error during evaluation

[simplelifetime]

Here, the task requires setting alarm at 9.45 PM, but the evaluation result tells me no alarm sets for 9.45AM. It seems that the model has correctly set the 9.45PM alarm on saturday and sunday. It that an error for evaluation? or are there anything that i've missed?

[simplelifetime]

same problem on this case:

[qykong]

Hi, both tasks are user-agent interaction tasks where the instruction is wrong by default (9.45 pm is unlikely to be a wake-up alarm). And the user-agent instruction was set to the following

MobileWorld/src/mobile_world/tasks/definitions/native/set_alarm_ask_user_3.py

Line 23 in 169b493

self.relevant_information = "I just gave the incorrect information by mistake, the alarm should be set for 9:45 AM on weekend."