Open Models Almanac
Methodsdata curationtokenization

single mention · data curation

Subtoken-averaged special-token embedding initialization

A new token's embedding is initialized as the mean of its constituent subtoken embeddings.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See tokenization for the methods filed beside it.

Also called subtoken averaging.

Filed under data curation › tokenization