Datasets:

Barrybw

/

ustc_test

like 0

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'filing_cabinet', 'tablelamp', 'tea', 'trashcan', 'drawing', 'orchid', 'floor', 'character', 'desk', 'bookshelf', 'computer', 'toaster', 'mat', 'stovefan', 'tvstand', 'window', 'mousepad', 'couch', 'wall', 'nightstand', 'wall_clock', 'cup', 'bed', 'food_donut', 'curtain', 'towel_rack', 'sink', 'napkin', 'bathtub'] Task:Go to sleep response: { "thoughts": Enter bedroom, turn on the light, and shut door. Change into night clothes. Turn off light. Turn down comforter, climb into bed, lay head on pillow, and cover up with comforter. Close eyes and drift off to sleep., "actions": ['[WALK] <bedroom>', '[WALK] <bed>', '[FIND] <bed>', '[LIE] <bed>', '[FIND] <sheets>', '[PULL] <sheets>'] }

Scene_Objects:['spectacles', 'sheets', 'filing_cabinet', 'food_salt', 'toilet', 'band-aids', 'walllamp', 'wallshelf', 'cpuscreen', 'hanger', 'floor', 'phone', 'computer', 'mop_bucket', 'stovefan', 'tvstand', 'deck_of_cards', 'closetdrawer', 'newspaper', 'coffe_maker', 'doorjamb', 'dresser', 'cupboard', 'door', 'food_steak', 'bag', 'sink', 'table', 'dishwasher', 'bathroom_cabinet'] Task:Put umbrella away response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'food_ice_cream', 'microwave', 'walllamp', 'bathroom', 'cat', 'trashcan', 'knifeblock', 'hanger', 'floor', 'faucet', 'shower', 'clothes_hat', 'wall', 'nightstand', 'tape', 'cup', 'food_food', 'water', 'tray', 'diary', 'pencil', 'kitchen_counter', 'headset', 'hairbrush', 'cupboard', 'hair', 'sink', 'instrument_guitar', 'soap'] Task:Drink response: { "thoughts": Walk to kitchen, grab a glass, walk to sink, turn on tap and pour water into glass, drink from glass., "actions": ['[WALK] <dining_room>', '[WALK] <cup>', '[FIND] <cup>', '[GRAB] <cup>', '[WALK] <sink>', '[WALK] <water>', '[FIND] <water>', '[GRAB] <water>', '[POUR] <water> <cup>', '[DRINK] <cup>'] }

Scene_Objects:['tablelamp', 'microwave', 'chair', 'walllamp', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'stovefan', 'ceilinglamp', 'window', 'oil', 'mousepad', 'couch', 'closetdrawer', 'food_jam', 'pillow', 'towel', 'dining_room', 'bowl', 'wall', 'nightstand', 'pasta', 'clothes_gloves', 'food_food', 'television', 'tray', 'hairbrush', 'clothes_dress'] Task:Read book response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'microwave', 'walllamp', 'detergent', 'cat', 'drawing', 'orchid', 'wallshelf', 'hanger', 'bathroom_counter', 'floor', 'desk', 'bookshelf', 'mat', 'tvstand', 'ceilinglamp', 'window', 'couch', 'dustpan', 'pillow', 'wall', 'wall_clock', 'television', 'tray', 'curtain', 'towel_rack', 'kitchen_counter', 'cupboard', 'table', 'light'] Task:Pet cat response: { "thoughts": I see my cat on the couch so I walk over, sit down and pet the cat., "actions": ['[WALK] <home_office>', '[WALK] <cat>', '[FIND] <cat>', '[TURNTO] <cat>', '[LOOKAT] <cat>', '[FIND] <couch>', '[SIT] <couch>', '[TOUCH] <cat>'] }

Scene_Objects:['food_ice_cream', 'tea', 'freezer', 'walllamp', 'orchid', 'cpuscreen', 'wallshelf', 'hanger', 'floor', 'ceiling', 'folder', 'laptop', 'tvstand', 'ceilinglamp', 'faucet', 'window', 'closetdrawer', 'towel', 'wall', 'box', 'video_game_controller', 'television', 'curtain', 'doorjamb', 'vacuum_cleaner', 'dresser', 'door', 'clothes_scarf', 'light', 'creditcard'] Task:Sweep hallway please response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['beer', 'microwave', 'dog', 'chair', 'cat', 'knifeblock', 'hanger', 'bathroom_counter', 'floor', 'phone', 'character', 'keyboard', 'clothes_jacket', 'ceilinglamp', 'shoe-shine_kit', 'tvstand', 'closetdrawer', 'oven', 'food_cheese', 'novel', 'toothbrush_holder', 'television', 'food_donut', 'console', 'controller', 'clothes_scarf', 'light', 'colander', 'board_game', 'creditcard'] Task:Pet cat response: { "thoughts": I walk to my cat, and I pet him, "actions": ['[WALK] <cat>', '[TOUCH] <cat>'] }

Scene_Objects:['sheets', 'pot', 'filing_cabinet', 'measuring_cup', 'microwave', 'band-aids', 'drawing', 'floor', 'ceiling', 'desk', 'bookshelf', 'envelope', 'button', 'toaster', 'mat', 'clothes_hat', 'wall', 'powersocket', 'oven', 'toy', 'novel', 'toothbrush_holder', 'after_shave', 'doorjamb', 'pencil', 'door', 'light', 'table', 'dishwasher', 'colander'] Task:Put toys away response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['blow_dryer', 'trashcan', 'hanger', 'floor', 'ceiling', 'bedroom', 'bookmark', 'toaster', 'stovefan', 'tvstand', 'food_jam', 'pillow', 'dining_room', 'bowl', 'wall', 'powersocket', 'clothes_gloves', 'novel', 'coffe_maker', 'coffee_filter', 'tray', 'after_shave', 'curtain', 'cd', 'chessboard', 'cupboard', 'sink', 'dishwasher', 'scrabble', 'coffee'] Task:Make coffee response: { "thoughts": Enter kitchen, find coffee maker, find coffee filter, pick up coffee filter, put it in coffee maker, find coffee, pour coffee into filter, turn on., "actions": ['[WALK] <dining_room>', '[WALK] <coffe_maker>', '[FIND] <coffe_maker>', '[FIND] <coffee_filter>', '[GRAB] <coffee_filter>', '[OPEN] <coffe_maker>', '[PUTBACK] <coffee_filter> <coffe_maker>', '[WALK] <coffee>', '[FIND] <coffee>', '[GRAB] <coffee>', '[POUR] <coffee> <coffe_maker>', '[SWITCHON] <coffe_maker>'] }

