Hello!
Just to make sure, we (datateam_tug) kept encountering some errors with Code Submission or docker submisssion, but we managed to submit our solution vie Run Uploads. Is this enough for a submission, or should we still somehow provide Docker / Github repo? Also, because we tried for a long time to solve all these submission problems, out submission is a little bit late after deadline, is it alright?
Best regards,
Anastasia
Dear Anastasia,
In which task do you participate?
For some code submissions are required, for others run uploads are fine.
An individualextension is no problem.
Best regards,
Maik
Dear Anastasia,
I found it, you participate in the multi-author task, or?
For this task we would ideadlly like to collect code submissions.
Could you please invite me to your github repository, I could make the submission from the repository.
(It is no problem when it takes a bit longer, you could in the meantime already work on the notebook paper)
Thanks in advance and best regards,
Maik
Dear Maik,
Yes, Multiauthor task, that’s right. I sent an Invite to you to our repository (it’s natriistorm/PAN26_submissoin). Thank you!
Best regards,
Anastasia
Dear Maik,
Do we need to wait for the leaderboard results to start writing the paper or can we proceed with the valid set results? If the answer to this question is yes, can you please tell me where I could find the leaderboard results? I am currently unable to do that.
Regards,
Shubhankar
Dear Shubhankar, Dear Anastasia,
please already start to write the notebook paper.
Ideally, you can focus on describing your approach and your idea.
I unblinded the evaluation results of the submissions that worked, so you should see the scores on the test set now. (we could still get the last submissions working, that is fine, the scores are currently only visible to you)
@AnastasiaAnastasia, I currently upload your submission, and I saw that you already have a valid submission to the test set now, so it likely worked now for you? (so I duplicate the submission? I still upload it, we could delete it later)
Thanks for your efforts!
Best regards,
Maik
Dear Maik,
Thank you, yes, my teammate also tried to upload the submission and it worked shortly after I wrote message here, so it’s duplicate and you can delete one of the submissions.
Best regards,
Anastasia
Oki doki,
thanks for the confirmation!
Best regards,
Maik
Dear Mike,
Regarding the test set scores, we have tried to view them but found that we are unable to see the specific scores as well—it is possible that the system only allows certain accounts or submission channels to access the score display. If you are also unable to view them, we may need to further check the permission settings of the evaluation system.
Thank you again for your continued coordination and support. Please keep us informed of any updates regarding the scores or the evaluation process.
Sincerely,
YDONG
The scores should be visible here. I just checked, I can see them also without login.
Or is an submission missing that I forgot to publish?
Best regards,
Maik
Ah, I confused the tasks, I was thinking we talk about the multi-author task here.
For GenAI Authorship verification, the results are still blinded.
The submissions are currently running on the Eloquent dataset, I think they should finish today, then we can publish and unblind the results, this will likely be today.
Best regards,
Maik
Dear Maik,
Thanks for clarifying. We understand now that the GenAI Authorship Verification results are still blinded, and the submissions are running on Eloquent today.
Will the submission of the paper/report be delayed as a result?
Please keep us posted once the results are unblinded.
Best,
YDONG
Hi, the paper deadline is not under my control, I can not make a comment on if it will be extended. (I would assume that it will be not extended)
Please focus your notebook paper on the description of your idea and approach, you could already work on this.
Best regards,
Maik
There are no requirements, the main focus of the notebook paper should be on describing the submitted approaches and/or the ideas behind them.
You do not need to include an evaluation, but you can, and when you want to do that, you could decide on which datasets you want to report what. I personally would not report results on the smoke-test dataset, as this is only to ensure that a submission produces valid outputs, but all other datasets should allow for meaningful interpretation.
Best regards,
Maik
Dear Maik,
we want for completeness of the results run an additional submission: for our current submission the approach on easy is different and we are curious if we’ve followed the approach from medium and hard, what would be the score on easy dataset. However now the resources of the system are very low and none of our new submissions are evaluated. I wonder, is it possible to increase the available resources for our check just for this evening for example? If no, it’s fine, we will then just report the existing submissions.
Kind regards,
Anastasia
Dear Anastasia,
Sorry for the delay in my response!
I added more powerful workers for execution again, you should be able to use A100 executors again.
(please also note that it is totally fine when the execution finishes after the deadline, we can still include it.)
Best regards,
Maik
Dear Maik,
Thank you very much! They already finished, can you please review the submussions and make scores for them visible? The submissions are for systems stone-valley and shy-snippet.
Kind regards,
Anastasia
Cool, I unblinded them 
Best regards,
Maik
