Agentic RAG
- Definition
- Agentic RAG is a form of retrieval-augmented generation in which an AI agent decides what to search, runs several retrieval steps across sources, checks the results, and retries until it can answer.
Showing 20 of 62 terms
AI Content Watermarking
- Definition
- AI content watermarking embeds a hidden, machine-readable signal in generated text, images, audio, or video so that the output can later be identified as AI-made.
AI Red Teaming
- Definition
- AI red teaming is the structured practice of attacking an AI system before and after deployment to find the prompts, inputs, and workflows that make it fail, leak data, or cause harm.
Artificial Intelligence (AI)
- Definition
- Artificial intelligence (AI) is software that performs tasks which normally require human judgment, such as recognizing patterns, understanding language, and choosing actions. In business it ranges from fraud scoring to agents working…
Benchmark
- Definition
- A benchmark is a standardized set of tasks and a scoring method used to compare AI models against each other, giving a repeatable measure of capability on a defined problem.
Chain-of-Thought Prompting
- Definition
- Chain-of-thought prompting asks a language model to write out its intermediate reasoning steps before giving a final answer, which improves accuracy on arithmetic, logic, and multi-step business rules.
Chatbot
- Definition
- A chatbot is a software program that holds a text or voice conversation with a person to answer questions or complete simple tasks. Modern chatbots run on large language models grounded in…
ChatGPT
- Definition
- ChatGPT is a conversational AI assistant from OpenAI, released in November 2022, that lets people interact with GPT language models through a chat interface on the web, desktop, and mobile.
Chunking
- Definition
- Chunking is the step in a retrieval-augmented generation pipeline that splits documents into smaller passages so they can be embedded, searched, and supplied to a language model within its context window.
Claude
- Definition
- Claude is a family of large language models and a conversational AI assistant developed by Anthropic, first released in March 2023 and offered through a chat product and an API.
Content Provenance
- Definition
- Content provenance is a verifiable record of where a piece of digital content came from and how it was edited, attached to the file as signed metadata.
Conversational AI
- Definition
- Conversational AI is the set of technologies that lets software hold a natural, multi-turn dialogue with people by text or voice, combining speech recognition, language understanding, dialogue management, and back-end integration.
Copilot (AI Assistant)
- Definition
- A copilot is an AI assistant embedded in a work application that drafts, summarizes, and answers questions from the content the user has open, while the person keeps control of the final…
Data Poisoning
- Definition
- Data poisoning is an attack that corrupts the data an AI system learns from or retrieves, so that the model later gives wrong answers, hides a backdoor, or favors the attacker.
Deepfake
- Definition
- A deepfake is synthetic audio, video, or imagery generated by AI to depict a real person saying or doing something they never did, usually convincingly enough to deceive a viewer.
Direct Preference Optimization (DPO)
- Definition
- Direct preference optimization (DPO) is a method for aligning a language model to human preferences by training directly on pairs of preferred and rejected responses, without a separate reward model.
Document Automation
- Definition
- Document automation is software that assembles contracts, letters, proposals, and reports from approved templates, business rules, and system data, and can use a language model to draft the passages that vary.
Embeddings
- Definition
- Embeddings are numeric vectors that represent the meaning of text, images, or other data so that items with similar meaning sit close together and can be compared, searched, and clustered by a…
Energy-Aware Routing (Green AI)
- Definition
- Energy-aware routing is a form of model routing that includes energy use in the decision, sending each request to the smallest model, hardware, or time slot that meets its quality needs.
Few-Shot Prompting
- Definition
- Few-shot prompting is the technique of including a small number of worked examples of input and desired output in a prompt so that the model copies their format and judgment on new…
Foundation Model
- Definition
- A foundation model is a large AI model trained on broad data that can be adapted to many different tasks through prompting, retrieval, or fine-tuning instead of being built for a single…
Function Calling
- Definition
- Function calling is a language model capability where the model, given descriptions of available functions, responds with a structured request to run one, and the application executes it and returns the result.
General-Purpose AI (GPAI) Model
- Definition
- A general-purpose AI (GPAI) model is the EU AI Act's legal term for a model, such as a large language model, that performs many distinct tasks and can be built into many…
Generative Adversarial Network (GAN)
- Definition
- A generative adversarial network is a pair of neural networks trained against each other, one producing synthetic data and the other judging whether it is real, until the output becomes convincing.
Generative AI
- Definition
- Generative AI is a class of AI models that produce new content, such as text, images, code, or audio, by learning the patterns in large collections of existing examples.
Generative Model
- Definition
- A generative model is a machine learning model that learns the distribution of its training data well enough to produce new examples that resemble it, such as text, images, audio, or tabular…
GPT (Generative Pre-Trained Transformer)
- Definition
- GPT, short for Generative Pre-Trained Transformer, is a family of large language models built on the transformer architecture and trained to predict the next token in a sequence of text.
GraphRAG (Knowledge Graph RAG)
- Definition
- GraphRAG is a retrieval-augmented generation approach that stores entities and their relationships in a knowledge graph so that a language model can answer questions requiring connections across many documents.
Grounding
- Definition
- Grounding is the practice of tying a generative AI model's answers to verified sources, such as retrieved documents or database records, so the output reflects real facts instead of the model's memory.
Guardrails
- Definition
- Guardrails are the controls placed around a generative AI model that check inputs and outputs against rules so the system stays within its intended scope, tone, and legal limits.
