Datasets:

bigscience

/

P3

like 184

Hi @bebr2 ,

For example, in "super_glue_cb_GPT_3_style_score_eval", I found the "inputs" from the first example and the second example are the same, but the "targets" from them are different.

You will find the field is_correct useful: it tells you which of the pair (input, target) is the correct one when you do rank evaluation. I think https://huggingface.co/datasets/bigscience/P3/discussions/5#6362ea0c2691058b19deb99a will be helpful in that regard

Then I check "super_glue_cb_GPT_3_style_", "targets_pretokenized" of which is "".

Can you give more details? What split are you looking at? For super glue tasks, labels for the test splits are not available.

Hi @bebr2 ,

For example, in "super_glue_cb_GPT_3_style_score_eval", I found the "inputs" from the first example and the second example are the same, but the "targets" from them are different.

You will find the field is_correct useful: it tells you which of the pair (input, target) is the correct one when you do rank evaluation. I think https://huggingface.co/datasets/bigscience/P3/discussions/5#6362ea0c2691058b19deb99a will be helpful in that regard

Then I check "super_glue_cb_GPT_3_style_", "targets_pretokenized" of which is "".

Can you give more details? What split are you looking at? For super glue tasks, labels for the test splits are not available.

Hi, @VictorSanh , I looked at the test split. I didn't check other splits. Now I know the train split and the validation split can help me test the model. Thank you so much.