How to interpret the feature importance output of xgboos model when using booster=“gblinear” in R?

Question

I use xgb.importance in R with a linear xgb model(booster="gblinear"). It gives me feature importance scores that look like the following figure:

I