I am trying to figure out how to interpret the what the weights of a binary text sentiment SGD classifier. The classifier was trained on Negative Sentiment as 0 and Positive