A Comparison Of LLM Finetuning Methods & Evaluation Metrics With Travel Chatbot Use Case

Abstract

This research compares large language model (LLM) fine-tuning methods, including Quantized Low Rank Adapter (QLoRA), Retrieval Augmented fine-tuning (RAFT), and Reinforcement Learning from Human Feedback (RLHF), and additionally compared LLM evaluation methods including End to End (E2E) benchmark method of "Golden Answers", traditional natural language processing (NLP) metrics, RAG Assessment (Rag