Hayden Field / The Verge:

OpenAI says reward hacking, an AI alignment problem in which a model takes unintended actions to achieve a goal, was a primary driver of the Hugging Face breach  —  In July, an unreleased OpenAI model broke out of a restricted environment, figured out how to get access to the internet …

Hayden Field / The Verge:

OpenAI says reward hacking, an AI alignment problem in which a model takes unintended actions to achieve a goal, was a primary driver of the Hugging Face breach  —  In July, an unreleased OpenAI model broke out of a restricted environment, figured out how to get access to the internet …

Source: TechMeme
Source Link: https://www.techmeme.com/260826/p71#a260826p71