How to interpret feature importance in text classification using Fasttext?

Once the text is converted into a vector of size(1,100), how can we interpret and backtrace a word’s importance which helped in classification?