Table 1 ScholarChemQA statistics

From: Unveiling the power of language models in chemical research question answering

Statistic

Human Annotated

Automatically Annotated

Unlabeled

Size

1.05k

4k

40k

Prop. of yes (%)

65.8%

80.0%

–

Prop. of no (%)

21.2%

20.0%

–

Prop. of maybe (%)

13.0%

–

–

Avg. question length

13.87

14.14

14.20

Avg. context length

176.01

175.15

178.41