type of AI model trained on vast amounts of text data that can generate, understand, and manipulate human language. Examples include GPT-4, Claude, Gemini, and LLaMA.
A type of AI model trained on vast amounts of text data to understand and generate human language.
A subset of AI where systems learn and improve from experience without being explicitly programmed. ML algorithms build mathematical models from training data to make predictions or decisions.
AI systems that can process and generate multiple types of data (text, images, audio, video) simultaneously. GPT-4V and Gemini are examples of multimodal models.
The branch of AI that helps computers understand, interpret, and generate human language. NLP powers chatbots, translation tools, sentiment analysis, and text summarization.
A computing system inspired by biological neural networks. It consists of interconnected nodes (neurons) organized in layers that process information and learn patterns from data.
A computing system inspired by biological neural networks that can learn patterns from data.
AI models and tools whose source code and weights are publicly available for anyone to use, modify, and distribute. Examples include LLaMA, Stable Diffusion, and Whisper.
When a machine learning model performs well on training data but poorly on new, unseen data. The model has essentially memorized the training examples rather than learning generalizable patterns.
The practice of designing and optimizing input prompts to get better outputs from AI models. Effective prompts include context, examples, constraints, and clear instructions.
The practice of designing effective prompts to get desired outputs from AI models.
A technique that combines information retrieval with text generation. The model first retrieves relevant documents from a knowledge base, then uses them as context to generate more accurate responses.
A technique that combines information retrieval with text generation for more accurate outputs.
A training technique where AI models are fine-tuned using human preferences and feedback. Humans rank model outputs, and the model learns to produce responses that align with human values.
A parameter that controls the randomness of AI model outputs. Lower temperature (0.0-0.3) produces more deterministic, focused responses. Higher temperature (0.7-1.0) produces more creative, varied outputs.
The basic unit of text that language models process. A token can be a word, part of a word, or a character. Most LLMs have a maximum context window measured in tokens (e.g., 128K tokens).
The basic unit of text that language models process, typically a word or part of a word.
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.
A deep learning architecture that uses self-attention mechanisms, the foundation of models like GPT.
The ability of an AI model to perform a task it was not explicitly trained on, without any examples. Modern LLMs demonstrate strong zero-shot capabilities across many tasks.
