Introduction to LLMs

This is a companion site for the course Introduction to LLMs for the Norwich Biosciences Institutes.

It’s a 2 hours walk-through roughly developed upon this by this 6 minutes presentation:

The core idea is understanding that LLMs are not intelligent, but overpowered stochastic parrots that can be used to generate text, code, and even images, but they can also confidently generate wrong information (but in a nice English!).

Ethical use of AI and privacy concerns are also discussed, with a focus on how to use AI responsibly in research.

The topics discussed include:

• A simple introduction to Machine Learning and basic principles of LLMs
• Some practical tips on using LLM chatbots (Claude, ChatGPT, Perplexity)
• How we can run some LLMs locally (in your computer, without sending any data to third parties)
• Literature mining specific tools: what works and what doesn’t
• Coding assistants and coding agents: what they can do and how to use them