HYU-NLP-EVAL/qwen3-4b-rar-medicine-onlinerubrics-dense-seed11-step-041

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:4BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Oct 1, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

HYU-NLP-EVAL/qwen3-4b-rar-medicine-onlinerubrics-dense-seed11-step-041 is a 4 billion parameter Qwen3 model, specifically a dense checkpoint from the RaR-Medicine OnlineRubrics project. This model is derived from a veRL checkpoint and is intended for research use, focusing on applications within the medical domain. Its primary use case involves tasks related to online rubrics in medicine, leveraging its specialized training for relevant evaluations.

Loading preview...

HYU-NLP-EVAL/qwen3-4b-rar-medicine-onlinerubrics-dense-seed11-step-041 Overview

This model is a 4 billion parameter variant of the Qwen3 architecture, developed as part of the RaR-Medicine OnlineRubrics project. It represents a dense checkpoint from the phase1-online-rubrics-medicine-full-dense-20260919-seed11 run, specifically at step 41 of its training process. The model is provided in BF16 format, suitable for inference.

Key Characteristics

  • Architecture: Based on the Qwen3 model family.
  • Parameter Count: 4 billion parameters, offering a balance between performance and computational efficiency.
  • Specialized Training: Fine-tuned within the RaR-Medicine OnlineRubrics project, indicating a focus on medical domain applications.
  • Checkpoint Type: A "dense" checkpoint, suggesting a comprehensive state of the model's training.
  • Origin: Derived from an original veRL (reinforcement learning) checkpoint, with the raw checkpoint files available in the original_checkpoint/ directory.

Good for

  • Medical Domain Research: Ideal for academic and research purposes in natural language processing within the medical field.
  • Online Rubrics Evaluation: Specifically designed for tasks related to online rubrics in medicine, potentially involving automated assessment or feedback generation.
  • Exploration of veRL Models: Useful for researchers interested in the application and performance of models trained with reinforcement learning techniques in specialized domains.

This model is designated for research use only.

Finetunes
1 models