Comment by p12tic

For all intents and purposes cache may not exist when the working set is 17B or 109B parameters. So it's still better that less parameters are activated for each token. 17B parameters works ~6x faster than 109B parameters just because less data needs to be loaded from RAM.