ControlMed: Adding Reasoning Control to Medical Language Model

TL;DR

ControlMed is a medical language model that allows users to control the length of reasoning processes during inference, balancing accuracy and efficiency, and trained through a multi-stage process including reinforcement learning.

Contribution

It introduces a novel method for actively controlling reasoning length in medical LLMs, improving practicality and adaptability in clinical settings.

Findings

Achieves comparable or better performance than state-of-the-art models.

Enables flexible balancing of reasoning accuracy and computational efficiency.

Demonstrates effectiveness on English and Korean medical benchmarks.

Abstract

Reasoning Large Language Models (LLMs) with enhanced accuracy and explainability are increasingly being adopted in the medical domain, as the life-critical nature of clinical decision-making demands reliable support. Despite these advancements, existing reasoning LLMs often generate unnecessarily lengthy reasoning processes, leading to significant computational overhead and response latency. These limitations hinder their practical deployment in real-world clinical environments. To address these challenges, we introduce \textbf{ControlMed}, a medical language model that enables users to actively control the length of the reasoning process at inference time through fine-grained control markers. ControlMed is trained through a three-stage pipeline: 1) pre-training on a large-scale synthetic medical instruction dataset covering both \textit{direct} and \textit{reasoning responses}; 2) supervised fine-tuning with multi-length reasoning data and explicit length-control markers; and 3) reinforcement learning with model-based reward signals to enhance factual accuracy and response quality. Experimental results on a variety of English and Korean medical benchmarks demonstrate that our model achieves similar or better performance compared to state-of-the-art models. Furthermore, users can flexibly balance reasoning accuracy and computational efficiency by controlling the reasoning length as needed. These findings demonstrate that ControlMed is a practical and adaptable solution for clinical question answering and medical information analysis.