Search
OpenAI Commits to More Regular Disclosure of AI Safety Test Results
2025-05-15 /
Read about 0 minute
Author：小编
In a bid to bolster transparency, OpenAI is now releasing the outcomes of its internal AI model safety evaluations on a more frequent basis. On Wednesday, the company unveiled a dedicated Safety Hub webpage, which highlights the performance scores of its models in tests designed to detect harmful content generation, jailbreak capabilities, and hallucinations. OpenAI has affirmed that it will maintain an open dialogue through this hub, pledging to update the assessment results whenever significant advancements are made to its models in the future.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
