sompe problems in the "gossipcop" dataset.

[mrakhgari]

Hi, I wanted to use the dataset you put in the repository. but the "gossip cop" dataset has a few problems. One problem is that the JSON library in python couldn't load the dataset. cause in some data instances, double quote notation (") has been removed. The number of samples is few, and we can fix them manually.
But the other problem is more critical. There is no bot_score field in this dataset, and it causes the LinguisticFeature to become useless data.