Materials for the workshop "How to code with LLMs" - CoDEx 2024.
The workshop slides are in the file LLM_Codex_Workshop, and the prompt engineering summary sheet is in prompt_engineering_cheat_sheet.
Hands-on Exercise
Here are the links to the different chatbots we will use in the hands-on exercise:
PART 1:
Use an LLM to generate code to determine which country consumed the least meat per capita in 2020.
- Use the data hosted here.
- Run the code yourself to make sure it works.
- This is what the columns mean:
- LOCATION is the country.
- SUBJECT is meat type (beef, pig, poultry, sheep).
- MEASURE is the metric reported.
- KG_CAP - kilograms per capita.
- THND_TONNE - thousand tonnes.
- TIME is the year.
- VALUE is the value for the indicator.
PART 2:
Was exercise 1 too easy? Try this!
Which country has the least annual meat consumption per capita the most years between 2000 and 2020?
For example, if country X consumed the least for 7 yrs, and country Y consumed the least for 13 years. Country Y would be the answer.