Huggingface Datasets do not match website datasets

[tbarton16]

My submission got flagged for missing entries. Turns out those entries are not in the HF version of zeroscrolls. To reproduce:

dataset = load_dataset("tau/zero_scrolls", task, split="test")
for d in dataset:
     if d['problem_id'] =="crs_R46013":
          print('contains') # will never print.  ```

[UriSha]

Hi,
Actually this entry does exist in the GovReport test set in HF.
The id is "crs_R46013" and pid is "crs_R46013_0".