> Need better submission failure feedback

I have submitted 13 times to the English task. 11 failed. 2 succeeded. The most common issue is
AssertionError: 499 / 498

I cannot replicate this issue with task_evaluate.py. But I believe it's an issue with my llm output having characters that break the csv reader.

I'd like to have a better submission failure feedback mechanism. The current setup is not only unhelpful for locating the output bug, but take a very long time to run.

Thanks.

Posted by: nirvanatear @ Nov. 7, 2024, 4:04 a.m.

Please, correct your headers:

Id;Text;Question;Answer

must be

ID;Text;Question;Answer

Posted by: lli-uam @ Nov. 7, 2024, 1:33 p.m.

Thank you. You are right that this is a header issue and not a output issue.

Posted by: nirvanatear @ Nov. 7, 2024, 4:46 p.m.
Post in this thread