Hello Tony,
The required columns for the subtask1 are not only "usage_id", "word" and "sense_id", but also "period" and "example" (used for sanity checks).
The example has all columns filled only because the example baseline simply modifies its input file. The columns orth, gloss, indices_target_token and date are not used for the evaluation of the subtask1.
The only required columns for the subtask2 are "usage_id", "word" and "gloss". (It is enough to have one gloss per word sense, no matter how many examples were there in the test data).
As for the deadline, unfortunately, the AXOLOTL'24 timeline is pretty tight, and we cannot extend the deadline any more.
However, after April 9, the Codalab competitions will still technically
remain open for submissions, so you will be able to evaluate your system.
You also can submit a paper describing your system to the LChange'24
submission deadline May 10), but you will have to specify that you
participated in the post-evaluation phase only.
Thanks for your interest!
With kind regards,
Maria.