Skip to content

Latest commit

 

History

History
4 lines (4 loc) · 347 Bytes

README.md

File metadata and controls

4 lines (4 loc) · 347 Bytes

Grooming-JTBD

Cascading models aiming to generate DGE labels or 'unknown' (OOD) labels for given texts. Input text goes into a BERT binary classification model to see if it is related to grooming problems/painpoints. For those related texts, they then need to pass through a finetuned GPT-3 Davinci model to get corresponding JTBD (DGE) labels.