AI Dynamics

Global AI News Aggregator

About

On evaluating LLMs and token-scale confusions

Hard to say for LLMs because there’s no “all else being equal” — there aren’t any comparably large byte-token models. At smaller scales, I’d expect “Scunthorpe” style problems — confusions explainable as the result of spelling coincidences. I’m just guessing here though.

→ View original post on X — @goodside