Open is Published a post -mortem On Recent psychophagei problems Strengthens the chatigpt with a default AI model, GPT-4O – cases that forced the company to return the model updated the model last week.
At the end of the week, after the GPT -4O model update, users on social media noted that Chat GPT began to respond in excessive verification and acceptable way. It quickly became a meme. Consumers published Chat GPT screenshots, praising all kinds of problems, Desperate Decision And Welfare.
In a post on X on Sunday, CEO Sam Altman Acknowledged The problem and said that the open will work on the reforms “ASAP”. Two days later, Altman Announced The GPT -4OO update was being brought back and working on “additional reforms” for the personality of the open model.
According to OpeniThe purpose of this update, the purpose of which was to make the model’s default personality “feels more intuitive and efficient”, was made a lot of awareness through “short -term impression” and how to commit consumer communication with Chat GPT over time.
We returned the GPT -4O update in Chat GPT last week as it was excessively flattering and pleased. Now you have access to the first version with more balanced behavior.
What happened about what happened, why it makes a difference, and how are we addressing psychophony: https://t.co/lohou7i7dc
– Openi (@Openi) 30 April, 2025
“As a result, the GPT -4O quarrels towards the reaction that was excessive but unpleasant,” Open said in a blog post. “Psychophantic interactions can cause anxiety, disturbing and anxiety. We are short and working to correct it.”
Open says it is implementing several reforms, including improving its basic model training techniques, and the system clearly indicates the GPT -4O to clearly remove the psychopathy. ۔ [the model’s] It states that “honesty and transparency,” and continue to increase its diagnosis “help identify problems beyond psychoscier.”
Openi also says he is experiencing “real -time views” to consumers with Chat GPT to “directly affect their interaction” and choose from multiple chat GPT personalities.
“[W]The company writes in its blog post that Chat is looking for new ways to add wider, democratic opinions in GPT’s default behavior. […] We also believe that consumers should have more control over the behavior of Chat GPT and, to the extent that it is safe and viable, if they do not agree with the default behavior, make adjustments.