Prompting, Instructions and the Future of Large Language Models

Presentation Link

Sam Witteveen and I started the Machine Learning Singapore Group (or MLSG for short) on MeetUp in February 2017 (it was previously named "TensorFlow and Deep Learning Singapore"), and the thirty-eighth MeetUp was our first 'in-person' event since the COVID era.

As we anticipated, everyone was extremely glad to get back to seeing each other in person - it felt almost surreal after being apart for so long : Lots of familiar faces to catch up with :-)

Many thanks are also due to the Google team, who not only allowed us to use Google's Developer Space, but were also kind enough to provide Pizza for the attendees, and also "Machine Learning Singapore"-branded swag for the early arrivers!

This month's event is covering some of the recent advances in the field of Large Language Models (LLMs):

Instructive Prompting - Sam Witteveen

Prompting LLMs has opened up a whole new world in NLP. In this talk Sam will go through some of the history of prompting in LLMs and how instructive prompting is being used in some of the latest models such as FLAN and BloomZ to massively improve results and how you can use them in your own project.

Tomorrow's World - My talk

Following on the current advances that Sam will talk about, this talk will highlight new research directions that may point a way forwards (other than "just train bigger models for longer"). This included :
- Having models learn from human preference;
- create their own data; and
- create their own prompts.

There's a video of me doing the talk on YouTube (this is a slightly longer version than the one at the MeetUp, partly due to the release of ChatGPT within days following the MeetUp...). Please Like and Subscribe!

The slides for my talk, which contain links to all of the reference materials and sources, are here :

If there are any questions about the presentation please ask below, or contact me using the details given on the slides themselves.