AI Literacy 101

Congressional Briefing

Ariel Fogel

November 5, 2025

How LLMs Actually Work

Let's watch an LLM in action

Transformer Explainer showing LLM processing

Prompt: "In today's congressional…"

Completed text showing 'In today's congressional debate'

The model picked "debate"

Probability distribution for next token completion

But "briefing" got 0% probability

Briefing shown with 0% probability

Safety vs. Security

Understanding risk through system architecture

Architecture 1: Standalone LLM

Not connected to external data or tools

Architecture 2: RAG

Retrieval-Augmented Generation: Connected to knowledge bases

Architecture 2: RAG

Architecture 2: RAG

Architecture 3: Agents

Architecture 3: Agents

Architecture 3: Agents

Architecture 3: Agents