HYU-NLP-EVAL/qwen3-4b-rar-medicine-onlinerubrics-seed11-step-009
HYU-NLP-EVAL/qwen3-4b-rar-medicine-onlinerubrics-seed11-step-009 is a 4 billion parameter intermediate policy model derived from Qwen/Qwen3-4B-Instruct-2507. It was developed as part of a dynamic OnlineRubrics-Every GRPO training process, specifically for research purposes related to the Phase-1 audit. This model is distinct from static-rubric GRPO training and is not intended for clinical decision-making or to make medical capability claims. Its primary use is for research, representing a historical policy state.
Loading preview...
HYU-NLP-EVAL/qwen3-4b-rar-medicine-onlinerubrics-seed11-step-009 Overview
This model is an intermediate policy checkpoint from a dynamic OnlineRubrics-Every GRPO training process, specifically OnlineRubrics RaR-Medicine: step 9, seed 11. It is based on the Qwen/Qwen3-4B-Instruct-2507 model, with its 'thinking' capabilities disabled. This particular checkpoint represents a historical policy state used during a Phase-1 audit.
Key Characteristics
- Research-focused: Developed exclusively for research purposes, particularly in the context of dynamic OnlineRubrics-Every GRPO training.
- Intermediate Policy: Represents a specific stage in a reinforcement learning training process.
- Base Model: Derived from Qwen/Qwen3-4B-Instruct-2507, a 4 billion parameter instruction-tuned model.
- No Clinical Claims: Explicitly states no downstream medical capability or safety claims; not validated for clinical decision-making.
Good for
- Auditing and Analysis: Ideal for researchers studying the evolution of policies within the OnlineRubrics-Every GRPO training framework.
- Understanding Training Dynamics: Useful for investigating intermediate states of dynamic reinforcement learning processes.
- Non-Clinical Research: Suitable for academic or research applications where medical claims or clinical validation are not required.