D-SCoRE: Document-Centric Segmentation and CoT Reasoning with Structured Export for QA-CoT Data Generation

TL;DR

D-SCoRE is a training-free framework that automatically generates diverse, high-quality question-answering datasets from text sources using LLMs and prompt engineering, enhancing domain adaptation for QA models.

Contribution

It introduces a novel, training-free method combining document segmentation, CoT reasoning, and structured export to create tailored QA datasets efficiently.

Findings

LLMs fine-tuned on D-SCoRE data outperform those trained on human data.

The method generates over 1,100 QA pairs per GPU-hour.

D-SCoRE enables rapid domain-specific QA dataset creation.

Abstract

The scarcity and high cost of high-quality domain-specific question-answering (QA) datasets limit supervised fine-tuning of large language models (LLMs). We introduce $\textbf{D-SCoRE}$, a training-free framework that leverages LLMs and prompt engineering to automatically generate diverse, rich QA datasets with Chain-of-Thought (CoT) from arbitrary textual sources. By integrating $\textbf{D}$ocument-centric processing, $\textbf{S}$egmentation, $\textbf{Co}$T $\textbf{R}$easoning, and structured $\textbf{E}$xport - along with multi-dimensional controls such as semantic role transformation, question type balancing, and counterfactual augmentation - D-SCoRE produces tailored QA pairs with enhanced diversity and relevance. LLMs fine-tuned on D-SCoRE-generated datasets outperform those trained on human-annotated QA data across most evaluated domains. Its efficiency and scalability enable rapid, high-performance domain-adaptive fine-tuning on consumer-grade hardware, generating over 1,100 high-quality QA pairs per GPU-hour end-to-end.