Scene_Objects:['filing_cabinet', 'measuring_cup', 'microwave', 'milk', 'knifeblock', 'bills', 'wallshelf', 'bathroom_counter', 'ceiling', 'phone', 'bedroom', 'keyboard', 'stovefan', 'laptop', 'shoe-shine_kit', 'shower', 'candle', 'clothes_hat', 'towel', 'dining_room', 'wall_clock', 'powersocket', 'newspaper', 'doorjamb', 'shaving_cream', 'vacuum_cleaner', 'dishrack', 'dishwasher', 'instrument_violin', 'creditcard'] Task:Vacuum response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'spectacles', 'ceilingfan', 'drawing', 'floor', 'ceiling', 'bedroom', 'remote_control', 'table_cloth', 'ceilinglamp', 'tvstand', 'balanceball', 'window', 'couch', 'form', 'pillow', 'wall', 'oven', 'box', 'newspaper', 'coffe_maker', 'television', 'oven_mitts', 'curtain', 'doorjamb', 'pencil', 'cupboard', 'maindoor', 'table', 'dishwasher'] Task:Change TV channel response: { "thoughts": Walk to living room. Walk to and sit at the sofa. Grab remote control and switch on TV., "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[SIT] <couch>', '[FIND] <remote_control>', '[GRAB] <remote_control>', '[FIND] <television>', '[SWITCHON] <television>'] }

Scene_Objects:['food_bread', 'beer', 'toilet', 'chair', 'freezer', 'food_butter', 'food_carrot', 'milk', 'drawing', 'floor', 'ceiling', 'toaster', 'mat', 'shower', 'mousepad', 'garbage_can', 'food_dessert', 'wall', 'nightstand', 'powersocket', 'oven', 'food_vegetable', 'food_cheese', 'photoframe', 'food_food', 'food_onion', 'doorjamb', 'laundry_detergent', 'door', 'light'] Task:Rotate stock in refrigerator response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'walllamp', 'trashcan', 'milk', 'knifeblock', 'orchid', 'kitchen_cabinet', 'floor', 'ceiling', 'character', 'desk', 'keyboard', 'toaster', 'stovefan', 'faucet', 'mousepad', 'couch', 'wall', 'pasta', 'oven', 'box', 'food_food', 'water', 'tray', 'pencil', 'kitchen_counter', 'sauce_pan', 'cupboard', 'instrument_guitar', 'bathroom_cabinet'] Task:Cook some food response: { "thoughts": Cook some pasta on the stove, "actions": ['[WALK] <dining_room>', '[WALK] <kitchen_cabinet>', '[FIND] <kitchen_cabinet>', '[OPEN] <kitchen_cabinet>', '[FIND] <pasta>', '[GRAB] <pasta>', '[FIND] <water>', '[GRAB] <water>', '[FIND] <sauce_pan>', '[POUR] <water> <sauce_pan>', '[FIND] <oven>', '[SWITCHON] <oven>', '[PUTBACK] <pasta> <sauce_pan>', '[FIND] <sauce_pan>', '[GRAB] <sauce_pan>', '[PUTBACK] <sauce_pan> <oven>'] }

Scene_Objects:['chair', 'ground_coffee', 'nail_polish', 'orchid', 'hanger', 'floor', 'ceiling', 'dvd_player', 'character', 'keyboard', 'food_cake', 'bookshelf', 'mail', 'toaster', 'plate', 'homework', 'pillow', 'dining_room', 'wall', 'bed', 'coffe_maker', 'coffee_filter', 'water', 'tray', 'curtain', 'kitchen_counter', 'cupboard', 'bathroom_cabinet', 'light', 'dishwasher'] Task:Make coffee response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['toilet', 'chair', 'walllamp', 'knifeblock', 'orchid', 'cpuscreen', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'alcohol', 'table_cloth', 'stovefan', 'ceilinglamp', 'mat', 'iron', 'faucet', 'mousepad', 'bowl', 'wall', 'nightstand', 'oven', 'mouse', 'cupboard', 'door', 'clothes_scarf', 'bathroom_cabinet', 'table', 'instrument_guitar'] Task:Work response: { "thoughts": For doing homework walk to home office. Walk to computer. Switch on Computer. Sit in a chair and do homework in computer., "actions": ['[WALK] <home_office>', '[WALK] <computer>', '[FIND] <computer>', '[WALK] <computer>', '[FIND] <chair>', '[SIT] <chair>', '[SWITCHON] <computer>'] }

Scene_Objects:['pot', 'juice', 'food_ice_cream', 'beer', 'food_oatmeal', 'microwave', 'toilet', 'chair', 'freezer', 'walllamp', 'ground_coffee', 'blow_dryer', 'food_carrot', 'dough', 'milk', 'drawing', 'floor', 'ceiling', 'phone', 'stovefan', 'mat', 'ceilinglamp', 'shower', 'faucet', 'bowl', 'wall', 'nightstand', 'oven', 'powersocket', 'cup', 'photoframe', 'coffe_maker', 'food_food', 'check', 'oven_mitts', 'tray', 'doorjamb', 'pencil', 'kitchen_counter', 'cupboard', 'door', 'sink', 'napkin', 'light', 'dishwasher', 'table'] Task:Cook some food response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'toilet', 'chair', 'walllamp', 'trashcan', 'cpuscreen', 'wallshelf', 'hanger', 'bathroom_counter', 'floor', 'ceiling', 'keyboard', 'tvstand', 'faucet', 'mousepad', 'closetdrawer', 'couch', 'food_jam', 'wall', 'nightstand', 'oven', 'newspaper', 'food_food', 'vacuum_cleaner', 'sink', 'bathtub', 'bathroom_cabinet', 'light', 'table', 'hands_both'] Task:Wash hands response: { "thoughts": I go to the bathroom, walk to the sink, turn on the faucet and then wash my hands under the water., "actions": ['[WALK] <bathroom>', '[WALK] <sink>', '[FIND] <faucet>', '[TURNTO] <faucet>', '[FIND] <hands_both>', '[WASH] <hands_both>'] }

