News for hackers. Stuff that matters.

A comment on LLMs can label data as well as human annotators, but 20 times faster. Parent comment

Re:LLMs can label data as well as human annotators, but 20 t... by nihit-desai on Sunday June 18, 2023 @04:51PM (#36384323)
Good question - one followup question there is value for who? If it is to train the LLM that is labeling, then I agree. If it is to train a smaller downstream model (e.g. finetune a pretrained BERT model) then the value is as good as coming from any human annotator and only a function of label quality
[ Parent | on HN ]
Re:LLMs can label data as well as human annotators, but 20 t... by voz_ on Sunday June 18, 2023 @04:56PM (#36384370)
Why retrain that smaller model from scratch tho? Just do a little transfer learning, or get creative and see if you can prune down to a smaller model algorithmically instead of doing the whole label and train rigamarole from scratch on what is effectively regurgitation.

I’m not sold this has directional value.

[ Parent | on HN ]
Re:LLMs can label data as well as human annotators, but 20 t... by nihit-desai on Sunday June 18, 2023 @05:45PM (#36384840)
Hmm, I'm not suggesting training a smaller model from scratch - in most cases you'd want to finetune a pretrained model (aka, transfer learning) for your specific usecase/problem domain.

The need for labeled data for any kind of training is a constant though :)

[ Parent | on HN ]