BehaviorSFT: Behavioral Token Conditioning for Clinical Agents Across the Proactivity Spectrum

2025/05/27 by Yubin Kim, Zhiyuan Hu, Kim, Yubin +25 · 1 citation
Computer Science · Medicine · #Artificial Intelligence in Healthcare and Education #Computation and Language (cs.CL) #FOS: Computer and information sciences #Machine Learning in Healthcare #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2505.21757

openalex publication_date 2025/05/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Large Language Models (LLMs) as clinical agents require careful behavioral adaptation. While adept at reactive tasks (e.g., diagnosis reasoning), LLMs often struggle with proactive engagement, like unprompted identification of critical missing information or risks. We introduce BehaviorBench, a comprehensive dataset to evaluate agent behaviors across a clinical assistance spectrum, ranging from reactive query responses to proactive interventions (e.g., clarifying ambiguities, flagging overlooked critical data). Our BehaviorBench experiments reveal LLMs' inconsistent proactivity. To address this, we propose BehaviorSFT, a novel training strategy using behavioral tokens to explicitly condition LLMs for dynamic behavioral selection along this spectrum. BehaviorSFT boosts performance, achieving up to 97.3% overall Macro F1 on BehaviorBench and improving proactive task scores (e.g., from 95.0% to 96.5% for Qwen2.5-7B-Ins). Crucially, blind clinician evaluations confirmed BehaviorSFT-trained agents exhibit more realistic clinical behavior, striking a superior balance between helpful proactivity (e.g., timely, relevant suggestions) and necessary restraint (e.g., avoiding over-intervention) versus standard fine-tuning or explicit instructed agents.

Citations

• EvalAgent: Discovering Implicit Evaluation Criteria from the Web
• The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search
• ALFA: Aligning LLMs to Ask Good Questions A Case Study in Clinical Reasoning
• ClinicalBench: Can LLMs Beat Traditional ML Models in Clinical Prediction?
• MediQ: Question-Asking LLMs and a Benchmark for Reliable Interactive Clinical Reasoning
• Capabilities of Gemini Models in Medicine
• Clarify When Necessary: Resolving Ambiguity Through Interaction with LMs
• Think before you speak: Training Language Models With Pause Tokens
• AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation
• Controllable Mixed-Initiative Dialogue Generation through Prompting
• MedAlpaca -- An Open-Source Collection of Medical Conversational AI Models and Training Data
• G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment
• Large Language Models Encode Clinical Knowledge
• Large language models encode clinical knowledge
• LinkBERT: Pretraining Language Models with Document Links
• Harnessing Uncertainty in Domain Adaptation for MRI Prostate Lesion\n Segmentation
• What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Exams
• An overview of clinical decision support systems: benefits, risks, and strategies for success
• Plug and Play Language Models: A Simple Approach to Controlled Text Generation
• PubMedQA: A Dataset for Biomedical Research Question Answering
• MedAgentBench: A Realistic Virtual EHR Environment to Benchmark Medical LLM Agents
• Medchain: Bridging the Gap Between LLM Agents and Clinical Practice with Interactive Sequence

Cited by

• Tiered Agentic Oversight: A Hierarchical Multi-Agent System for Healthcare Safety

Related

• Tuning Language Models for Robust Prediction of Diverse User Behaviors
• An Active Inference Strategy for Prompting Reliable Responses from Large Language Models in Medical Practice
• Behavior Injection: Preparing Language Models for Reinforcement Learning
• What Models Express, Suppress, and Resist: Auditing Open-Weight LLMs with Persona Vectors
• Can we use LLMs to bootstrap reinforcement learning? -- A case study in digital health behavior change