oof-baroomf/csrsef-thinking-20260325T081327Z-it01-pubmedqa

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Mar 25, 2026Architecture:Transformer Warm

The oof-baroomf/csrsef-thinking-20260325T081327Z-it01-pubmedqa is a 4 billion parameter language model created by oof-baroomf, merged using the Task Arithmetic method with a Qwen3-4B-Instruct-2507 base. This model integrates components from Qwen/Qwen3-4B-Thinking-2507 and a PubMedQA-specific instruction-tuned model, suggesting an optimization for reasoning tasks, potentially within the biomedical domain. It is designed for applications requiring focused reasoning capabilities, leveraging its 32768 token context length.

Loading preview...

Model Overview

The oof-baroomf/csrsef-thinking-20260325T081327Z-it01-pubmedqa is a 4 billion parameter language model developed by oof-baroomf. It was created using the Task Arithmetic merge method, leveraging mergekit to combine distinct model capabilities. The base model for this merge was Qwen/Qwen3-4B-Instruct-2507.

Key Capabilities

This model is a result of merging two primary components:

  • Qwen/Qwen3-4B-Thinking-2507: This component likely contributes to general reasoning and instruction-following capabilities.
  • /workspace/csrsef/runs/20260325T081327Z/iteration_01/pubmedqa/instruct_merged: This indicates a strong specialization towards the PubMedQA dataset, suggesting enhanced performance on question-answering tasks within the biomedical literature domain.

Use Cases

Given its specialized merge, this model is particularly well-suited for:

  • Biomedical Question Answering: Excelling in tasks related to medical research papers and clinical queries, as indicated by its PubMedQA-specific component.
  • Reasoning Tasks: Benefiting from the "Thinking" component, it can be applied to problems requiring logical inference.
  • Instruction Following: The Instruct base model ensures robust adherence to given instructions.

Developers should consider this model for applications where accurate and context-aware responses within the biomedical field are critical, especially for question-answering and information extraction from scientific texts.

Finetunes
1 models