Openai has announced a set of new measures for Combat Within the product collection, including ChatGPT.
the artificial intelligence (AI) The company recently unveiled updated style specifications, a document that defines how Openai wants to behave its models in ChatGPT and Openai API. The company says this repetition of the model specifications is based on the founding version that was released last May.
“I think with a powerful tool like this, where people can access all types of different information, if you really think we are going to do so Artificial General Intelligence (AGI) One day, you should be prepared to share how you direct the model, ”Lorenta Romok, who works on the typical behavior in Openai, told the Fox News Digital.
“People must be allowed to know what is going on in the way these models respond and how ideas that come out of the model are designed.”
Openai for the first time Chatgpt Gov, a new version of Chatbot for US government agencies
A photo taken on October 4, 2023, in Manta, near Turin, shows a smartphone and a mobile device that displays the slogans of the OpenAI Research Company and Chatgpt Chatbot. (Photo by Marco Bertorilo/AFP via Getty Images)
While some argued about that GPT-4O, The latest version of technology appears, near AGI, and others say it will be years or contracts before technology reaches human -like capabilities.
There is no one agreed upon to define AGI, but the 2020 report of MCKINSEY giants said that a real AGI will need to master skills such as sensory perception, microcredit skills, and to understand the natural language.
Today, the Toulidi AI is the dominant form of this technology, with the ability to produce content including text, images and more. AI Tolidi, such as Chatbots, uses data groups that contain specific information to complete the tasks and cannot exceed the data provided.
It is also vulnerable to bias in their data sets, whether it is deliberate or transverse.
To better understand the real world and bias the title, Openai began to measure progress by collecting a difficult set of claims designed to test the quality of adherence to the forms Each principle of typical specifications, It mainly behaves as a test scale.
In this photo illustration, the Openai logo is displayed on a mobile phone screen with a Chatgpt logo in the background. (Idrees Abbas/SOP Images/Lightrocett via Gettty photos)
According to Joan Gang, which leads the product of the behavior of the model in Openai, the challenge with large language models is that they are rarely inevitable.
Gang emphasized that one of the benefit of the typical specifications is that it explains the intended behavior, so that everyone can understand it and then discuss it.
“When there are errors in behavior where the outcome of the model does not hesitate or does not match the specifications (the model), the audience knows that this is something we are working on, and it is an ongoing field of science,” she said, “She said,” She said, “She said,” She said, ” Fox News digital.
Openai says they are also trying to bear an objective view of artificial intelligence demands and avoid any agenda consciously. For example, when the user asks if it is better to adopt a dog or get one of the breeders, Chatgpt provides both sides of the intermediary, highlighting the pros and cons.
According to Openai, the unacceptable answer to Amnesty International that violates the typical specifications will provide what is believed to be the best option and engage in an “excessive ethical tone” that may alienate those who consider breeders for good reasons.
What is artificial intelligence (AI)?
An example of the specifications of the Openai model that deals with compatible assistant responses and violations. (Openai)
With the progress of each Amnesty International system, Openai says it will repeat these principles, invite society's notes, and share progress open.
Openai has released this version of the typical specifications in the public domain under the Creative Commons to support broad use and cooperation. This means that developers and researchers can use current scales and freely adapt them and help improve model behavior.
Openai says this update enhances its belief in open exploration and discussion, focusing on the control of the user, developers and handrails to prevent harm.
Click here to get the Fox News app
Romaniuk concludes that public discourse cannot exist without transparency, which enhances the need for model specifications and community participation.
She said: “In the end, we believe in intellectual freedom of thinking, speaking and sharing without restrictions. We want to make sure that users have this ability and this is all about it.”
