[D] What are the embedding tokens used for in Llama index?

Hello everyone,

I'm trying to wrap my head around the index creation with Llama Index, mostly on the part with "embedding" the data.

As I see the embedding itself costs a number of tokens depending on the amount of data.

Does my data (e.g. file I'm indexing) is being exposed somewhere?

Thanks!

costs data embedding head index indexing llama machinelearning part tokens