Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Discrepancy in rm_scores Length for Mistral-7B-SPPO Iter1 Dataset #27

Open
srzhu97 opened this issue Oct 9, 2024 · 1 comment
Open

Comments

@srzhu97
Copy link

srzhu97 commented Oct 9, 2024

I noticed data-mistral-7b-instruct-sppo-iter1 column rm_scores is a list length of 7 where as there are only 5 generated responses. data-mistral-7b-instruct-sppo-iter2 and data-mistral-7b-instruct-sppo-iter3 looks correct and both have length = 5

@MeckyWu
Copy link
Collaborator

MeckyWu commented Oct 9, 2024

We included the original two responses in the UltraFeedback dataset and the 5 generated responses in the first iteration's dataset.

This is not done in the released Github repo though. Our official code only generates responses at each iteration. To reproduce the results it is recommended to run the pipeline in the official code directly. @angelahzyuan

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants