Satron comments on Preventing Language Models from hiding their reasoning