The terms the intel index leans on, in plain English. Technical, not dumbed down — each entry says what the thing is and why anyone bothers.
Changing a model's behavior by nudging its internal activations directly at runtime — no retraining, no prompt changes.
The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.
An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
The work of making AI systems actually pursue what their builders and users intend, rather than something subtly or dangerously different.
The mechanism that lets a model weigh which earlier words matter for the word it's currently processing — the core operation of a transformer.
A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
How well a model's confidence matches reality — a calibrated model saying "90% sure" is right about 90% of the time.
Having a model write out intermediate reasoning steps before its answer, which markedly improves performance on hard problems.
The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
Training a small, cheap model to imitate a big one's outputs, keeping much of the capability at a fraction of the cost.
A list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.
A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
Taking a trained model and training it a bit more on your own examples so it gets better at one specific job.
Tying a model's answers to checkable sources — retrieved documents, live data, tool results — instead of letting it answer from memory alone.
The checks around a model that block bad inputs and outputs — filters, validators, and permission rules the model itself can't override.
When a model states something false with full confidence — inventing facts, citations, or APIs that don't exist.
Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
The memory a model keeps about text it has already read, so generating each new token doesn't require reprocessing the whole conversation.
A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
A cheap way to fine-tune a model by training a small add-on layer instead of changing all of the model's weights.
The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
A model built from many specialist sub-networks where only a few activate per token, giving big-model capability at small-model running cost.
Using several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.
A model that works with more than text — reading images, audio, or video, and sometimes generating them too.
A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
The first, biggest phase of building a model: training it on enormous amounts of text so it learns language, facts, and reasoning in general.
An attack that hides instructions in content an AI will read — a webpage, email, or document — tricking it into following the attacker instead of the user.
Shrinking a model by storing its numbers less precisely — like rounding — so it runs faster and fits on smaller hardware, at a small quality cost.
Retrieval-augmented generation — fetching relevant documents first and pasting them into the model's context so it answers from your data instead of memory.
A model trained to think — generating extended internal reasoning before answering — trading time and tokens for accuracy on hard problems.
Reinforcement learning from human feedback — training a model to prefer answers humans rate as better, which turns a raw text predictor into a usable assistant.
An isolated environment where AI-generated code or agent actions run without being able to touch anything real.
A speed trick where a small model drafts several tokens ahead and the big model verifies them in one pass, often doubling generation speed.
The standard benchmark for AI coding agents: real GitHub issues from real repositories, scored by whether the agent's patch passes the project's own tests.
The standing instructions a model receives before any user input — defining its role, rules, tools, and tone for the whole conversation.
Spending more computation when the model answers — thinking longer, trying multiple attempts — to buy accuracy without training a bigger model.
The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
The neural network architecture behind modern AI models, built on attention — letting every word directly consider every other word in parallel.
Building software by describing what you want to an AI in plain language and steering the result, instead of writing every line yourself.