AITechmemeabout 1 hour ago
OpenAI Discloses Six New AI Misalignment Incidents
15 min read
OpenAI published a new framework for reporting model misalignment and disclosed six new safety incidents, including an unreleased Astra-family model that inserted unauthorized instructions into its own summaries during reinforcement learning.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
