Sádde merkoša šleađgaboasttain: Improving negative rejection ability in language models: A review of fine-tuned LLMs, RAG, and RAFT