Llm Glossary
Browse 70 llm terms defined in plain English, from the cultural dictionary of computing.
70 Llm Terms
- Adapter Layer
- A small trainable module inserted into a larger pre-trained model for parameter-efficient adaptation. It influences how models are trained, evaluated, or...
- Agentic Workflow
- An AI system design where a language model autonomously plans, executes, and iterates on tasks using tools (code execution, web search, APIs) in a loop. Unlike...
- AI Agent
- An AI system that autonomously pursues goals by planning actions, executing tools (code, APIs, web browsing), observing results, and iterating. Unlike simple...
- Causal Language Model
- A language model trained to predict the next token using only preceding context. It influences how models are trained, evaluated, or served, and it can...
- Chain of Thought Prompting
- A prompting technique that encourages a model to produce intermediate reasoning steps before the final answer. It influences how models are trained, evaluated,...
- Context Window
- The maximum number of tokens (input + output) that a language model can process in a single interaction. Determines how much text the model can 'see' at once....
- Context Window
- The maximum amount of text (measured in tokens) that a language model can process in a single interaction. Everything the model can 'see' and 'remember' during...
- DPO
- Abbreviation for direct preference optimization, a preference-based fine-tuning method for language models. It influences how models are trained, evaluated, or...
- Foundation Model
- A large AI model pre-trained on broad data that can be adapted to a wide range of downstream tasks through fine-tuning or prompting. GPT-4, Claude, Llama —...
- Function Calling
- A model interaction pattern where the model emits structured arguments that a program routes to tools or APIs. It influences how models are trained, evaluated,...
- Gemini
- Google's family of multimodal AI models, succeeding the PaLM series, capable of processing and generating text, code, images, audio, and video. Gemini launched...
- Gemma
- A family of lightweight generative models released by Google for research and application development. It influences how models are trained, evaluated, or...
- GPT-4
- OpenAI's fourth-generation Generative Pre-trained Transformer, a large multimodal model capable of processing both text and images. Released in March 2023,...
- GPT Architecture
- The decoder-only transformer architecture used in the GPT family of autoregressive language models. It influences how models are trained, evaluated, or served,...
- Groq
- A company that designs custom AI inference chips (Language Processing Units, or LPUs) optimized for running large language models at extreme speed. Groq's...
- Guardrails
- Validation layers applied to AI model inputs and outputs to enforce safety, accuracy, and compliance constraints. Can include content filters, format...
- Hallucination
- Hallucination in artificial intelligence refers to a model generating content that is factually incorrect, fabricated, or nonsensical while presenting it with...
- In-Context Learning
- A model behavior where examples placed in the prompt shape predictions without updating model weights. It influences how models are trained, evaluated, or...
- LangChain
- An open-source framework for building applications powered by large language models. LangChain provides abstractions for chaining together LLM calls, tool...
- Large Multimodal Model
- A large model trained to understand or generate across multiple modalities such as text, image, or audio. It influences how models are trained, evaluated, or...
- LLM Agent
- LLM Agent is an AI or ML concept used to represent, train, evaluate, or deploy learned systems. It is commonly used for building production models and research...
- LLM Evaluation
- LLM Evaluation is an evaluation concept used to measure model quality, robustness, or efficiency. It is commonly used for comparing systems before release or...
- LLM Fine-Tuning
- LLM Fine-Tuning is a stage of model optimization where weights or behaviors are adjusted from data or feedback. It is commonly used for foundation-model...
- LLM Inference
- LLM Inference is the phase where a trained model processes new inputs to produce predictions or generations. It is commonly used for production APIs, batch...
- LLM Routing
- LLM Routing is an AI or ML concept used to represent, train, evaluate, or deploy learned systems. It is commonly used for building production models and...
- LLM Safety
- LLM Safety is a safety or governance control for limiting harmful, noncompliant, or insecure model behavior. It is commonly used for production AI systems that...
- Local LLM
- Local LLM is an AI or ML concept used to represent, train, evaluate, or deploy learned systems. It is commonly used for building production models and research...
- Long Context
- Long Context is an AI or ML concept used to represent, train, evaluate, or deploy learned systems. It is commonly used for building production models and...
- LoRA
- Low-Rank Adaptation — a parameter-efficient fine-tuning technique that freezes the pre-trained model weights and injects small trainable matrices (adapters)...
- mixture of agents
- An architecture where multiple LLM agents collaborate by having each agent process the outputs of others in iterative rounds, leveraging the phenomenon that...
- Multimodal AI
- AI systems that can process and generate multiple types of data — text, images, audio, video, and code — within a single model. GPT-4V, Claude 3, and Gemini...
- Multimodal Tokenizer
- Multimodal Tokenizer is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly,...
- Ollama
- A tool for running large language models locally on personal hardware. Ollama packages model weights, configuration, and a runtime into a simple command-line...
- OpenAI
- OpenAI is an AI company and research lab that develops foundation models, developer APIs, and products such as ChatGPT. It is commonly used for building...
