Human and Large Language Model Generated Abstractive Summaries for Technical Discord Conversations

In the ever-evolving realm of software development, the ability to condense information through summarization is crucial for effective information management. This investigation explores various abstractive summarization models to evaluate their efficacy in consolidating developer conversations from Discord, a prominent collaborative platform. Our focus is centered on comparing the summaries generated by these models with human-crafted ones, utilizing Part-of-Speech (POS) tag analysis to assess their proximity to human accuracy and contextual relevance.