PaECTER embeddings for all patent families as of Patstat 2023 Fall release. File "PaECTER_EPO_patstat-2023-fall.parquet" contains embeddings for all patent families with an EPO member (3.9M). File "PaECTER_USPTO_patstat-2023-fall.parquet" contains embeddings for all patent families with a USPTO member (9.2M). We encoded one representative member of the patent family according to the following rule: has English abstract Authority Priority rule: EP > WO > US > CN > JP > KR > GB > CA > TW > DE > ES > RU > AU > FR > MX > UA > NZ (for authorities not listed, the earliest member takes priority) if multiple documents qualify, the first listed (according to Patstat) takes priority