A research team at OpenAI has developed a method to improve the comprehensibility of text while maintaining the accuracy of language models, and has published the ...