NEWS
OpenAI releases new framework to monitor and explain model misalignmentsOpenAI introduced a new framework to track, investigate, and publicly disclose misaligned behaviors in its models. This framework includes standards and timelines stating that behaviors can be disclosed to the public even if they have not been fully explained or resolved. This is a news summary compiled for the AiPulse newsletter — not the original article. The original article is at aihot.news. Read the original article AI Critique → English (automatic translation) |