Evaluating the performance of state-of-the-art esg domain-specific pre-trained large language models in text classification against existing models and traditional machine learning techniques