Questions about the items in the prompt dataset

[JunyiChE]

Hi,
Very impressive and interesting work.

However, I have one question about the prompt dataset. The number of prompts in the dataset uploaded seems much less than that claimed in the paper.

For example, in the table 3 of the paper, it is said Amazon prompt dataset contains 2,252,890 prompts. However, in the repository, I can only find 1600 items in the training set and 160 items in the test set. Is there anything I missed or misunderstood?

Thanks for your clarification and looking forward to your kind reply!

[JunyiChE]

Table that describes the prompts number in the paper