It's about weights/parameters, not representations. | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		mlpro 33 days ago \| parent \| context \| favorite \| on: The universal weight subspace hypothesis It's about weights/parameters, not representations.

tsurba 33 days ago [–]

True, good point, maybe not a straightforward consequence to extend to weights.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact