con-cord/Mod1_2-with-ref
VISIONConcurrent Unit Cost:1Model Size:4.3BQuant:BF16Context Size:32kPublished:Jul 17, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold
con-cord/Mod1_2-with-ref is a 4.3 billion parameter medical LLM-as-a-Judge model based on Gemma-3-4B. It is specifically fine-tuned for evaluating generated medical responses against predefined clinical criteria, utilizing expert reference answers. This model excels at assessing the quality and accuracy of medical text, making it suitable for automated medical content review and quality assurance.
Loading preview...
Mod1_2-with-ref: Medical LLM-as-a-Judge
This model, developed by con-cord, is a specialized 4.3 billion parameter medical LLM-as-a-Judge, built upon the Gemma-3-4B architecture. It is fine-tuned using PEFT/LoRA and TRL frameworks to perform automated evaluation of medical responses.
Key Capabilities
- Medical Response Evaluation: Designed to assess the quality and adherence of generated medical answers to specific clinical evaluation criteria.
- Reference-Based Assessment: This particular version (
with-ref) leverages expert reference answers to enhance the accuracy and reliability of its evaluations. - Gemma-3-4B Foundation: Benefits from the robust capabilities of the Gemma-3-4B base model, adapted for medical domain tasks.
Good For
- Automated quality assurance of medical text generation.
- Evaluating the clinical accuracy and relevance of AI-generated medical content.
- Applications requiring an LLM to act as a judge for medical information, especially when reference answers are available.