Scientific Understanding Benchmark (SUB)

Emerging

1papers using it

2023first seen

The Scientific Understanding Benchmark (SUB) is a set of tests designed to evaluate and compare the levels of scientific understanding in both humans and artificial intelligence by assessing abilities such as information retrieval, explanation production, and inferential reasoning.

🔎 Find this dataset

Papers using Scientific Understanding Benchmark (SUB) (1)

Towards a Benchmark for Scientific Understanding in Humans and Machines2023 · 1 cites