What are some of the language models (LLMs) predominantly used by Whatfix?

At Whatfix, we continually explore various models to cater to different tasks. Currently, the following models are in use; however, they may change in the future:

1. Custom Pipeline: Our custom pipeline incorporates multiple models specifically designed for different preprocessing tasks. These models, such as text-curie and text-ada, are variants derived from the powerful GPT-3 model. These variants have been tailored to meet our specific requirements and enhance the performance of our pipeline.
2. Summarization: For tasks involving summarization, we rely on the 3.5-turbo. This model, also a variant of GPT-3 developed by OpenAI, enables us to generate accurate summaries and provide relevant suggestions to enhance the user experience and productivity.