The logistic regression is well-recognized for its capability in dealing with binary dependent variable. It uses a cumulative logistic function transformation to bind the output between 0 and 1. In other words, the logistic model effectively turns real values into probabilities. Suppose we now have 3 features, and the functional form $y_i$ is estimated as: $y_i = \alpha + \beta_1 x_{1i} + \beta_2 x_{2i} + \beta_3 x_{3i}$ What is the probability that $y_i = 0$?