Scene_Objects:['filing_cabinet', 'food_salt', 'chair', 'bathroom', 'orchid', 'cpuscreen', 'bathroom_counter', 'desk', 'keyboard', 'button', 'piano_bench', 'ceilinglamp', 'stereo', 'floor_lamp', 'oil', 'mousepad', 'closetdrawer', 'tape', 'love_seat', 'newspaper', 'bed', 'check', 'video_game_controller', 'mouthwash', 'curtain', 'lighter', 'diary', 'doorjamb', 'dresser', 'sink'] Task:Read response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'tablelamp', 'chair', 'freezer', 'cat', 'drawing', 'cpuscreen', 'floor', 'ceiling', 'phone', 'desk', 'bookshelf', 'keyboard', 'computer', 'toaster', 'stereo', 'faucet', 'picture', 'mousepad', 'couch', 'wall', 'nightstand', 'powersocket', 'cup', 'doorjamb', 'mouse', 'kitchen_counter', 'door', 'home_office', 'light'] Task:Listen to music response: { "thoughts": I go to the bedroom and sit at my desk. I turn my radio on and listen to music., "actions": ['[WALK] <bedroom>', '[WALK] <light>', '[FIND] <light>', '[SWITCHON] <light>', '[WALK] <desk>', '[FIND] <chair>', '[SIT] <chair>', '[FIND] <stereo>', '[SWITCHON] <stereo>'] }

Scene_Objects:['blender', 'milk', 'drawing', 'orchid', 'needle', 'hanger', 'bathroom_counter', 'floor', 'phone', 'computer', 'bookmark', 'toaster', 'mat', 'tvstand', 'electric_shaver', 'clothes_socks', 'closetdrawer', 'couch', 'nightstand', 'wall', 'bed', 'keys', 'photoframe', 'tray', 'centerpiece', 'dresser', 'bag', 'soap', 'clothes_shirt', 'teeth'] Task:Read book response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'cpuscreen', 'wallshelf', 'bathroom_counter', 'floor', 'stovefan', 'plate', 'shower', 'picture', 'faucet', 'dish_soap', 'pillow', 'wall', 'powersocket', 'cutting_board', 'tape', 'cup', 'coffe_maker', 'food_food', 'television', 'lighter', 'centerpiece', 'mouse', 'kitchen_counter', 'chessboard', 'dresser', 'cupboard', 'bathroom_cabinet', 'table', 'dishwasher'] Task:Wash dishes with dishwasher response: { "thoughts": I go to the kitchen and open the dishwasher. I fill it up with dirty dishes and glasses. Then I put dish soap in the dishwasher, close it and turn it on., "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[FIND] <dishwasher>', '[OPEN] <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <cup>', '[GRAB] <cup>', '[PUTBACK] <cup> <dishwasher>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[PUTBACK] <dish_soap> <dishwasher>', '[CLOSE] <dishwasher>', '[SWITCHON] <dishwasher>'] }

Scene_Objects:['food_bread', 'filing_cabinet', 'toilet', 'freezer', 'dough', 'drawing', 'hanger', 'floor', 'ceiling', 'mat', 'shoe-shine_kit', 'shower', 'clothes_hat', 'bowl', 'nightstand', 'wall', 'powersocket', 'wall_clock', 'photoframe', 'light_bulb', 'food_food', 'coffe_maker', 'razor', 'coffee_filter', 'doorjamb', 'hairbrush', 'door', 'home_office', 'light', 'soap'] Task:Put away groceries response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['dry_pasta', 'pot', 'tablelamp', 'microwave', 'chair', 'freezer', 'orchid', 'drawing', 'floor', 'ceiling', 'phone', 'toaster', 'mat', 'ceilinglamp', 'faucet', 'stovefan', 'shower', 'food_peanut_butter', 'wall', 'powersocket', 'oven', 'cup', 'coffe_maker', 'tray', 'doorjamb', 'coffee_pot', 'towel_rack', 'kitchen_counter', 'cupboard', 'door', 'sink', 'table', 'dishwasher', 'light'] Task:Wipe down counter response: { "thoughts": Clean counter, "actions": ['[WALK] <dining_room>', '[WALK] <rag>', '[FIND] <rag>', '[GRAB] <rag>', '[WALK] <sink>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[RINSE] <rag>', '[SQUEEZE] <rag>', '[SWITCHOFF] <faucet>', '[WALK] <kitchen_counter>', '[WIPE] <kitchen_counter>'] }

Scene_Objects:['food_pizza', 'tablelamp', 'walllamp', 'orchid', 'knifeblock', 'wallshelf', 'floor', 'ceiling', 'bedroom', 'remote_control', 'envelope', 'computer', 'ceilinglamp', 'tvstand', 'window', 'closetdrawer', 'couch', 'pillow', 'nightstand', 'wall', 'pasta', 'oven', 'love_seat', 'television', 'razor', 'curtain', 'dresser', 'cupboard', 'bathroom_cabinet', 'table'] Task:Watch TV response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['food_bread', 'tablelamp', 'blender', 'tea', 'chair', 'freezer', 'drawing', 'orchid', 'hanger', 'floor', 'shampoo', 'stamp', 'bedroom', 'desk', 'computer', 'mat', 'stereo', 'mousepad', 'food_jam', 'dining_room', 'wall', 'wall_clock', 'bench', 'photoframe', 'food_food', 'dresser', 'napkin', 'home_office', 'table', 'spoon'] Task:Eat dinner response: { "thoughts": In our dining room, my family eats dinner. Once the food has been prepared, we gather at the dining room table and prepare to eat., "actions": ['[WALK] <dining_room>', '[WALK] <food_food>', '[FIND] <food_food>', '[GRAB] <food_food>', '[WALK] <dining_room>', '[WALK] <table>', '[PUTBACK] <food_food> <table>', '[WALK] <dining_room>', '[WALK] <food_food>', '[FIND] <food_food>', '[GRAB] <food_food>', '[WALK] <dining_room>', '[WALK] <table>', '[PUTBACK] <food_food> <table>', '[WALK] <chair>', '[PULL] <chair>', '[SIT] <chair>', '[TURNTO] <food_food>', '[LOOKAT] <food_food>', '[TURNTO] <food_food>', '[LOOKAT] <food_food>'] }

