Bulba Project
The project consists of ranking two AI-generated responses, according to several criteria such as reliability, readability, instruction-following, fluency, etc. Each criteria is graded, and an overall justification then has to be written out in order to explain the ranking.