BioBasicsBench is a new benchmark developed in collaboration with the biotechnology company Ligo that tests on 100 tasks derived from real experiments how AI agents interpret raw lab data, identify errors, and recognize when evidence does not support a conclusion. The tool aims to track model development and scientific progress, and to reduce biosafety and biosecurity risks.
AI-generated text
BioBasicsBench: new benchmark for measuring AI agents' analysis of wet-lab routine
BioBasicsBench is a new benchmark developed in collaboration with the biotechnology company Ligo that tests on 100 tasks derived from real experiments how AI agents interpret raw lab data, identify errors, and recognize when evidence does not support a conclusion.
