logoalt Hacker News

ramity • today at 3:33 AM • 1 reply • view on HN

I'm passionate about and talk about this topic a lot. A common misconception is the idea that temperature or entropy must be 0 when seeded entropy enables both creativity and consistency. Attention mechanisms, layer dropout, initial weights, all of it. It can all be driven by seeded entropy. These systems can and have been served in a deterministic manner, but there are a litany of economic drivers against doing so.


Replies

khimaros • today at 4:21 AM

and can still be achieved locally with llama.cpp for example. one extra json parameter.