audit-icu-gpt2-131_6M

This repo contains the model weights for audit-icu-gpt2-131_6M, a tabular language model built on the gpt2 architecture for evaluating the cross-entropy of Epic EHR audit log event sequences. This model was originally designed to calculate cross-entropies but can also be used for generation.

The code to train and perform inference this model is available here. More details about how to use this model can be found there.

Model Details

More details can be found in the model card of our paper in Appendix B here.

Please cite our paper if you use this model in your work:

@misc{warner2023autoregressive,
      title={Autoregressive Language Models For Estimating the Entropy of Epic EHR Audit Logs},
      author={Benjamin C. Warner and Thomas Kannampallil and Seunghwan Kim},
      year={2023},
      eprint={2311.06401},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}