Writing Like the Best: Exemplar-Based Expository Text Generation

TL;DR

This paper introduces a novel task of exemplar-based expository text generation, proposing a RePA framework that improves coherence and adaptiveness in generated texts using large language models.

Contribution

The paper presents the RePA framework with adaptive imitation and recurrent planning, advancing expository text generation by addressing coherence and topic adaptation challenges.

Findings

RePA outperforms existing methods in factual accuracy and relevance.

New evaluation metrics effectively measure imitation and adaptiveness.

RePA demonstrates strong performance across diverse datasets.

Abstract

We introduce the Exemplar-Based Expository Text Generation task, aiming to generate an expository text on a new topic using an exemplar on a similar topic. Current methods fall short due to their reliance on extensive exemplar data, difficulty in adapting topic-specific content, and issues with long-text coherence. To address these challenges, we propose the concept of Adaptive Imitation and present a novel Recurrent Plan-then-Adapt (RePA) framework. RePA leverages large language models (LLMs) for effective adaptive imitation through a fine-grained plan-then-adapt process. RePA also enables recurrent segment-by-segment imitation, supported by two memory structures that enhance input clarity and output coherence. We also develop task-specific evaluation metrics--imitativeness, adaptiveness, and adaptive-imitativeness--using LLMs as evaluators. Experimental results across our collected three diverse datasets demonstrate that RePA surpasses existing baselines in producing factual, consistent, and relevant texts for this task.