Sparse autoencoders have become a popular technique to identify interpretable concepts in language models. They have been successfully applied to several models of varying sizes, including both open and commercial ones, and have become one of the main avenues for interpretability research. A number of approaches have been proposed to extract latent representations from models, as well as automatically provide natural language explanations for the concepts they supposedly represent. Despite these advances, little attention has been given to applying sparse autoencoders to Italian language models, possibly due to the small number of Italian models and the costs associated with leveraging sparse autoencoders, including training and parsing a large number of features. In this work, we present an initial step toward addressing this gap. We train a sparse autoencoder on the residual stream of a language model, release the weights, and leverage an automated interpretability pipeline based on large language models to evaluate both the quality of the latent representations and provide explanations for some of them. We show that, albeit the approach has several limitations, some concepts can be identified in the model's weights.