Hello, I am working with the new implementation of GPT-5 as shown in screenshot 1, but I must report that I am repeatedly encountering inaccuracies in the translation of texts, so that in the final output, terms in the original language, in this case English, can be found continuously, as shown in screenshot 2. This does not change even when changing the style from Authoritative to Informational. I am reporting this because I do not believe it is a desired or expected effect, and it is frustrating to have to manually correct all the generated texts.