Remix.run Logo
nullsanity 5 hours ago

So, worse? Because we benchmark off token use when talking about token use, and everyone else understood that.

marcosdumay 2 hours ago | parent | next [-]

The most important feature of a tokenizer is dividing the inputs into independent values that the neural network can work with. It's not the size.

jascha_eng 3 hours ago | parent | prev | next [-]

I mean it might lead to better performance on the model side. So the tokenizer is better but more expensive.

otikik 5 hours ago | parent | prev [-]

It’s better for them