, which combines elements of supervised and unsupervised techniques. This method employs a small quantity of labeled data and a bigger level of unlabeled data, thereby improving upon learning accuracy though lowering the need for labeled data, that may be time and labor intense to obtain.Analyzed products show a 70B parameter LLM leveraging RAG for