ELSER for long texts

I am using Elastic Search 8.11 and want to use ELSER to incorporate semantic search. But fields I have to perform search is long text fields.

I found this link saying

If your data set contains long documents, divide them into smaller segments before ingestion if you need the full text to be searchable.

But I am not able to figure out and find reference that how exactly that is supposed to be done.

Can someone share more details on this?