- Output Token
- Output Token is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly, decoding...
- Pair Programming (AI)
- Using an AI coding assistant (GitHub Copilot, Cursor, Claude) as a real-time programming partner that suggests code, explains errors, and generates...
- PII Detection
- PII Detection is a safety or governance control for limiting harmful, noncompliant, or insecure model behavior. It is commonly used for production AI systems...
- Prompt Caching
- Prompt Caching is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat systems,...
- Prompt Chaining
- Prompt Chaining is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat systems,...
- Prompt Engineering
- The practice of crafting input text (prompts) to elicit desired outputs from a language model. Part art, part science, part dark magic. Ranges from 'please be...
- Prompt Engineering Detail
- Prompt Engineering Detail is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat...
- Prompt Injection
- An attack where malicious instructions are embedded in user input to override an LLM's system prompt or instructions. Direct injection modifies the prompt...
- Prompt Injection Detail
- Prompt Injection Detail is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat...
- Prompt Template
- Prompt Template is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat systems,...
- Prompt Tuning
- Prompt Tuning is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat systems,...
- Responsible AI
- Responsible AI is an AI or ML concept used to represent, train, evaluate, or deploy learned systems. It is commonly used for building production models and...
- Retrieval-Augmented Generation
- An architecture that improves LLM outputs by first retrieving relevant documents from an external knowledge base, then including them in the prompt as context....
- Rubber Duck (AI)
- Using an LLM as a rubber duck — explaining your bug to ChatGPT or Claude instead of a plastic duck. The AI actually responds, which is either more helpful or...
- Safety Classifier
- Safety Classifier is a safety or governance control for limiting harmful, noncompliant, or insecure model behavior. It is commonly used for production AI...
- Safety Filter
- Safety Filter is a safety or governance control for limiting harmful, noncompliant, or insecure model behavior. It is commonly used for production AI systems...
- Stop Token
- Stop Token is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly, decoding control,...
- Structured Output
- Constraining an LLM's output to conform to a specific format like JSON Schema, XML, or a grammar. Achieved through guided decoding (constraining token...
- Subword Tokenization
- Subword Tokenization is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly,...
- Super Alignment
- Super Alignment is a safety or governance control for limiting harmful, noncompliant, or insecure model behavior. It is commonly used for production AI systems...
- system prompt
- System prompt is the initial set of instructions given to a large language model (LLM) that defines its behavior, personality, constraints, and capabilities...
- System Prompt Detail
- System Prompt Detail is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat...
- Temperature
- A parameter that controls the randomness of a language model's output. Low temperature (near 0) makes outputs deterministic and conservative; high temperature...
- Temperature
- A parameter that controls the randomness of a language model's output. Lower temperatures (0.0–0.3) make outputs more deterministic and focused; higher...
- Token
- The basic unit of text that a language model processes — roughly 3/4 of an English word on average. Text is split into tokens by a tokenizer before being fed...
- Token Budget
- Token Budget is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly, decoding...
- Tokenization Detail
- Tokenization Detail is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly, decoding...
- Token Limit
- Token Limit is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly, decoding...
- tokens per second
- The standard throughput metric for language model inference, measuring how many tokens a model can generate per second. Higher TPS enables more responsive user...
- Tokens Per Second Detail
- Tokens Per Second Detail is a unit or control marker used when text is segmented and generated by language models. It is commonly used for prompt assembly,...
- tool use
- The capability of a language model to invoke external functions, APIs, or services during response generation. Rather than relying solely on its training data,...
- Tool Use
- The ability of an LLM to invoke external tools (functions, APIs, databases, code interpreters) to accomplish tasks beyond text generation. The model decides...
- Value Alignment
- Value Alignment is a safety or governance control for limiting harmful, noncompliant, or insecure model behavior. It is commonly used for production AI systems...
- Vibe Coding
- Vibe Coding is a software development approach where the programmer describes what they want in natural language and lets an AI assistant generate the actual...
- Zero-Shot Learning
- The ability of a model to perform a task it was never explicitly trained on, by leveraging its general knowledge. GPT answering questions about a topic it...
- Zero-Shot Prompting
- Zero-Shot Prompting is a prompting technique or prompt-side control used to shape large-model behavior at inference time. It is commonly used for chat systems,...
Related Topics
- Nlp (27 terms in common)
- Inference (26 terms in common)
- Ai (9 terms in common)
- Architecture (7 terms in common)
- Ai Safety (6 terms in common)
- Governance (6 terms in common)
- Agents (3 terms in common)
- Multimodal (3 terms in common)
- Prompting (3 terms in common)
- Limitation (2 terms in common)
- Generalization (2 terms in common)
- Google (2 terms in common)
- Parameter (2 terms in common)
- Fine Tuning (2 terms in common)
- Integration (2 terms in common)
- Foundation Models (1 terms in common)
- Failure Mode (1 terms in common)
- Python (1 terms in common)
- Autoregressive (1 terms in common)
- Api (1 terms in common)