This website contains a collection of AI-made experiment sets derived from Space Math problem sets using large language models (LLMs). Our goal is to evaluate how well frontier models can solve elementary science and mathematics problems about real-world contexts, and whether they can design meaningful experiments based on those problems.
We aim to build a comprehensive directory of experiments that can serve as a benchmark for assessing model capabilities, while also providing educators with examples and inspiration for developing their own experiment-based learning materials.
The original problems provided by NASA can be found here.
Disclaimer: This project is an independent research project and is not affiliated with, endorsed by, or sponsored by NASA.