Machine learning is a subsection of artificial intelligence that has revolutionized recent decades with its remarkable performance when compared to the conventional state-of-the-art methods in the fields of science and information technology. However, a trained machine learning model is regarded as a black box model due to the non-transparent nature of its internal workings. This lack of transparency reduces the interpretability of trained machine learning models, making them less desirable in life-critical applications such as aerodynamics or the medical industry. Because of their uncertainty, these models cannot be employed until their internal workings have been thoroughly explored. As a result, understanding their internal working knowledge is critical. To solve this problem, visualization provides the ability to understand the internal structure and workflow of trained machine learning models. Meanwhile, the interpretability of neural networks is tedious due to the large number of parameters. In particular, the internal hidden state information of the recurrent neural network models is yet to be explored. Hence, this thesis introduces a visual analytics tool for analyzing and examining the internal hidden state information of recurrent neural network models. This tool allows an interactive exploration of the hidden state information, which is processed by the neural layers when an input sequence is provided to it. This tool aids in exploring the behavior of a recurrent neural network model when demonstrated with turbulence data from an aerodynamics domain. To achieve better interpretability, we assess and compare the training of the gated recurrent unit model for a number of hyperparameter configurations using various visualization approaches such as dimensionality reduction, heat maps, and parallel coordinates plots.