ClinicalRewardModel-Qwen2_5-7B

This is a 5-head reward model fine-tuned on clinical decision-making data using the Qwen2.5-7B-Instruct backbone. It is trained to evaluate clinical vignette–based multiple-choice questions across five expert-defined clinical criteria:

• Clinical Plausibility
• Clinical Utility
• Quality of Decision Path
• Alignment to Decision Path
• Correctness of the Suggested Answer

Each head independently scores one dimension on a 1–5 Likert scale. This model supports fine-grained quality filtering for guideline-based QA datasets such as MedGUIDE-MCQA-8K.