ECG Events
ECG Webinar series: Experimenting with GPT and Generative AI for evaluation practice
18 October 2023 - 18 October 2023
Host: Independent Evaluation Group (IEG), World Bank Group
|
Venue: Virtual webinar
Over the past few months, there has been a lot of hype around Generative AI, ChatGPT, and other Large Language Models. Oscillating between awe and catastrophism, many opinions have been voiced. The Independent Evaluation Group’s Methods Team decided to take a dispassionate approach by doing what it does best, setting up clear experiments, and assessing the models as objectively as possible. In this webinar ,IEG shared with participants how to set up the experiments, what they found and what they recommend.
Participants discussed and learned about the promises, the risks, and the results of leveraging GPT, other Large Language Models and Generative AI for evaluation practice.
After the webinar, participants were able to:
- Understand what Large Language Models are, and what they are not.
- Know how to set-up experiments to test LLMs on various evaluative tasks.
- Identify use cases for evaluation practice based on concrete example.
- Assess the risk-cost-benefit of using LLMs for various evaluation tasks.
Presenters:
- Estelle Raimondo (Program Manager and Methods Advisor, Independent Evaluation Group)
- Harsh Anuj (Data Scientist, Independent Evaluation Group)
Contact: Estelle Raimondo: eraimondo [at] worldbank.org (eraimondo[at]worldbank[dot]org)
Related
Related Videos
Experimenting with GPT and Generative AI for evaluation practice