Hello, I have noticed that our scores were incorrectly presented. The score for Embed-RL-2B should be 0.6680, and the score for Embed-RL-4B should be 0.6805.

#116
No description provided.
TIGER-Lab org

Hi @ZoengHouNaam , thank you for your submissions.

We have renamed MMLongBench-page and ViDoSeek-page to MMLongBench-page-fixed and ViDoSeek-page-fixed, so please kindly check that the two datasets are named correctly in your submitted json file.

Please let us know if the scores are still incorrect after fixing. We can find another way to fix that.

ZoengHouNaam changed pull request status to closed

Sign up or log in to comment