Paper ID: 2311.09136

Rescue: Ranking LLM Responses with Partial Ordering to Improve Response Generation

Yikun Wang, Rui Zheng, Haoming Li, Qi Zhang, Tao Gui, Fei Liu

Customizing LLMs for a specific task involves distinguishing effective responses from erroneous ones. This skill can be developed using supervised fine-tuning with extensive human preference data. However, obtaining expert-annotated preference data is expensive for most tasks. In this paper, we present a novel method to optimize LLMs using ranking metrics. This method trains the model to prioritize the best responses from a pool of candidates created for a particular task. Rather than a traditional full ordering, we advocate for a partial ordering, as achieving consensus on the perfect order of candidate responses can be challenging. Our partial ordering is more robust, less sensitive to noise, and can be achieved with limited human annotations or through heuristic methods. We test our system's improved response generation ability using benchmark datasets, including the latest multi-document question answering task. We conduct ablation studies to understand crucial factors, such as how to gather candidate responses for specific tasks, determine their most suitable order, and balance supervised fine-tuning with ranking metrics. Our approach, named \textsc{Rescue}, suggests a promising avenue for enhancing LLMs' contextual understanding via response ranking.

Submitted: Nov 15, 2023