According to Heise Online, 70.000 questions were submitted, among which 3000 questions made it into the final dataset of questions from all scientific disciplines.
Beside a prize money of $500, the authors of the top-rated 550 questions are included as co-authors of the publication. As of today, no state-of-the-art LLM model is able to solve even 10% of the questions present in the dataset, the New York Times writes When A.I. Passes This Test, Look Out.
The only contributors from RUB are Prof. Christian Stump and Prof. Alexander Ivanov, congratulations!