Hallucination
- Definition
- A hallucination is output from a generative AI model that is fluent and confident but false, such as an invented fact, citation, number, or claim that an action was completed.
Hybrid Search
- Definition
- Hybrid search combines keyword matching with vector similarity in one retrieval step, so that exact identifiers are matched literally while related passages are still found by meaning.
Image Generation
- Definition
- Image generation is the use of AI models, mostly diffusion models, to create new images from text prompts or reference pictures. Businesses use it for marketing visuals, product mockups, and synthetic training…
Jailbreak
- Definition
- A jailbreak is a prompt or conversation designed to make a language model ignore its safety training and operating instructions so that it produces content or takes actions it was built to…
Knowledge Base (AI)
- Definition
- An AI knowledge base is the curated collection of documents, records, and procedures that a generative AI system retrieves from when it answers questions about a specific company or domain.
Knowledge Cutoff
- Definition
- The knowledge cutoff is the date after which a language model's training data stops, so the model has no built-in knowledge of events, products, rules, or prices that appeared later.
Knowledge Graph
- Definition
- A knowledge graph is a database that stores facts as entities connected by named relationships, such as a supplier that provides a part used in a product, so systems can query and…
Large Language Model (LLM)
- Definition
- A large language model (LLM) is a neural network trained on very large amounts of text to predict the next token, which lets it draft, summarize, translate, classify, and answer questions in…
LLM-as-a-Judge
- Definition
- LLM-as-a-judge is an evaluation method in which a language model scores or ranks the outputs of another model against a rubric, replacing or supplementing human review at scale.
Modality
- Definition
- A modality is a type of data that an AI model can take in or produce, such as text, images, audio, video, or sensor readings. Models are often described by the modalities…
Model Routing
- Definition
- Model routing is the practice of sending each AI request to the most suitable model, local or cloud, based on the task type, cost, latency requirement, and sensitivity of the data involved.
Multimodal AI
- Definition
- Multimodal AI is AI that can take in, reason over, or produce more than one type of data, such as text, images, audio, and video, within a single model or system.
Natural Language Generation (NLG)
- Definition
- Natural language generation (NLG) is the production of readable text by software from data, instructions, or other text. It is the output half of NLP and the basis of most generative AI…
Permission-Aware Retrieval
- Definition
- Permission-aware retrieval is the practice of restricting the documents a generative AI system can retrieve to those the requesting user is authorized to see, enforced before the model receives any text.
Prompt
- Definition
- A prompt is the text sent to a generative AI model to get a response, including the instructions, the question, any supplied documents or examples, and the required output format.
Prompt Chaining
- Definition
- Prompt chaining is the technique of breaking a task into a sequence of smaller prompts, where the output of each step becomes the input to the next, with checks between steps.
Prompt Engineering
- Definition
- Prompt engineering is the discipline of designing, testing, and maintaining the instructions given to a generative AI model so that it produces correct, consistent output for a specific business task.
Prompt Injection
- Definition
- Prompt injection is an attack in which instructions hidden in content a language model processes, such as an email, document, or web page, cause it to act against its operator's intent.
Prompt Template
- Definition
- A prompt template is a reusable prompt with placeholders that an application fills with the user's question, retrieved documents, or record data at run time, so every request is built the same…
Reasoning Model
- Definition
- A reasoning model is a large language model trained to work through a problem in intermediate steps before giving its final answer, trading speed and cost for accuracy on harder tasks.
Reinforcement Learning from Human Feedback (RLHF)
- Definition
- Reinforcement learning from human feedback (RLHF) is a training method that teaches a language model to produce answers people prefer by learning from human rankings of its outputs.
Reranking
- Definition
- Reranking is a second scoring pass that takes the candidate passages returned by a search and reorders them with a more accurate model so the best evidence reaches the language model first.
Retrieval-Augmented Generation (RAG)
- Definition
- Retrieval-augmented generation (RAG) is a technique that finds relevant passages from a company's own documents and supplies them to a language model so that it answers from current, verifiable sources.
Semantic Search
- Definition
- Semantic search is a retrieval method that finds documents by meaning rather than by matching exact words, using embeddings to return the passages closest to what the user asked.
Streaming (LLM Responses)
- Definition
- Streaming is the delivery of a language model's response token by token as it is generated, so the user sees text appear immediately instead of waiting for the full answer.
Structured Outputs
- Definition
- Structured outputs are model responses that conform to a predefined format such as a JSON schema, so that downstream software can parse the result without interpreting free text.
Summarization
- Definition
- Summarization is the use of a language model to condense a long text into a shorter one that keeps the important points, in a format and length set by the prompt.
System Prompt
- Definition
- A system prompt is the standing set of instructions an application gives a language model before any user message, defining the model's role, rules, tone, tools, and the boundaries of what it…
Temperature and Sampling
- Definition
- Temperature and sampling are the settings that control how a language model picks each next token, from always choosing the most likely option to drawing from a wider spread of possibilities.
Unstructured Data
- Definition
- Unstructured data is information without a predefined data model, such as documents, emails, images, audio, and chat transcripts, which must be interpreted before software can analyze it.
Vector Database
- Definition
- A vector database is a data store built to hold embedding vectors with their metadata and to find the vectors nearest to a query quickly, which makes it the retrieval layer for…
Vector Search
- Definition
- Vector search finds the stored items whose embedding vectors are closest to a query vector, which returns results by similarity of meaning instead of by matching the exact words.
No terms match your search.
Try a broader keyword or another category.