英语日课原文存档 · 返回课文

OpenAI

Measuring how AI agents handle everyday wet-lab science

源站完整页面未能保存,以下为课文收录的文本。

Everyday wet-lab work involves failed experiments, ambiguous images and conflicting observations. … The benchmark’s 100 tasks test whether agents can interpret raw experimental data, identify errors, and recognize when the evidence cannot support a conclusion. … Getting these tasks right matters because misreading an image or overlooking a problem with the controls used to check an experiment’s results can lead scientists to draw the wrong conclusion. … The results show progress on routine lab analysis, with room for further improvements in interpreting images, checking controls and applying experimental context.