synthetic-instruct-gptj-pairwise pairwise data how to pre-process for train data

#9
by chaochaoli - opened

All models are train on these dataset with a same split seed across datasets (if validation split wasn't available)

1、webgpt_comparisons
2、summarize_from_feedback
3、synthetic-instruct-gptj-pairwise
4、anthropic_hh-rlhf
all these data have different format,how to Processed into a unified form?
thks

hi

Sign up or log in to comment