Media Outlet:
Nature
Publication Date:
Description:
MIT researchers have demonstrated how an AI model “can interpret the raw internal values of some other models well enough to predict what different parts of the systems do. It can then edit the models to make improvements—including boosting their mathematical abilities,” writes Nature’s Elizabeth Gibney. “The authors hope that, with development, the findings could boost the efficiency of AI models,” notes Gibney. “We may need much less data, much less compute and much less time to teach our models a specific skill,” says graduate student Rikiya Takehi.