princeton-nlp / MQuAKE

[EMNLP 2023] MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions
https://arxiv.org/abs/2305.14795
MIT License
99 stars 7 forks source link

Request for source code #8

Open msumitaml opened 9 months ago

msumitaml commented 9 months ago

How do you train the MQuake-CF Dataset on ROME, MEMIT, please provide the source code. Then, how do you evaluate?

AOZMH commented 9 months ago

Same question, especially for the evaluation code for the edit/instance-wise accuracy. Thanks for the authors in advance!

a3616001 commented 9 months ago

Hi @msumitaml and @AOZMH , our code base is based on the MEMIT code with little modification. I will work on releasing it soon when I get more bandwidth.