Hey!
Thanks for the comprehensive bench :D
While running EnterpriseOps-Gym, we found the following issues:
task_20251126_055301_905_8e9e30d7_6aab00b8 is missing the list_team_apps tooltask_20251128_120055_076_8e9e30d7_e5dcfa22 is missing the remove_channel_member tooltask_20251222_113047_572_64a4ebb9_cd2083a8 points to the Drive instead the ITSMtask_20251222_120551_813_1628b966_8e21a82f points to Email instead of ITSM
The first two will log a warning, but this makes them hard to solve / unsolvable; while the latter two are unsolvable due to the grading function being wrong.
Also: Is it possible to get the HF dataset updated with the changes in data/revised? We ran into #8 ourselves, which is solved in this repo, but not in the HF dataset, which we use for the ground truth.
Hey!
Thanks for the comprehensive bench :D
While running EnterpriseOps-Gym, we found the following issues:
task_20251126_055301_905_8e9e30d7_6aab00b8is missing thelist_team_appstooltask_20251128_120055_076_8e9e30d7_e5dcfa22is missing theremove_channel_membertooltask_20251222_113047_572_64a4ebb9_cd2083a8points to the Drive instead the ITSMtask_20251222_120551_813_1628b966_8e21a82fpoints to Email instead of ITSMThe first two will log a warning, but this makes them hard to solve / unsolvable; while the latter two are unsolvable due to the grading function being wrong.
Also: Is it possible to get the HF dataset updated with the changes in
data/revised? We ran into #8 ourselves, which is solved in this repo, but not in the HF dataset, which we use for the ground truth.