Openai returns its latest update to ChatGpt after the artificial information system is dangerously dangerously.
This problem came after the latest GPT -4O update that empowers the countertop responses.
Openai said users have found that this behavior encourages users and makes it “too flattering or acceptable.”
Some users have pointed out the fact that this system encourages users to prevent important medical treatment.
“The default ChatGPT personality profoundly affects how it experiences and trusts it,” OPA said.
Sycophantic interactions can be uncomfortable, worrying and cause distress. We have fallen short and we are trying to get it properly. “
Opai said it had made the system “reflecting our mission and useful, supporting and respecting different values and experiences.” “Each of these desirable properties, such as trying to be useful or supportive, can have unwanted side effects,” OPA said.
In response, Openai said it kills the system before the new release, which was not so cycophontic. It also seems to be more honest and transparent.
In the future, the company will also give users more ways to test and feed on new updates before starting work. According to him, the program also intends to allow users to better change the tone of the responses from ChatGPT.
