A research team has a language model that was pre-trained on a large dataset of movie reviews to classify them as having 'positive' or 'negative' sentiment. They now want to use this model for a new project: classifying short medical summaries into one of three categories: 'Cardiology', 'Neurology', or 'Oncology'. Which of the following describes the most effective procedure for adapting the pre-trained model to this new task?

You are tasked with adapting a language model, which was initially trained to identify the sentiment of customer reviews, for a new purpose: classifying news articles into categories like 'Sports', 'Technology', and 'Politics'. Arrange the following steps in the correct chronological order to successfully adapt the model for this new classification task.