Scene_Objects:['pot', 'filing_cabinet', 'food_salt', 'chair', 'bathroom', 'drawing', 'food_egg', 'floor', 'desk', 'bookshelf', 'computer', 'mat', 'stereo', 'shower', 'couch', 'food_peanut_butter', 'wall', 'powersocket', 'oven', 'cup', 'doorjamb', 'mouse', 'kitchen_counter', 'pajamas', 'controller', 'door', 'clothes_scarf', 'light', 'dishwasher', 'board_game'] Task:Pull up carpet response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'toilet', 'freezer', 'food_carrot', 'drawing', 'hanger', 'floor', 'ceiling', 'shampoo', 'stovefan', 'mat', 'shower', 'window', 'nightstand', 'wall', 'powersocket', 'food_cheese', 'photoframe', 'bed', 'newspaper', 'food_food', 'drinking_glass', 'water', 'curtain', 'doorjamb', 'dresser', 'food_steak', 'door', 'home_office', 'light'] Task:Drink response: { "thoughts": im feeling thirsty im going to kitchen and open the freezer where i can find some cold water and im going to drink it to end up my thirst, "actions": ['[WALK] <dining_room>', '[WALK] <freezer>', '[FIND] <freezer>', '[OPEN] <freezer>', '[FIND] <drinking_glass>', '[GRAB] <drinking_glass>', '[WALK] <water>', '[FIND] <water>', '[GRAB] <water>', '[POUR] <water> <drinking_glass>', '[DRINK] <drinking_glass>'] }

Scene_Objects:['spectacles', 'pot', 'scissors', 'microwave', 'chair', 'walllamp', 'knifeblock', 'food_egg', 'hanger', 'floor', 'ceiling', 'phone', 'keyboard', 'toaster', 'plate', 'faucet', 'clothes_socks', 'food_kiwi', 'wall', 'powersocket', 'clothes_gloves', 'coffe_maker', 'after_shave', 'mouse', 'kitchen_counter', 'dresser', 'cupboard', 'sink', 'bathroom_cabinet', 'dishwasher'] Task:Do dishes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'food_salt', 'microwave', 'ceilingfan', 'orchid', 'bathroom_counter', 'bookshelf', 'purse', 'stovefan', 'mat', 'iron', 'electric_shaver', 'food_noodles', 'mousepad', 'couch', 'clothes_pants', 'pillow', 'powersocket', 'keys', 'coffe_maker', 'oven_mitts', 'wine', 'shaving_cream', 'pencil', 'kitchen_counter', 'vacuum_cleaner', 'dresser', 'home_office', 'light', 'tooth_paste'] Task:Hang up car keys response: { "thoughts": Open the front door. Enter the entrance hall. Take out car keys from purse. Hang car keys up on hook on the wall., "actions": ['[WALK] <home_office>', '[WALK] <purse>', '[FIND] <purse>', '[OPEN] <purse>', '[FIND] <keys>', '[GRAB] <keys>', '[CLOSE] <purse>'] }

Scene_Objects:['filing_cabinet', 'microwave', 'jelly', 'chair', 'detergent', 'food_butter', 'knifeblock', 'wallshelf', 'kitchen_cabinet', 'floor', 'ceiling', 'button', 'toaster', 'mat', 'plate', 'faucet', 'stereo', 'pillow', 'wall', 'powersocket', 'cup', 'food_food', 'diary', 'pencil', 'kitchen_counter', 'cupboard', 'door', 'sink', 'bathroom_cabinet', 'light'] Task:Get glass of water response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'chair', 'walllamp', 'orchid', 'cpuscreen', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'table_cloth', 'stovefan', 'ceilinglamp', 'mat', 'toaster', 'deck_of_cards', 'window', 'tvstand', 'mousepad', 'pillow', 'wall', 'pasta', 'light_bulb', 'check', 'mouthwash', 'wooden_spoon', 'mouse', 'door', 'table', 'coffee'] Task:Browse internet response: { "thoughts": Turn on computer, pull up browser, browse the internet, "actions": ['[WALK] <home_office>', '[WALK] <computer>', '[FIND] <computer>', '[SWITCHON] <computer>', '[TURNTO] <computer>', '[LOOKAT] <computer>'] }

Scene_Objects:['fork', 'chair', 'freezer', 'knife', 'drawing', 'needle', 'wallshelf', 'floor', 'phone', 'ceiling', 'dvd_player', 'food_bacon', 'ceilinglamp', 'laptop', 'plate', 'shower', 'tvstand', 'tooth_paste', 'wall', 'tape', 'cup', 'novel', 'food_food', 'television', 'coffee_filter', 'hair', 'cupboard', 'table', 'board_game', 'clothes_shirt'] Task:Eat response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'mop', 'toilet', 'ceilingfan', 'drawing', 'orchid', 'bathroom_counter', 'floor', 'ceiling', 'character', 'video_game_console', 'food_cake', 'button', 'ceilinglamp', 'tvstand', 'balanceball', 'stereo', 'window', 'couch', 'wall', 'powersocket', 'tape', 'cup', 'video_game_controller', 'television', 'mouthwash', 'curtain', 'hairdryer', 'table', 'coffee'] Task:Play games response: { "thoughts": Turn on the television. Turn on the video game system. Insert desired game into gaming system. Follow directions on game to and push needed buttons on video game remote controller., "actions": ['[WALK] <home_office>', '[WALK] <television>', '[FIND] <television>', '[SWITCHON] <television>', '[FIND] <video_game_console>', '[SWITCHON] <video_game_console>', '[FIND] <video_game_controller>', '[GRAB] <video_game_controller>', '[FIND] <button>', '[PUSH] <button>', '[TURNTO] <television>', '[WATCH] <television>'] }

Scene_Objects:['pot', 'food_bread', 'beer', 'toilet', 'freezer', 'walllamp', 'drawing', 'knifeblock', 'floor', 'ceiling', 'desk', 'toaster', 'mat', 'stovefan', 'shower', 'faucet', 'food_orange', 'nightstand', 'wall', 'powersocket', 'oven', 'photoframe', 'coffe_maker', 'tray', 'doorjamb', 'kitchen_counter', 'cupboard', 'door', 'sink', 'light'] Task:Make toast response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'microwave', 'dog', 'bathroom', 'drawing', 'orchid', 'wallshelf', 'floor', 'ceiling', 'desk', 'bookshelf', 'keyboard', 'computer', 'toaster', 'mat', 'faucet', 'shower', 'couch', 'wall', 'powersocket', 'newspaper', 'keys', 'wooden_spoon', 'doorjamb', 'centerpiece', 'kitchen_counter', 'door', 'home_office', 'light', 'shoe_rack'] Task:Turn on light response: { "thoughts": Switch on light in home office, "actions": ['[WALK] <home_office>', '[WALK] <light>', '[FIND] <light>', '[PLUGIN] <light>', '[SWITCHON] <light>', '[WALK] <home_office>'] }

