Domaino1s: Guiding LLM Reasoning for Explainable Answers in High-Stakes Domains

TL;DR

Domaino1s enhances large language models' reasoning in high-stakes domains by fine-tuning with domain-specific datasets, employing tree search for solution exploration, and introducing a new explainability metric, leading to improved performance and transparency.

Contribution

This work introduces Domaino1s, a novel framework combining supervised fine-tuning, tree search, and a new explainability metric to improve LLM reasoning and explainability in high-stakes domains.

Findings

Outperforms existing models in stock investment and legal QA tasks.

Provides more explainable and confident answers in high-stakes domains.

Demonstrates the effectiveness of Selective Tree Exploration and PROOF-Score metrics.

Abstract

Large Language Models (LLMs) are widely applied to downstream domains. However, current LLMs for high-stakes domain tasks, such as financial investment and legal QA, typically generate brief answers without reasoning processes and explanations. This limits users' confidence in making decisions based on their responses. While original CoT shows promise, it lacks self-correction mechanisms during reasoning. This work introduces Domain$o1$s, which enhances LLMs' reasoning capabilities on domain tasks through supervised fine-tuning and tree search. We construct CoT-stock-2k and CoT-legal-2k datasets for fine-tuning models that activate domain-specific reasoning steps based on their judgment. Additionally, we propose Selective Tree Exploration to spontaneously explore solution spaces and sample optimal reasoning paths to improve performance. We also introduce PROOF-Score, a new metric for evaluating domain models' explainability, complementing traditional accuracy metrics with richer assessment dimensions. Extensive experiments on stock investment recommendation and legal reasoning QA tasks demonstrate Domaino1s's leading performance and explainability. Our code is available at https://github.com/Hyalinesky/Domaino1s.