EasyEdit
  • ✨Overview
  • 🔧Installation
  • PACKAGE
    • Editor
    • Hparams
    • Models
      • IKE
      • KN
      • MEMIT
      • SERAC
      • ROME
      • MEND
    • Evaluate
    • Trainer
Powered by GitBook
On this page
  • Model Editing
  • Colab Tutorial

Overview

NextInstallation

Last updated 1 year ago

EasyEdit is a Python package built upon for edit Large Language Models (LLM) like GPT-J, Llama, GPT-NEO, GPT2, T5(support models from 1B to 65B), altering the behavior of LLMs efficiently without negatively impacting performance across other inputs.

@misc{yao2023editing,
      title={Editing Large Language Models: Problems, Methods, and Opportunities}, 
      author={Yunzhi Yao and Peng Wang and Bozhong Tian and Siyuan Cheng and Zhoubo Li and Shumin Deng and Huajun Chen and Ningyu Zhang},
      year={2023},
      eprint={2305.13172},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

Model Editing

Deployed models may still make unpredictable errors. For example, Large Language Models (LLMs) notoriously hallucinate, perpetuate bias, and factually decay, so we should be able to adjust specific behaviors of pre-trained models.

Model editing aims to adjust an initial base model's (fθf_{\theta}fθ​) behavior on the particular edit descriptor [xe,ye][x_e, y_e][xe​,ye​] efficiently, such as(The president of USA: Donald Trump -> Joe Biden):

  • xex_exe​: Who is the president of the US

  • yey_eye​: Joe Biden

without influencing the model behavior on unrelated samples. The ultimate goal is to create an edited model (fθ′f_{\theta}'fθ′​).

Colab Tutorial

You can go colab directly

If you find EasyEdit useful, please kindly cite this .

✨
paper
colab link
PyTorch
License: MIT
Static Badge
An Easy-to-use Knowledge Editing Framework for Large Language Models.