Scene_Objects:['pot', 'juice', 'microwave', 'freezer', 'blow_dryer', 'ground_coffee', 'cat', 'floor', 'legs_both', 'face', 'toaster', 'laptop', 'homework', 'closetdrawer', 'towel', 'feet_both', 'arms_both', 'basket_for_clothes', 'newspaper', 'photoframe', 'coffe_maker', 'mouthwash', 'centerpiece', 'doorjamb', 'headset', 'dresser', 'cupboard', 'napkin', 'soap', 'clothes_shirt'] Task:Try yourself off response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'freezer', 'ground_coffee', 'detergent', 'drawing', 'bathroom_counter', 'floor', 'ceiling', 'character', 'bedroom', 'stovefan', 'toaster', 'faucet', 'shower', 'closetdrawer', 'candle', 'dining_room', 'wall', 'oven', 'photoframe', 'bed', 'coffe_maker', 'coffee_filter', 'tray', 'doorjamb', 'kitchen_counter', 'cupboard', 'door', 'sink', 'table'] Task:Make coffee response: { "thoughts": Make coffee in the kitchen., "actions": ['[WALK] <dining_room>', '[WALK] <coffe_maker>', '[FIND] <coffe_maker>', '[PLUGIN] <coffe_maker>', '[OPEN] <coffe_maker>', '[FIND] <coffee_filter>', '[GRAB] <coffee_filter>', '[PUTBACK] <coffee_filter> <coffe_maker>', '[FIND] <ground_coffee>', '[GRAB] <ground_coffee>', '[PUTBACK] <ground_coffee> <coffe_maker>', '[CLOSE] <coffe_maker>', '[SWITCHON] <coffe_maker>'] }

Scene_Objects:['filing_cabinet', 'tea', 'measuring_cup', 'band-aids', 'knifeblock', 'cpuscreen', 'dvd_player', 'folder', 'desk', 'bookmark', 'stereo', 'faucet', 'dustpan', 'microphone', 'oven', 'cup', 'bed', 'photoframe', 'check', 'coffe_maker', 'television', 'food_rice', 'pencil', 'shredder', 'controller', 'home_office', 'light', 'table', 'scrabble', 'electrical_outlet'] Task:Shred receipts response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['fork', 'filing_cabinet', 'chair', 'drawing', 'bills', 'knife', 'floor', 'ceilinglamp', 'stereo', 'window', 'couch', 'food_jam', 'form', 'towel', 'dining_room', 'wall', 'basket_for_clothes', 'photoframe', 'novel', 'food_food', 'check', 'television', 'oven_mitts', 'razor', 'tray', 'pencil', 'dresser', 'pajamas', 'table', 'soap'] Task:Read book response: { "thoughts": Sit and read, "actions": ['[WALK] <home_office>', '[WALK] <filing_cabinet>', '[OPEN] <filing_cabinet>', '[FIND] <novel>', '[GRAB] <novel>', '[CLOSE] <filing_cabinet>', '[WALK] <couch>', '[SIT] <couch>', '[READ] <novel>'] }

Scene_Objects:['tablelamp', 'chair', 'jelly', 'walllamp', 'bathroom', 'knife', 'wallshelf', 'floor', 'ceiling', 'bookmark', 'ceilinglamp', 'mat', 'window', 'mousepad', 'couch', 'pillow', 'wall', 'nightstand', 'food_food', 'television', 'toothbrush_holder', 'wooden_spoon', 'kettle', 'shoes', 'cupboard', 'conditioner', 'clothes_dress', 'door', 'napkin', 'light'] Task:Read book response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'chair', 'orchid', 'drawing', 'cpuscreen', 'floor', 'ceiling', 'phone', 'desk', 'keyboard', 'food_cake', 'thread', 'computer', 'laptop', 'mat', 'ceilinglamp', 'tvstand', 'mousepad', 'wall', 'nightstand', 'oven', 'pasta', 'television', 'tray', 'doorjamb', 'mouse', 'headset', 'chessboard', 'cupboard', 'door'] Task:Watch movie response: { "thoughts": Sit in my deck chair with my laptop, find a good movie on You Tube to watch, plug in earphones and listen to movie quietly so I don't disturb my upstairs neighbor., "actions": ['[WALK] <home_office>', '[WALK] <laptop>', '[FIND] <laptop>', '[GRAB] <laptop>', '[FIND] <chair>', '[SIT] <chair>', '[SWITCHON] <laptop>', '[FIND] <keyboard>', '[TYPE] <keyboard>', '[FIND] <headset>', '[GRAB] <headset>', '[PUTON] <headset>', '[TURNTO] <laptop>', '[LOOKAT] <laptop>'] }

Scene_Objects:['fork', 'filing_cabinet', 'scissors', 'juice', 'food_salt', 'microwave', 'chair', 'food_butter', 'toothbrush', 'drawing', 'wallshelf', 'bathroom_counter', 'floor', 'button', 'brush', 'ceilinglamp', 'plate', 'wall', 'tape', 'food_food', 'check', 'food_donut', 'wooden_spoon', 'rag', 'door', 'clothes_scarf', 'napkin', 'table', 'light', 'dishwasher'] Task:Eat response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'mop', 'microwave', 'freezer', 'walllamp', 'detergent', 'toothbrush', 'orchid', 'drawing', 'clothes_underwear', 'floor', 'thread', 'mat', 'closetdrawer', 'wall', 'nightstand', 'toy', 'standingmirror', 'light_bulb', 'novel', 'check', 'tray', 'cards', 'mouse', 'pencil', 'vacuum_cleaner', 'headset', 'maindoor', 'napkin', 'light'] Task:Read book response: { "thoughts": Walk to the Living Room. Look at the Cabinet. Find a Textbook. Grab the textbook. Open the Textbook. Read the textbook., "actions": ['[WALK] <home_office>', '[WALK] <filing_cabinet>', '[FIND] <filing_cabinet>', '[TURNTO] <filing_cabinet>', '[LOOKAT] <filing_cabinet>', '[FIND] <novel>', '[GRAB] <novel>', '[READ] <novel>'] }

