Retrieve-Plan-Generation: An Iterative Planning and Answering Framework for Knowledge-Intensive LLM Generation

TL;DR

The paper introduces the Retrieve-Plan-Generation framework, an iterative approach that improves knowledge-intensive LLM outputs by planning and selecting relevant information, reducing errors and off-topic content.

Contribution

It proposes a novel iterative planning and answer framework with multi-task prompt-tuning, enhancing relevance and accuracy in knowledge-intensive LLM tasks.

Findings

RPG outperforms baselines across 5 tasks.

Iterative planning improves answer relevance.

Multi-task prompt-tuning enables efficient implementation.

Abstract

Despite the significant progress of large language models (LLMs) in various tasks, they often produce factual errors due to their limited internal knowledge. Retrieval-Augmented Generation (RAG), which enhances LLMs with external knowledge sources, offers a promising solution. However, these methods can be misled by irrelevant paragraphs in retrieved documents. Due to the inherent uncertainty in LLM generation, inputting the entire document may introduce off-topic information, causing the model to deviate from the central topic and affecting the relevance of the generated content. To address these issues, we propose the Retrieve-Plan-Generation (RPG) framework. RPG generates plan tokens to guide subsequent generation in the plan stage. In the answer stage, the model selects relevant fine-grained paragraphs based on the plan and uses them for further answer generation. This plan-answer process is repeated iteratively until completion, enhancing generation relevance by focusing on specific topics. To implement this framework efficiently, we utilize a simple but effective multi-task prompt-tuning method, enabling the existing LLMs to handle both planning and answering. We comprehensively compare RPG with baselines across 5 knowledge-intensive generation tasks, demonstrating the effectiveness of our approach.