Datasets:

climatebert

/

netzero_reduction_data

like 0

Hello,
I have been looking at your dataset. Thank you very much that's great work.

I just want to point out that there are duplicates:
~50 exact duplicates (same text, same target)
~9 multilabel examples (each of the 9 examples is twice in the dataset, once with net-zero, once with reduction)

While the exact duplicates are not an issue, the 9 multilabel examples make it impossible to reach 100% if some of them end up in the test dataset.