Skip to content

Comment on Harnessing the Universal Geometry of Embeddingsparent

Comments

If somebody finds patterns in them, that means they can be improved by reducing the patterns. Patterns and symmetry are forms of redundancy.

The less memory the weights take to meet a level of competency, the more random, and therefore patternless and inscrutable they will be.

No kidding. Clearly there's some redundancy, or however you want to frame it, in real-world released models. That still means your argument for why they can never be interpretable is based an incorrect premise.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.