Scene_Objects:['tablelamp', 'crayon', 'chair', 'freezer', 'walllamp', 'bathroom', 'woman', 'orchid', 'drawing', 'floor', 'ceiling', 'phone', 'computer', 'ceilinglamp', 'tvstand', 'window', 'closetdrawer', 'couch', 'pillow', 'nightstand', 'wall', 'child', 'powersocket', 'check', 'coffe_maker', 'television', 'curtain', 'pencil', 'dresser', 'table'] Task:Watch movie response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'pot', 'juice', 'cd_player', 'toilet', 'freezer', 'trashcan', 'bedroom', 'remote_control', 'button', 'clothes_jacket', 'toaster', 'laptop', 'ceilinglamp', 'plate', 'homework', 'dining_room', 'broom', 'bench', 'tape', 'keys', 'coffe_maker', 'television', 'towel_rack', 'hair', 'cupboard', 'maindoor', 'sink', 'clothes_skirt', 'vase'] Task:Listen to music response: { "thoughts": I walk to my living room. Find CD player. Plugin the CD player to Power. Also find the remote control. Push the Botton, "actions": ['[WALK] <home_office>', '[WALK] <cd_player>', '[FIND] <cd_player>', '[PLUGIN] <cd_player>', '[FIND] <remote_control>', '[GRAB] <remote_control>', '[FIND] <button>', '[PUSH] <button>'] }

Scene_Objects:['food_bread', 'toilet', 'walllamp', 'bills', 'drawing', 'cpuscreen', 'bathroom_counter', 'ceiling', 'cleaning_solution', 'mop_bucket', 'table_cloth', 'faucet', 'shower', 'iron', 'clothes_socks', 'pillow', 'wall', 'food_food', 'mirror', 'wine', 'doorjamb', 'coffee_pot', 'paper_towel', 'cupboard', 'conditioner', 'door', 'sink', 'bathroom_cabinet', 'dishwasher', 'scrabble'] Task:Clean mirror response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['food_ice_cream', 'microwave', 'toilet', 'chair', 'bathroom_counter', 'floor', 'laptop', 'faucet', 'shower', 'towel', 'wall', 'wall_clock', 'nightstand', 'cup', 'bed', 'photoframe', 'water', 'tray', 'doorjamb', 'towel_rack', 'dresser', 'door', 'sink', 'bathtub', 'bathroom_cabinet', 'light', 'table', 'soap', 'scrabble', 'hands_both'] Task:Wash hands response: { "thoughts": Go into the bathroom and go over to the sink. Grab soap and turn on the water. Wet your hands and rub them together so that the soap lathers. Rub the soap all over your hands thoroughly. Rinse off the soap with water and turn off the sink. Dry your hands with a towel., "actions": ['[WALK] <bathroom>', '[WALK] <sink>', '[FIND] <soap>', '[GRAB] <soap>', '[FIND] <water>', '[GRAB] <water>', '[FIND] <hands_both>', '[POUR] <water> <hands_both>', '[FIND] <hands_both>', '[SCRUB] <hands_both>', '[PUTOBJBACK] <soap>', '[RINSE] <hands_both>', '[FIND] <towel>', '[GRAB] <towel>', '[WIPE] <hands_both>'] }

Scene_Objects:['pot', 'tablelamp', 'chair', 'band-aids', 'walllamp', 'detergent', 'orchid', 'drawing', 'floor', 'ceiling', 'phone', 'video_game_console', 'computer', 'ceilinglamp', 'tvstand', 'mat', 'plate', 'window', 'closetdrawer', 'couch', 'pillow', 'nightstand', 'wall', 'photoframe', 'video_game_controller', 'television', 'curtain', 'dresser', 'table', 'dishwasher'] Task:Play games response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['fork', 'water_glass', 'pot', 'tablelamp', 'microwave', 'drawing', 'knife', 'food_egg', 'bathroom_counter', 'floor', 'ceiling', 'desk', 'toaster', 'stovefan', 'plate', 'shower', 'dish_soap', 'mousepad', 'wall', 'oven', 'photoframe', 'coffe_maker', 'tray', 'towel_rack', 'kitchen_counter', 'cupboard', 'spoon', 'coffee_cup', 'dishwasher', 'light'] Task:Wash dishes with dishwasher response: { "thoughts": I go to the kitchen and put plates, cutlery and glasses into the dishwasher, put the soap, close the machine and switch it on., "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[FIND] <dishwasher>', '[OPEN] <dishwasher>', '[FIND] <coffee_cup>', '[GRAB] <coffee_cup>', '[PUTBACK] <coffee_cup> <dishwasher>', '[FIND] <coffee_cup>', '[GRAB] <coffee_cup>', '[PUTBACK] <coffee_cup> <dishwasher>', '[FIND] <pot>', '[GRAB] <pot>', '[PUTBACK] <pot> <dishwasher>', '[FIND] <pot>', '[GRAB] <pot>', '[PUTBACK] <pot> <dishwasher>', '[FIND] <fork>', '[GRAB] <fork>', '[PUTBACK] <fork> <dishwasher>', '[FIND] <fork>', '[GRAB] <fork>', '[PUTBACK] <fork> <dishwasher>', '[FIND] <knife>', '[GRAB] <knife>', '[PUTBACK] <knife> <dishwasher>', '[FIND] <knife>', '[GRAB] <knife>', '[PUTBACK] <knife> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <spoon>', '[GRAB] <spoon>', '[PUTBACK] <spoon> <dishwasher>', '[FIND] <spoon>', '[GRAB] <spoon>', '[PUTBACK] <spoon> <dishwasher>', '[FIND] <water_glass>', '[GRAB] <water_glass>', '[PUTBACK] <water_glass> <dishwasher>', '[FIND] <water_glass>', '[GRAB] <water_glass>', '[PUTBACK] <water_glass> <dishwasher>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[PUTBACK] <dish_soap> <dishwasher>', '[CLOSE] <dishwasher>', '[SWITCHON] <dishwasher>'] }

Scene_Objects:['spectacles', 'tablelamp', 'microwave', 'chair', 'wallshelf', 'hanger', 'floor', 'phone', 'food_bacon', 'piano_bench', 'stovefan', 'ceilinglamp', 'shower', 'mousepad', 'closetdrawer', 'powersocket', 'television', 'coin', 'food_donut', 'ice', 'rag', 'pencil', 'shoes', 'dresser', 'clothes_dress', 'clothes_scarf', 'sink', 'bathroom_cabinet', 'light', 'home_office'] Task:Hang up jacket response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'fork', 'pot', 'sheets', 'food_salt', 'detergent', 'food_butter', 'drawing', 'cpuscreen', 'wallshelf', 'bathroom_counter', 'floor', 'shampoo', 'desk', 'food_cake', 'mat', 'faucet', 'food_kiwi', 'wall', 'powersocket', 'toy', 'novel', 'food_food', 'curtain', 'dresser', 'sink', 'bathroom_cabinet', 'dishwasher', 'soap', 'hands_both'] Task:Wash hands response: { "thoughts": Going to the sink and scrubbing hands, "actions": ['[WALK] <bathroom>', '[WALK] <sink>', '[FIND] <sink>', '[FIND] <hands_both>', '[RINSE] <hands_both>', '[SCRUB] <hands_both>', '[RINSE] <hands_both>'] }

