📄️ Zero-Shot Text Classification
One of the powerful features of LLMs is the ability to perform text classification without being re-trained. For that, the only requirement is that the labels must be descriptive.
📄️ Few-Shot Text Classification
Few-shot text classification is a task of classifying a text into one of the pre-defined classes based on a few examples of each class. For example, given a few examples of the class positive, negative, and neutral, the model should be able to classify a new text into one of these classes.
📄️ Dynamic Few-Shot Text Classification
Dynamic Few-Shot Classification is an extension of Few-Shot Text Classification that is more suitable for larger datasets. Instead of using a fixed set of examples for each class, it constructs a dynamic subset for each sample on the fly. This allows to efficiently utilize the limited contex window of the model and save the number of consumed tokens.
📄️ Tunable Text Classification
Tunable estimators allow to fine-tune the underlying LLM for a classification task. Usually, tuning is performed directly in the cloud (e.g. OpenAI, Vertex), therefore it is not required to have a GPU on your local machine. However, be aware that tuning can be costly and time-consuming. We recommend to first try the in-context learning estimators, and only if they do not provide satisfactory results, to try the tunable estimators.