// GLOSSARY -- AGENTIC AI

What is Large Language Model (LLM)?

1 min read Updated

A Large Language Model (LLM) is a neural network trained on vast text corpora that can understand, generate, and reason about natural language, serving as the foundation for modern AI agents and assistants.

WHY IT MATTERS

Large Language Models are the engines behind the current AI revolution. Models like GPT-4, Claude, Gemini, and Llama are trained on trillions of tokens, learning patterns of language, reasoning, and knowledge that emerge at scale.

What makes LLMs transformative for agents is their generality. A single model can understand instructions, reason about complex tasks, generate code, parse structured data, and make decisions — all capabilities needed for autonomous agents.

The key limitation: LLMs are probabilistic. They predict the most likely next token, which means they can be confidently wrong. For financial applications, this means you need deterministic guardrails around LLM-driven decisions.

FREQUENTLY ASKED QUESTIONS

How large is 'large' in LLM?
Modern frontier models have hundreds of billions of parameters. Even 'small' models today (7B-13B parameters) are enormous by historical standards.
Can LLMs reason or just pattern-match?
LLMs demonstrate emergent reasoning capabilities, but whether this constitutes 'true reasoning' vs sophisticated pattern matching is an open research question.
Why do LLMs hallucinate?
LLMs generate text by predicting likely continuations. When training data is sparse or contradictory, the model generates plausible-sounding but factually incorrect text.

FURTHER READING

Enforce policies on every tool call

Intercept is the open-source MCP proxy that enforces YAML policies on AI agent tool calls. No code changes needed.

npx -y @policylayer/intercept
github.com/policylayer/intercept →
// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.