# Paraphrased rules vs templated rules

> On the open-world task, Jev's accuracy ranges from 78.1% (paraphrased (NatLang)) to 86.1% (templated), a spread of 8.0 points. Whether the theory comes from ProofWriter's NatLang set, where people reworded the templated sentences, or from the templated sets.. Every model, both tasks, with 95% intervals.
