Glad to hear your additional description for your model. Your LIME results appear fine. Your model has 10 variables. There are only 7 variables significant for that particular data point. LIME simulates data around the target data point that you want to interpret, then build a surrogate linear model. Every data point will have its own set of variables out of your 10 variables. To explain further, I use the 40-variable equation in my post https://medium.com/analytics-vidhya/an-explanation-for-explainable-ai-xai-d56ae3dacd13. Although the global equation has 40 variables, to a particular local point, it may be just a few variables matter. For convenience, I also copy the same 40-variable eq. in this LIME post. Let me know if I address your questions. Happy modeling!