List of 36 "Hardest" Dataset in Table 4

[JerryLife]

We are planning to use tabzilla as benchmark in our recent paper, but encounter some difficulty to directly compare our results with that in the paper.

There are 36 hardest datasets provided in the Table 4 of tabzilla paper. However, the openml_easy_import_list.txt provides 195 datasets from openml. I tried to search the name in Table 4 in the downloaded 195 datasets, but many of them do not have a precise match. Could you please provide the list of IDs for the 36 hardest datasets?