[GPT-4 becomes stupid? FAP, June 2 / PRNewswire-FirstCall-Asianet /-- with the increasing use of GPT-4, a large number of users have reported that the response quality of the large model has declined recently, especially in program generation, the code generated by GPT-4 is often wrong. A few days ago, Logan Kilpatrick, the ambassador for the promotion of OpenAI developers, responded to the question, saying that since the release of GPT-4 on March 14, the ontology of the big model has been static and there is no contamination of the model by a large amount of external data. At the same time, he also admitted that because of the instability of the large model itself, the answers to the similar cues were inconsistent.