https://eugeneyan.com/writing/finetuning/
utterances-bot opened this issue · comments
Out-of-Domain Finetuning to Bootstrap Hallucination Detection
How to use open-source, permissive-use data and collect less labeled samples for our tasks.
Interesting. Thanks for sharing. Do you have any suggestions if the source document is longer?
If your source document is larger, you could probably try the same approach but using a model with larger context size.
This is a great article, thanks for sharing this approach Eugene.