ChatGPT 4.5 makes logical mistake when asked not to show its thinking

medium.com

2 pointslogicallee1 comment

At State of Utopia, we recently found this vulnerability in logical thinking that arises in ChatGPT 4.5 when asked not to show its thinking.