Dispersion loss counteracts embedding condensation in small language models

That sounds very similar to what we known in self-supervised learning to representation collapse. Wonder if we could copy some of the anti collapse mechanisms from SSL into GPT... after all, they are ways to increment the differential entropy. However, I'm not sure if it could be useful after all: any pure function cannot produce more entropy than the entropy it receives... and natural language as text has much less entropy than other domains...

It makes sense to me that distributing across more parameters results in models that can be quant more heavily (information theory - more bits available)

I wonder if anyone has figured out how the information is compressed and calculated the amount of information an LLM can hold depending on its size

Anyone with a billion dollars want to try this and report back?