Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
OPENAI launched a new web page called the safety assessment hub Publicly share information related to things such as the hallucination rates of your models. The hub will also highlight whether a model produces harmful content, how well it behaves as indicated and attempted jailbreaks.
The technological company says that this new page will provide additional transparency on Openai, a company which, for the context, has been faced multiple prosecution alleging that he has illegally used material protected by copyright to form his AI models. Oh, yeah, and it’s worth mentioning this The New York Times asks the technological company Accidental deleted evidence In the case the newspaper’s plagiarism against her.
The security assessment center is intended to develop OPENAI system cards. They only describe the safety measures of a development development, while the center should provide current updates.
“While the science of AI evaluation is evolving, we aim to share our progress in the development of more evolutionary means to measure the capacity and security of the model”, the OpenAi states in its ad. “By sharing a subset of our safety assessment results here, we hope that this will not only facilitate the understanding of the safety performance of OpenAi systems over time, but will also support community efforts to increase transparency through the field.” Openai adds that his work to have more proactive communication in this area of the company.
Presentation of the safety assessment center – a resource to explore security results for our models.
While the system cards share security measures at launch, the HUB will be updated periodically as part of our efforts to communicate in a proactively security.https://t.co/c8ngmxlc2y
– Openai (@openai) May 14, 2025
Interested parties can examine each of the center sections and see information on relevant models, such as GPT-4.1 through 4.5. OPENAI notes that the information provided in this center is only an “snapshot” and that interested parties should examine its system cards. Evaluations and other versions for more details.
One of the big but For the entire security assessment, OpenAi is the entity that makes these tests and the choice of information to be shared publicly. Consequently, there is no way to guarantee that society will share all its problems or concerns with the public.
(Tagstotranslate) llm
Source link
