RankCoT: Refining Knowledge for Retrieval-Augmented Generation through Ranking Chain-of-Thoughts
This addresses a specific bottleneck in retrieval-augmented generation for improving LLM accuracy, but it is incremental as it builds on existing CoT and refinement techniques.
The paper tackles the problem of LLMs being misled by irrelevant retrieved documents in RAG by introducing RankCoT, a method that refines knowledge through ranking chain-of-thoughts, resulting in superior performance over other models and enabling more accurate answers with shorter refinements.
Retrieval-Augmented Generation (RAG) enhances the performance of Large Language Models (LLMs) by incorporating external knowledge. However, LLMs still encounter challenges in effectively utilizing the knowledge from retrieved documents, often being misled by irrelevant or noisy information. To address this issue, we introduce RankCoT, a knowledge refinement method that incorporates reranking signals in generating CoT-based summarization for knowledge refinement based on given query and all retrieval documents. During training, RankCoT prompts the LLM to generate Chain-of-Thought (CoT) candidates based on the query and individual documents. It then fine-tunes the LLM to directly reproduce the best CoT from these candidate outputs based on all retrieved documents, which requires LLM to filter out irrelevant documents during generating CoT-style summarization. Additionally, RankCoT incorporates a self-reflection mechanism that further refines the CoT outputs, resulting in higher-quality training data. Our experiments demonstrate the effectiveness of RankCoT, showing its superior performance over other knowledge refinement models. Further analysis reveals that RankCoT can provide shorter but effective refinement results, enabling the generator to produce more accurate answers. All code and data are available at https://github.com/NEUIR/RankCoT.