%0 Generic
%A de Winter, Joost
%A Dodou, Dimitra
%A Eisma, Yke Bauke
%D 2024
%T Supplementary data for the paper 'System 2 thinking in OpenAI’s o1-preview model: Near-perfect performance on a mathematics exam'
%U 
%R 10.4121/2e663686-f656-4ff2-bb21-567ba4d4f03e.v1
%K Generative Artificial Intelligence
%K Reasoning
%K ChatGPT
%X <p>The processes underlying human cognition are often divided into two systems: System 1, which involves fast, intuitive thinking, and System 2, which involves slow, deliberate reasoning. Previously, large language models were criticized for lacking the deeper, more analytical capabilities of System 2. In September 2024, OpenAI introduced the O1 model series, specifically designed to handle System 2-like reasoning. While OpenAI's benchmarks are promising, independent validation is still needed. In this study, we tested the O1-preview model twice on the Dutch ‘Mathematics B’ final exam. It scored a near-perfect 76 and 74 out of 76 points. For context, only 24 out of 16,414 students in the Netherlands achieved a perfect score. By comparison, the GPT-4o model scored 66 and 62 out of 76, well above the Dutch average of 40.63 points. The O1-preview model completed the exam in around 10 minutes, while GPT-4o took 3 minutes, and neither model had access to the exam figures. Although O1-preview had the ability to achieve a perfect score, its performance showed some variability, as it made occasional mistakes with repeated prompting. This suggests that the self-consistency method, where the consensus output is selected, could improve accuracy. We conclude that while OpenAI’s new model series holds great potential, certain risks must be considered.</p>
%I 4TU.ResearchData