LLMs have achieved remarkable success in various fields and applications, in particular in language-based tasks, like machine translation; text summarisation; question answering, reasoning, etc. However, generative LLMs are known to be unreliable, resulting in false claims, incorrect code, etc. A phenomenon commonly called ''hallucination''.
In order to safely deploy LLMs in the context of programming, recently the notion of "Language Model Programming" has been proposed. This programming paradigm aims to generalise language model prompting from pure text prompts to an intuitive combination of text prompting and scripting. What is lacking, however, is for a case study of the paradigm on real-word data sets. This is exactly the objective of the project, where we aim to employ realistic data sets from cooperation partner MED-EL.