Scene_Objects:['food_salt', 'toilet', 'freezer', 'ground_coffee', 'milk', 'orchid', 'ceiling', 'phone', 'character', 'keyboard', 'food_bacon', 'stovefan', 'ceilinglamp', 'tvstand', 'closetdrawer', 'couch', 'dining_room', 'bowl', 'nightstand', 'pasta', 'food_cheese', 'bed', 'novel', 'check', 'food_food', 'coin', 'tray', 'doorjamb', 'diary', 'table'] Task:Look out window response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'chair', 'drawing', 'hanger', 'floor', 'phone', 'ceilinglamp', 'plate', 'iron', 'candle', 'pillow', 'wall', 'bench', 'powersocket', 'oven', 'food_food', 'video_game_controller', 'check', 'wooden_spoon', 'after_shave', 'doorjamb', 'shoes', 'dresser', 'cupboard', 'clothes_dress', 'door', 'conditioner', 'sink', 'table', 'light'] Task:Eat dinner response: { "thoughts": Put food on the table, sit down, and eat the food., "actions": ['[WALK] <dining_room>', '[WALK] <plate>', '[FIND] <plate>', '[GRAB] <plate>', '[FIND] <food_food>', '[GRAB] <food_food>', '[PUTBACK] <food_food> <plate>', '[WALK] <dining_room>', '[WALK] <table>', '[FIND] <table>', '[PUTBACK] <plate> <table>', '[FIND] <chair>', '[SIT] <chair>', '[EAT] <plate>'] }

Scene_Objects:['food_bread', 'tablelamp', 'microwave', 'chair', 'walllamp', 'floor', 'ceiling', 'character', 'bedroom', 'stovefan', 'ceilinglamp', 'faucet', 'oil', 'closetdrawer', 'couch', 'food_peanut_butter', 'pillow', 'dining_room', 'wall', 'nightstand', 'cutting_board', 'photoframe', 'television', 'mouse', 'pencil', 'hairbrush', 'pajamas', 'bathroom_cabinet', 'light', 'teeth'] Task:Read book response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'ceilingfan', 'drawing', 'food_egg', 'floor', 'ceiling', 'bookshelf', 'remote_control', 'button', 'ceilinglamp', 'tvstand', 'balanceball', 'window', 'couch', 'food_jam', 'pillow', 'dining_room', 'wall', 'nightstand', 'photoframe', 'television', 'food_rice', 'curtain', 'kitchen_counter', 'hair', 'home_office', 'table', 'light', 'dishwasher', 'dirt'] Task:Change TV channel response: { "thoughts": I go to the living room where the TV is located. I plug in the TV to the outlet. I search for the remote control. I point the remote control directly to the TV, turn it on and push the number of the channel that I want., "actions": ['[WALK] <home_office>', '[WALK] <television>', '[FIND] <television>', '[PLUGIN] <television>', '[FIND] <remote_control>', '[TURNTO] <television>', '[POINTAT] <television>', '[SWITCHON] <television>', '[FIND] <button>', '[PUSH] <button>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'chair', 'walllamp', 'drawing', 'needle', 'cpuscreen', 'bills', 'floor', 'ceiling', 'phone', 'desk', 'bookshelf', 'keyboard', 'computer', 'mat', 'ceilinglamp', 'mousepad', 'tooth_paste', 'couch', 'pillow', 'towel', 'wall', 'nightstand', 'powersocket', 'television', 'coin', 'razor', 'doorjamb', 'centerpiece', 'mouse', 'door', 'light', 'board_game'] Task:Re arrange office response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['glue', 'crayon', 'freezer', 'nail_polish', 'food_carrot', 'milk', 'cpuscreen', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'ceilinglamp', 'tvstand', 'mousepad', 'dining_room', 'wall', 'powersocket', 'cup', 'bed', 'food_food', 'doorjamb', 'mouse', 'pajamas', 'controller', 'door', 'spoon', 'napkin', 'light', 'dirt'] Task:Get glass of milk response: { "thoughts": Get an empty glass. Take milk from refrigerator and open it. Pour milk into glass., "actions": ['[WALK] <dining_room>', '[WALK] <cup>', '[FIND] <cup>', '[GRAB] <cup>', '[WALK] <freezer>', '[OPEN] <freezer>', '[FIND] <milk>', '[GRAB] <milk>', '[POUR] <milk> <cup>', '[PUTOBJBACK] <milk>', '[CLOSE] <freezer>'] }

Scene_Objects:['dry_pasta', 'pot', 'filing_cabinet', 'food_salt', 'tea', 'chair', 'food_carrot', 'drawing', 'floor', 'ceiling', 'phone', 'dvd_player', 'bookshelf', 'stovefan', 'mat', 'mousepad', 'food_noodles', 'dining_room', 'wall', 'powersocket', 'oven', 'cup', 'check', 'rag', 'doorjamb', 'cd', 'hairbrush', 'dresser', 'door', 'light'] Task:Turn on light response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'toilet', 'chair', 'walllamp', 'trashcan', 'ceilingfan', 'drawing', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'phone', 'desk', 'keyboard', 'bookshelf', 'computer', 'laptop', 'shoe-shine_kit', 'balanceball', 'picture', 'ceilinglamp', 'mousepad', 'wall', 'nightstand', 'newspaper', 'photoframe', 'doorjamb', 'mouse', 'dresser', 'cupboard', 'door', 'light'] Task:Work response: { "thoughts": I need to finish some work on my computer in my home office to meet a deadline for tomorrow., "actions": ['[WALK] <home_office>', '[WALK] <light>', '[FIND] <light>', '[SWITCHON] <light>', '[FIND] <keyboard>', '[GRAB] <keyboard>', '[FIND] <desk>', '[OPEN] <desk>', '[PUTBACK] <keyboard> <desk>', '[FIND] <mouse>', '[GRAB] <mouse>', '[FIND] <mousepad>', '[PUTBACK] <mouse> <mousepad>', '[PLUGIN] <keyboard>', '[PLUGIN] <mouse>', '[FIND] <computer>', '[SWITCHON] <computer>', '[FIND] <chair>', '[SIT] <chair>', '[TURNTO] <computer>', '[TYPE] <keyboard>', '[GRAB] <mouse>', '[TOUCH] <mouse>', '[LOOKAT] <computer>'] }

