replicate experiments on the LLama model.

[xyzCS]

Hello, thank you for sharing your code. Now I want to replicate your experiments on the LLama model, mainly to experiment on the compound_words dataset. However, the data you provided should be based on the Pythia model, so I need to rebuild it for the LLama dataset. I noticed that there is a file called make_feature_datasets.py, but when I tried to run this file to build the data, I found that the file "PILE_TEST_PATH = '/home/gridsan/groups/maia_mechint/datasets/pile-test.hf'" is missing. Can you please provide this file? Or if I need to replicate your experiments on LLama, do you have any suggestions?