Exploring the Capabilities and Limitations of Large Language Models in the Electric Energy Sector
Large Language Models (LLMs) as chatbots have drawn remarkable attention thanks to their versatile capability in natural language processing as well as in a wide range of tasks. While there has been great enthusiasm towards adopting such foundational model-based artificial intelligence tools in all sectors possible, the capabilities and limitations of such LLMs in improving the operation of the electric energy sector need to be explored, and this article identifies fruitful directions in this regard. Key future research directions include data collection systems for fine-tuning LLMs, embedding power system-specific tools in the LLMs, and retrieval augmented generation (RAG)-based knowledge pool to improve the quality of LLM responses and LLMs in safety-critical use cases.
PDF Abstract
