chrislimbe/pubmedqa-recursive-llm-degradation-qwen2.5-0.5b

PubMedQA Recursive LLM Degradation — Qwen2.5-3B

This repository contains synthetic biomedical question-answering data and model predictions generated as part of a study of recursive fine-tuning and model degradation.

Base Model

Qwen/Qwen2.5-3B

Source Dataset

The experiments use the PubMedQA dataset: qiaoxin/PubMedQA

This repository contains generated/derived research artifacts and does not redistribute the original PubMedQA dataset in its entirety.

Experimental Conditions

• —Recursive
• —Human Control

Generations

• —G0
• —G1
• —G2
• —G3

Random Seeds

• —42
• —123

Repository Structure

Repository Structure

seeds/
├── seed_42/
│   ├── human_control/
│   │   ├── synthetic/
│   │   └── predictions/
│   └── recursive/
│       ├── synthetic/
│       └── predictions/
│
└── seed_123/
    ├── human_control/
    │   ├── synthetic/
    │   └── predictions/
    └── recursive/
        ├── synthetic/
        └── predictions/

Contents

Synthetic Data

Generated biomedical question-answering data from each experimental condition and generation.

Predictions

Model predictions produced during evaluation for each generation and experimental condition.

Intended Use

These artifacts are provided for research, analysis, and reproducibility.

Citation

If you use these artifacts in your research, please cite the associated research paper when available.