SeongryongJung/Qwen3-8B-Physics-GRPO-TR

TEXT GENERATIONConcurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 3, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

SeongryongJung/Qwen3-8B-Physics-GRPO-TR is an 8 billion parameter Qwen3-based language model specifically fine-tuned for physics-related tasks. Utilizing the GRPO (Generalized Reinforcement Learning from Policy Optimization) method, this model demonstrates specialized performance on the SciKnowEval physics dataset, achieving a validation mean@16 score of 72.97%. It is optimized for accurate responses in physics domains, making it suitable for scientific question answering and research assistance.

Loading preview...

Model Overview

SeongryongJung/Qwen3-8B-Physics-GRPO-TR is an 8 billion parameter language model built upon the Qwen3 architecture, specifically fine-tuned for physics-related tasks. This model leverages the GRPO (Generalized Reinforcement Learning from Policy Optimization) training method to enhance its performance in scientific domains.

Key Capabilities

  • Physics Specialization: Fine-tuned on the SciKnowEval physics dataset, demonstrating strong performance in this specific scientific area.
  • GRPO Training: Utilizes the GRPO method with a batch size of 32, optimizing for improved accuracy in physics problem-solving.
  • Performance Metrics: Achieved a peak validation mean@16 score of 72.97% on the SciKnowEval physics test set after 100 training steps.
  • Context Length: Supports a maximum prompt length of 2048 tokens and a maximum response length of 8192 tokens, with a maximum model length of 10240 tokens.

Good For

  • Scientific Question Answering: Excels at answering questions and providing information within the field of physics.
  • Physics Research Assistance: Can be used to aid researchers and students in understanding complex physics concepts and solving problems.
  • Specialized Applications: Ideal for applications requiring high accuracy and domain-specific knowledge in physics, where general-purpose models might fall short.