29,259 technical terms and definitions
programmable read-only memory, fusible link, one-time programmable, field-programmable ROM, write-once memory
PROM, or programmable read-only memory, is a read-only memory that is manufactured blank and programmed once by the user after fabrication, rather than having its data fixed…
metrics, monitoring
Prometheus is the open-source monitoring and alerting toolkit that collects time-series metrics by scraping HTTP endpoints on a pull-based architecture — serving as…
mlops
Prometheus is an open-source monitoring and alerting toolkit that collects, stores, and queries time-series metrics data. It has become the de facto standard for monitoring…
prompting, instruction
Prompt engineering is the systematic methodology of formulating, structuring, and optimizing natural language inputs and contextual conditioning to steer the conditional…
continual learning
Prompt-based continual learning is continual adaptation that uses learned prompts or prefix tokens to encode task-specific behavior - Task behavior is steered through prompt…
optimization
Prompt caching is a technique that stores and reuses the processed prefix of prompts (particularly system prompts and common instructions) to avoid redundant computation on…
inference
Prompt caching is the technique that stores reusable prompt processing artifacts so repeated prompts can skip full prefill computation - it accelerates inference for recurring…
optimization
Prompt Caching is a performance optimization that reuses previously computed prompt-prefix representations - It is a core method in modern semiconductor AI serving…
prompting
Prompt chaining is the workflow pattern where outputs from one prompt stage become inputs to subsequent stages in a multi-step pipeline - chaining decomposes complex tasks into…
prompting techniques
Prompt Chaining is a workflow pattern that links multiple prompts sequentially so each step feeds the next stage - It is a core method in modern LLM workflow execution. What Is…
text splitting, long document
Prompt chunking is the method that splits long text into manageable token segments and processes them in structured passes - it extends effective prompt capacity beyond a single…
Prompt Completion and Prediction Introduction & Motivation Prompt Completion and Prediction addresses a central problem in generative generation and completion-relevance-quality…
prompting
Prompt composition is the systematic assembly of prompt components such as instructions, examples, retrieved context, and user query into a single coherent input - composition…
prompting techniques
Prompt Compression is techniques that reduce prompt token count while preserving essential task instructions and context - It is a core method in modern LLM execution workflows.…
generative models
Prompt embeddings is the vector representations produced from prompt text that carry semantic information into the generative model - they are the internal control signal that…
prompting techniques, few shot prompting, rag prompts, system prompt
prompt engineering is the systematic design of model instructions, context, examples, tools, and output constraints to elicit useful behavior from foundation models. It is a fast…
prompting
Advanced prompt engineering encompasses systematic techniques for eliciting optimal responses from large language models beyond basic instruction formatting. Key methods include…
prompt engineering context optimization, prompt engineering, engineering context, cot llm rag, prompt engineering context, engineering context optimization, context optimization
Prompt Engineering and Context Optimization Introduction & Motivation Effective prompts dramatically improve LLM outputs for engineering tasks. Prompt engineering combines domain…
prompting
Prompt engineering for RAG is the design of instructions, context formatting, and response constraints that guide the model to use retrieved evidence correctly - prompt quality…
Prompt Engineering for Multimodal Models Introduction & Motivation Prompt Engineering for Multimodal Models addresses a central problem in multimodal learning…
prompt engineering optimization, prompt engineering, engineering optimization
Prompt Engineering Optimization Introduction & Motivation Prompt Engineering Optimization addresses a central problem in nlp process control…
prompt engineering, engineering techniques, glue llm
Prompt Engineering Techniques Introduction & Motivation Prompt engineering: craft inputs for optimal LLM output. Critical for zero-shot and few-shot learning. Applications…
prompt engineering
Prompt ensemble is the technique of combining predictions from multiple different prompts for the same task — leveraging the diversity of prompt formulations to produce more…
jailbreak, llm security, adversarial prompts, red teaming, guardrails, safety bypass, input sanitization
Prompt injection exploits a specific underlying design fact about how large language models actually work: a model's own explicit instructions and the data that model is actually…
ai safety
