names of the features

[mishabounty]

Hi there! I use the CORA dataset to train the model and explain it using model interpretation methods (torch_geometric.explain). But in order to fully understand the operation of the model, the names of the features that are encoded in the tensor data.x (cora.x) are missing. Where can I find a dictionary with the names of the encoded features?
For example, as an explanation, I get the importance of CORA features (top 10). But I want to know the real names of the signs (1263, 1355, ...).

[rusty1s]

These are sadly not available, since this dataset is only available as a pre-processed version. Here, the features represent a bag-of-words from the abstract of papers.

[mayabechlerspeicher]

Could you please explain why the fact the dataset is available as a pre-processed version and the features represent a bag-of-words from the abstract of papers makes the features not available?

[rusty1s]

Because I haven't done the pre-processing and I don't have access to the mapping from the 0/1 matrix to the raw words.

[mayabechlerspeicher]

Thank you.
Do you know where the mapping is available? I am having trouble finding it.

[rusty1s]

I don't really know, sorry.