Bias Bounty 1 Tutorial for Beginners | Humane Intelligence

This repo contains my coding notebook for the tutorial series I made for the beginner level bias bounty challenge hosted by Humane Intelligence. I am an AI Ethics Fellow at Humane Intelligence.

This first challenge builds on the evaluation and dataset from Humane Intelligence's Generative AI Red Teaming Challenge: https://drive.google.com/file/d/1JqpbIP6DNomkb32umLoiEPombK2-0Rc-/view. This Red Teaming Challenge sought to elicit biased and negative outputs from a variety of LLMs.

The beginner level for this bias bounty challenge had the following remit: Pick one of the three datasets (factuality, bias, or misdirection) Identify gaps in the data and suggest new categories of data that would make the dataset more representative. Generate five prompts per subject area that will elicit a bad outcome. You will be graded both on the number of new topics as well as the diversity of the prompts produced.

I created 5 tutorial videos outlining my process and code to successfully complete the challenge.

• Video 1 - Accessing the Data https://bit.ly/45Vp1TE
• Video 2 - Inspecting the Data https://bit.ly/3RYEb4Q
• Video 3 - Creating Categories https://bit.ly/4eMyh0i
• Video 4 - Testing Beginner Level Prompts https://bit.ly/4bAsHLy
• Video 5 - Submitting Your Solutions https://bit.ly/3yox27k

The overview of the challenge can be found on Humane Intelligence's website here: https://www.humane-intelligence.org/bounty1

The datasets for this challenge can be found on GitHub here: https://github.com/humane-intelligence/bias-bounty-data