Defending against prompt injection remains a genuinely active, still largely unsolved security research area, since no current defense technique, input sanitization…
ai safety
Prompt Injection is an attack technique that embeds malicious instructions in untrusted input to override intended model behavior - It is a core method in modern AI safety…
ai safety
Prompt injection attacks is the adversarial technique where untrusted input contains instructions intended to override or subvert system-defined model behavior - it is a primary…
system
Prompt Injection Defense What is Prompt Injection? Attacks where user input manipulates LLM behavior, bypassing intended instructions. Attack Types Attack Example Direct injection…
ai safety
Prompt injection defense is the set of architectural and prompt-level controls designed to prevent untrusted text from overriding trusted instructions or triggering unsafe actions…
ai safety
Prompt Leaking is the attack technique that extracts hidden system prompts, instructions, and confidential configurations from AI applications — enabling adversaries to reveal…
prompting techniques
Prompt Mining is the extraction of high-performing prompt patterns from existing corpora, logs, or historical experiments - It is a core method in modern LLM execution workflows.…
ai safety
Prompt moderation is the pre-inference safety process that evaluates user prompts for harmful intent, policy violations, or attack patterns before model execution - it reduces…
prompt engineering
Prompt Optimization is the systematic, automated process of improving prompts through search, gradient-based methods, or LLM-guided rewriting rather than manual trial-and-error…
prompting techniques
Prompt Optimization is the systematic search for prompt formulations that maximize task performance under defined metrics - It is a core method in modern LLM execution workflows.…
prompt engineering, templates, few-shot, chain of thought, role prompting
Prompt engineering patterns are reusable templates and techniques for structuring LLM interactions — providing proven approaches like few-shot examples, chain-of-thought…
prompting techniques
Prompt Sensitivity is the degree to which model outputs change in response to small prompt wording or formatting variations - It is a core method in modern LLM execution…
prompting
Prompt syntax is the formal text structure and token conventions used by a generation system to interpret user instructions - it determines how phrases, separators, weights…
prompting techniques
Prompt Template is a reusable prompt artifact with placeholders and fixed instruction blocks for repeatable model interactions - It is a core method in modern LLM workflow…
prompting
Prompt templates is the reusable prompt structure with parameterized fields that standardizes model inputs across repeated tasks - templates improve consistency, maintainability…
multimodal ai
Prompt-to-Prompt is a diffusion editing technique that modifies generated content by changing prompt text while preserving layout - It allows semantic edits without rebuilding…
generative models
Prompt-to-Prompt Editing is a text-guided image editing technique for diffusion models that modifies generated images by manipulating the cross-attention maps between text tokens…
generative models
Prompt truncation is the automatic removal of tokens beyond encoder context length when prompt input exceeds model limits - it is a common but often hidden behavior that can…
fine-tuning
Prompt tuning learns continuous "soft prompts" while keeping the base model frozen. Mechanism: Prepend learned embedding vectors to input, these vectors trained via…
prompting techniques
Prompt Tuning is a parameter-efficient adaptation method that learns virtual prompt embeddings while keeping base model weights frozen - It is a core method in modern LLM…
prefix tuning, soft prompt, learnable prompt, p tuning, prompt based fine tuning
Prompt Tuning and Prefix Tuning are the parameter-efficient fine-tuning methods that prepend small sequences of learnable "soft" token embeddings to the input or intermediate…
Soft Prompt Tuning for Extreme Parameter Efficiency Introduction & Motivation Soft Prompt Tuning for Extreme Parameter Efficiency addresses a central problem in transfer learning…
version control
Prompt Versioning and Management Why Version Prompts? Prompts are code. Track changes, roll back issues, and collaborate effectively. Git-Based Versioning Metadata Format Prompt…
prompt engineering
Prompt weighting assigns different importance levels to different parts of a text prompt. Syntax examples: AUTOMATIC1111 uses (word:weight), Midjourney uses ::weight suffix…
generative models
Prompt weighting is the method of assigning relative importance to prompt tokens or phrase groups to prioritize selected concepts - it helps resolve conflicts when multiple…
computer vision
Promptable Segmentation is a paradigm where segmentation masks are generated based on user inputs — allowing users to interactively define what to cut out using points, bounding…