Welcome

Install + Setup

Take a look at the workshop website and go through the setup instructions: https://github.com/chendaniely/2025-06-05-llm

Url is at the bottom of all the slides.

  1. Clone this repo
  2. Install your R + Python packages
  3. Download at least one of the Ollama models, I provided a few to pick from. Feel free to pick any other one.
  4. (Optional) use the .env.template file to provide your API key into .env

Note

If you pay for Claude, OpenAI, etc access with their web/desktop application, this is a separate purchase for the API key. Depending on your usage, you may even find that paying for the API key could be cheaper!

Passing along what I learned

Joe will do a better job than I can, but I can demo you code today.

https://www.youtube.com/watch?v=owDd1CJ17uQ

Poll: Experience with LLMs

  1. Used an LLM before (ChatGPT/Claude/Ollama desktop/web application)?

  2. Used it for a homework assignment?

  3. MDS: Using for capstone?

  4. Tasks outside of school work?

  5. Skeptical about LLMs/AI (1-2 out of 5)? Why?

  6. Neutral about LLMs/AI (3 out of 5)? Why?

  7. Enthusiastic about LLMs/AI (4-5 out of 5)? Why?

Today

  • MDS: taught you how things work behind the scenes (transformers!)
  • Today, we will treat LLMs as black boxes
  • Practical introduction
  • Get some hands on practice to demystify using them
    • MDS: Maybe you can throw in something extra for the end of your capstone projects?

Goal

Quick Start course on LLMs. You will leave having used a Chat API.

  • MDS: if you made something cool from what you learned today, share it in #llm-workshop-hackathon
    • Doesn’t have to be “successful”
    • We’re just here trying to make cool things
    • Share so we can learn about the limitations
  • DSCI 100: We can always chat during office hours

Security

  • DO NOT send proprietary code or data to any LLM, unless you are sure IT policies allow it
  • Local models (e.g., Ollama) typically perform worse than frontier models