ZIRCONIAN

SESSION 01 · LIVE · FREE

How LLMs Actually Work

The foundational session every professional AI user needs. 26 sections. No coding required. Free forever.

26

Sections

~3h

Total time

Free

No credit card

Session 01 is the entry point for every IMAnthropic learner. It covers how language models actually work - not the hype, the mechanics. By the end, you'll understand why prompt structure matters, how tokens relate to cost, and what to do differently tomorrow.

WHAT'S COVERED

26 sections. Every LLM fundamental a professional needs.

01What is a language model?
02Tokens: the currency of AI
03How transformers process text
04Attention mechanism explained
05RLHF and alignment
06Context windows and memory
07Temperature and sampling
08System prompts vs user prompts
09Chain-of-thought reasoning
10Few-shot and zero-shot
11Hallucination: why it happens
12Model selection for your use case
13Cost vs quality tradeoffs
14Enterprise thinking about AI
15Prompt structure fundamentals
16Role-based prompting
17Output formatting
18Iteration and refinement
19Common failure modes
20When not to use AI
21Data handling and privacy
22API basics (no code)
23Claude vs GPT vs Gemini
24Building an AI habit
25Domain application exercise
26Your first efficiency baseline

YOUR DELIVERABLE

Write your first structured prompt using chain-of-thought framework.

Every IMAnthropic session ends with something you can use at work tomorrow. Session 01's deliverable: a chain-of-thought prompt for one of your recurring work tasks - and your first token efficiency baseline.

After Session 01.

Stay free

The community is open. #problem-statements is running. Continue exploring at no cost.

Upgrade to Practitioner - $40/month

Unlock Sessions 02–08, Claude API access, your domain cohort, and 4 mentor sessions.

Session 01 is open right now. Free.