Will the SFT dataset be released?

[ycchen-tw]

Hi, thank you for sharing this great project!

I'm interested in your work "SU-01: Achieving Gold-Medal-Level Olympiad Reasoning via Simple and Unified Scaling." After reading the repo, I would like to know if there's any plan to release the SFT (Supervised Fine-Tuning) dataset you used or created for this project?

Open sourcing such a dataset would greatly benefit the research community. Thank you for considering this request!

[yafuly]

Hi,

Thank you for your interest in our SU-01 project and for your kind words.

Regarding the SFT dataset, due to company policy, we are not able to open-source the data in the short term.

That said, we have described the data construction pipeline in detail in the paper, including the key design principles and filtering procedures. We hope these details can still be helpful for reproducing or adapting the methodology in your own research :)