Appendix C - AlignBot

C.1. Task Details

Task	Description
put the apple in fridge	Put the apples in the fridge. If there is a bowl on the table, need to put the apples in the bowl before placing them in the fridge.
boil water	Fill the pot with water and boil it. Turn off the stove after completing the task.
cut carrot	Cut carrots on cutting board.
heat the pot	Heat a pot on the stove. Turn off the stove after completing the task.
heat the potato	Microwave potatoes. If there is a bowl on the table, need to put the potatoes in the bowl before placing them in the Microwave.
put the apple in bowl	Put the apples on the table into the bowl.
put all the fruits in the bowl	Put the fruits on the table into the bowl.
put pear in the Drawer	Put the pears into the white drawer.
Saute carrot slices	Cut carrots on cutting board. Then fry the carrot slices in the pan.
Organize items	Place items on the counter at designated locations.
put things in drawer	Place items on the counter at designated drawer.
give me something to drink	According to the user's preferences, get the user's favorite drinks.
give me some snacks	According to the user's preferences, get the user's favorite snacks.
give me my cup	According to the user's preferences, get the user's favorite cups.
give me salt	Give the user the container with salt.
give me sugar	Give the user the container with sugar.
give me vinegar	Give the user the bottle with vinegar.
give me soy sauce	Give the user the bottle with soy sauce.
Organize the table	Put items on the table onto the shelves.
put things in box	Put items on the floor into designated boxes.

C.2. Human Rater Evaluation Criteria

Points	Rating Guidelines
0	Cues are so irrelevant to the scene that it has no corrective effect at all.
1	Although there are some errors, cues contain partially correct item status reminders, personalized reminders, or planning step reminders.
2	There are no errors, but cues only contain partially correct item status reminders, personalized reminders or planning step reminders.
3	Cues accurately include item status reminders, personalized reminders and planning step reminders.

C.3. Detailed Experimental Results

Table 1

Method	put apple in fridge	boil water	cut carrot	heat pot	heat potato	put apple in bowl	put fruits in bowl	put pear in Drawer	Saute carrot slices
GPT4o only	0%	13.3%	36.7%	46.7%	3.3%	40%	56.7%	0%	6.7%
GPT4o with memory base	33.3%	26.7%	40%	16.7%	13.3%	36.7%	46.7%	6.7%	13.3%
LLaMA2-7b with GPT4o	13.3%	13.3%	3.3%	3.3%	10%	40%	6.7%	20%	20%
AlignBot	90%	83.3%	96.7%	96.7%	80%	93.3%	73.3%	93.3%	56.7%

Table 2

Method	Organize items	put things in drawer	give something to drink	give snacks	give cup	give sugar	give vinegar	Organize table	put things in box
GPT4o only	0%	0%	30%	33.3%	30%	3.3%	20%	0%	0%
GPT4o with memory base	10%	3.3%	40%	36.7%	43.4%	80%	83.3%	40%	26.7%
LLaMA2-7b with GPT4o	3.3%	6.7%	56.7%	40%	96.7%	96.7%	93.3%	43.4%	80%
AlignBot	70%	76.7%	96.7%	90%	96.7%	96.7%	96.7%	86.7%	93.3%