con-cord/Mod1_2-no-ref
VISIONConcurrency Cost:1Model Size:4.3BQuant:BF16Ctx Length:32kPublished:May 27, 2026License:apache-2.0Architecture:Transformer Open Weights Warm
con-cord/Mod1_2-no-ref is a 4.3 billion parameter medical LLM-as-a-Judge model, based on the Gemma-3-4B architecture. It is specifically fine-tuned for evaluating generated medical responses against predefined clinical criteria without the use of expert reference answers. This model is optimized for automated assessment of medical text quality and adherence to clinical standards.
Loading preview...
Overview
con-cord/Mod1_2-no-ref is a specialized medical Large Language Model (LLM) designed to function as an "LLM-as-a-Judge." Built upon the Gemma-3-4B base architecture, this 4.3 billion parameter model is fine-tuned for the critical task of evaluating medical responses.
Key Capabilities
- Medical Response Evaluation: The primary function is to assess the quality and clinical adherence of generated medical answers.
- No Reference Answers: This specific version (Mod1_2-no-ref) is engineered to perform evaluations without relying on expert reference answers, making it suitable for scenarios where such references are unavailable or impractical.
- Clinical Criteria Adherence: Designed to judge responses based on predefined clinical evaluation criteria.
Good For
- Automated quality assurance of AI-generated medical text.
- Scenarios requiring objective evaluation of medical responses where expert references are not provided.
- Developing systems that need to assess the clinical soundness of language model outputs in a medical context.