SIGNATURE WORK
CONFERENCE & EXHIBITION 2024

As artificial intelligence (AI) large language models achieve significant breakthroughs in natural language processing (NLP), the application of these models for text sentiment analysis has attracted wide attention across academia and industry. Prior research utilized traditional dictionary-based tools for comprehensive sentiment analysis of social media discourse surrounding CryptoPunks, yet it did not incorporate the more advanced AI analytical methodologies. This study leverages cutting-edge, open-source large language model technologies for a nuanced examination of the sentiment within CryptoPunks’ social media conversations. This analysis not only corroborates the findings of earlier empirical investigations but also assesses the reliability of AI technologies in sentiment analysis of non-fungible token (NFT) related social media discourse. Our findings indicate that large language models tend to identify a higher prevalence of neutral sentiments in social media discussions compared to traditional NLP tools. This underscores the significant influence that the selection of NLP tools exerts on the outcomes of sentiment analyses. The insights derived from this study offer valuable contributions to the interdisciplinary fields of artificial intelligence, blockchain technology, and social media, furnishing both theoretical and practical directions for enhancing sentiment analysis methodologies in the context of NFT applications with large language models.