siacus/Llama-32-3B-100-climate-fasrc

LlaMig: Exploring Migration Research with Large Language Models

Overview

The ability of migration scholars to synthesize knowledge is increasingly hindered by the rapid expansion of the field — both in volume and interdisciplinarity.

To address this challenge, we develop LlaMig (Large Language Model for Migration Research), an open-source, locally deployable framework designed to transform large volumes of scholarly text into a structured, queryable database. LlaMig (this model) is a fine-tuned version of the meta-llama/Llama-3.2-3B-Instruct model.

The companion github repository contains:

• Scripts for fine-tune Llama-3.2-3B
• Exploratory analysis of LlaMig’s classifications of published migration articles
• Tools for mapping trends and themes in migration research

Model Details

• Developed by: Stefano M. Iacus, Haodong, Qi
• Finetuned from model: meta-llama/Llama-3.2-3B-Instruct
• Code Repository: https://github.com/HaodongQi/deep_literature_review_migration
• Paper: Deep literature reviews: an application of fine-tuned language models to migration research

Uses

Scholars in migration studies and climate induced migration can use this model to classify literature on the topic.

Bias, Risks, and Limitations

Potential ahallucinations, always check the results.

Citation [optional]

If you use this model, please cite:

Iacus, S.M., Qi, H., and Han, J. (2025). Deep literature reviews: an application of fine-tuned language models to migration research. ArXiv. DOI: 10.48550/arXiv.2504.13685

BibTeX

@misc{LlaMig2025,
      title={Deep literature reviews: an application of fine-tuned language models to migration research}, 
      author={Stefano M. Iacus and Haodong Qi and Jiyoung Han},
      year={2025},
      eprint={2504.13685},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2504.13685}, 
}