Hi!
I understand that you used a subset of 15k examples of the Anthropic/hh-rlhf dataset to fine-tune this model.How did you make this split and is this dataset available anywhere?
Thank you!
· Sign up or log in to comment