Feminist AI workshop - measuring bias in LLMs

​In this workshop, we will be investigating the bias on some open source LLM models. We will work in groups, come up with ways to measure bias in LLMs and there experiments will be performed with multiple open source LLM models.

We expect most participants to have experience in coding, especially in Python and work with GitHub repositories. It is preferable if participants are familiar with Hugging Face and running local LLM models.

​More information about the workshop will be provided closer to the event. Here is a brief outline:

​18:00 - doors open
18:30 - Introduction
18:45 - Discuession of bias-measuring methods
19:30 - coding and implementation of experiments
20:45 - lighting talks about findings

​We will be working in groups, so if you think you may not have experience with all of the above, it is ok to participate.