Openai explains why Chatgpt Sycophant is very very

Openai has Spread on Recent sycophance issues With the default AI style that runs ChatGPT, GPT-4O – The issues that forced the company to a decline in update of the model that was released last week.

During the weekend, after updating the GPT-4O, users on social media indicated that Chatgpt began to respond in an excessive way to verify healthy and acceptable. Soon Mimi became. Publishing users from ChatGPT screen coincides with all kinds of problems, dangerous Decisions and Ideas.

On Sunday’s publication, CEO Sam Al -Taman Recognized The problem said that Openai will work on repairs “as soon as possible.” Two days later, Taman Declare The GPT-4O update has been retracted and that Openai was working on “additional repairs” of the model’s personality.

According to OpenaiThe update, which was aimed at making the virtual model character “easier and effective”, was informed more than that through “short -term reactions” and “he did not completely explain how user interactions develop with Chatgpt over time.”

Openai wrote in a blog post: “As a result, GPT -4O was granted towards responses that were very supportive but deceitful.” “Sikovantic reactions can be uncomfortable, disturbing, causing distress. We have shortened and worked to correct it.”

Openai says it is implementing many repairs, including improving basic typical training technologies and system demands to direct GPT-4O on SYCOPHANCY. (The system’s claims are the initial guidelines that direct the comprehensive behavior of the model and color in the interactions.) As the company says from the safety handrails “to increase honesty (model)”.

Openai also says it explores ways to allow users to grant “actual time comments” direct effect on their reactions “with Chatgpt and choose from multiple” characters “.

The company wrote in its blog publication: “(w) Explore new ways to integrate broader and democratic notes in the virtual behaviors.” “We also believe that users should have more control of how ChatGPT behaves, to the extent that it is safe and possible, make adjustments if they do not agree with the virtual behavior.”