Agentic Adversarial QA for Improving Domain-Specific LLMs

TL;DR

This paper introduces an adversarial question-generation framework that enhances domain-specific large language models by creating challenging, semantically rich questions, leading to improved accuracy with fewer synthetic samples.

Contribution

It proposes a novel adversarial question-generation method that improves domain adaptation of LLMs by focusing on interpretive reasoning and sample efficiency.

Findings

Achieves higher accuracy with fewer synthetic samples.

Outperforms baseline methods on LegalBench subsets.

Enhances interpretive reasoning in domain-specific LLMs.

Abstract

Large Language Models (LLMs), despite extensive pretraining on broad internet corpora, often struggle to adapt effectively to specialized domains. There is growing interest in fine-tuning these models for such domains; however, progress is constrained by the scarcity and limited coverage of high-quality, task-relevant data. To address this, synthetic data generation methods such as paraphrasing or knowledge extraction are commonly applied. Although these approaches excel at factual recall and conceptual knowledge, they suffer from two critical shortcomings: (i) they provide minimal support for interpretive reasoning capabilities in these specialized domains, and (ii) they often produce synthetic corpora that are excessively large and redundant, resulting in poor sample efficiency. To overcome these gaps, we propose an adversarial question-generation framework that produces a compact set of semantically challenging questions. These questions are constructed by comparing the outputs of the model to be adapted and a robust expert model grounded in reference documents, using an iterative, feedback-driven process designed to reveal and address comprehension gaps. Evaluation on specialized subsets of the LegalBench corpus demonstrates that our method achieves greater accuracy with substantially fewer synthetic samples.