Scene_Objects:['pot', 'filing_cabinet', 'toilet', 'knifeblock', 'wallshelf', 'cpuscreen', 'bathroom_counter', 'floor', 'ceiling', 'desk', 'bookshelf', 'mat', 'tvstand', 'faucet', 'couch', 'pillow', 'wall', 'powersocket', 'bed', 'clothes_gloves', 'novel', 'toilet_paper', 'doorjamb', 'coffee_pot', 'mouse', 'kitchen_counter', 'door', 'napkin', 'bathroom_cabinet', 'light'] Task:Turn off light response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['beer', 'tea', 'chair', 'walllamp', 'ground_coffee', 'bills', 'orchid', 'bathroom_counter', 'floor', 'phone', 'ceiling', 'mail', 'ceilinglamp', 'mat', 'iron', 'stereo', 'faucet', 'clothes_socks', 'clothes_hat', 'basket_for_clothes', 'food_rice', 'curtain', 'lighter', 'doorjamb', 'towel_rack', 'vacuum_cleaner', 'dishrack', 'cd', 'napkin', 'home_office'] Task:Turn on light response: { "thoughts": I go to bedroom and locate a lamp. This lamp can be turned on by touch, as it has capacitive sensing. So I do it., "actions": ['[WALK] <bedroom>', '[WALK] <floor_lamp>', '[FIND] <floor_lamp>', '[TOUCH] <floor_lamp>', '[SWITCHON] <floor_lamp>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'microwave', 'detergent', 'knife', 'bills', 'bathroom_counter', 'phone', 'folder', 'bedroom', 'character', 'computer', 'toaster', 'stovefan', 'mat', 'window', 'mousepad', 'pillow', 'oven', 'check', 'coffe_maker', 'razor', 'after_shave', 'mouse', 'kitchen_counter', 'shoes', 'dresser', 'pajamas', 'sink', 'bag'] Task:Get dressed response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['fork', 'tablelamp', 'food_ice_cream', 'walllamp', 'bathroom', 'ceilingfan', 'trashcan', 'woman', 'drawing', 'orchid', 'food_egg', 'bathroom_counter', 'floor', 'bedroom', 'man', 'ceilinglamp', 'window', 'couch', 'wall', 'bench', 'cup', 'newspaper', 'novel', 'check', 'television', 'towel_rack', 'clothes_dress', 'table', 'clothes_skirt', 'creditcard'] Task:Hang with friends response: { "thoughts": Sit with friends. Talk, "actions": ['[WALK] <home_office>', '[WALK] <man>', '[FIND] <man>', '[TURNTO] <man>', '[LOOKAT] <man>', '[GREET] <man>', '[FIND] <woman>', '[TURNTO] <woman>', '[LOOKAT] <woman>', '[GREET] <woman>', '[WALK] <couch>', '[SIT] <couch>', '[TURNTO] <man>', '[LOOKAT] <man>', '[GREET] <man>', '[TURNTO] <woman>', '[LOOKAT] <woman>', '[GREET] <woman>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'band-aids', 'bathroom', 'cat', 'orchid', 'wallshelf', 'hanger', 'floor', 'phone', 'character', 'bedroom', 'purse', 'computer', 'table_cloth', 'tvstand', 'powersocket', 'toy', 'clothes_gloves', 'coffe_maker', 'razor', 'food_onion', 'centerpiece', 'headset', 'chessboard', 'cupboard', 'door', 'home_office', 'table', 'laser_pointer'] Task:Open door response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'freezer', 'ceilingfan', 'drawing', 'knifeblock', 'bathroom_counter', 'floor', 'ceiling', 'phone', 'keyboard', 'bookshelf', 'faucet', 'picture', 'window', 'dustpan', 'towel', 'wall', 'nightstand', 'photoframe', 'keys', 'toothbrush_holder', 'food_donut', 'mouse', 'kitchen_counter', 'sink', 'bathroom_cabinet', 'table', 'vase', 'soap', 'hands_both'] Task:Wash hands response: { "thoughts": Cleaning hands, "actions": ['[WALK] <bathroom>', '[WALK] <faucet>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[FIND] <soap>', '[GRAB] <soap>', '[FIND] <hands_both>', '[SCRUB] <hands_both>', '[PUTOBJBACK] <soap>', '[RINSE] <hands_both>', '[SWITCHOFF] <faucet>', '[FIND] <towel>', '[GRAB] <towel>', '[WIPE] <hands_both>', '[PUTOBJBACK] <towel>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'chair', 'ground_coffee', 'woman', 'orchid', 'wallshelf', 'cpuscreen', 'hanger', 'shampoo', 'floor', 'ceiling', 'bookshelf', 'food_noodles', 'closetdrawer', 'mousepad', 'couch', 'towel', 'wall', 'food_food', 'television', 'kettle', 'mouse', 'cards', 'vacuum_cleaner', 'dresser', 'door', 'hairdryer', 'dishwasher', 'clothes_shirt'] Task:Greet guests response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['dry_pasta', 'fork', 'filing_cabinet', 'chair', 'cat', 'drawing', 'hanger', 'floor', 'character', 'tea_bag', 'mat', 'picture', 'electric_shaver', 'couch', 'food_jam', 'towel', 'dining_room', 'wall', 'oven', 'coffe_maker', 'tray', 'curtain', 'doorjamb', 'diary', 'cards', 'cupboard', 'door', 'bathroom_cabinet', 'light', 'tooth_paste'] Task:Read book response: { "thoughts": pick up the book open the book and read, "actions": ['[WALK] <bedroom>', '[WALK] <novel>', '[GRAB] <novel>', '[READ] <novel>'] }

Scene_Objects:['tablelamp', 'chair', 'walllamp', 'orchid', 'floor', 'ceiling', 'character', 'computer', 'ceilinglamp', 'tvstand', 'shower', 'faucet', 'window', 'deck_of_cards', 'closetdrawer', 'couch', 'dustpan', 'pillow', 'food_kiwi', 'nightstand', 'wall', 'cup', 'coffe_maker', 'television', 'food_rice', 'curtain', 'dresser', 'napkin', 'table', 'light'] Task:Watch TV response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'walllamp', 'bathroom', 'drawing', 'cpuscreen', 'wallshelf', 'floor', 'ceiling', 'character', 'tea_bag', 'computer', 'table_cloth', 'toaster', 'mat', 'tvstand', 'stovefan', 'electric_shaver', 'couch', 'pillow', 'wall', 'basket_for_clothes', 'newspaper', 'coffee_table', 'photoframe', 'video_game_controller', 'television', 'doorjamb', 'dresser', 'cupboard', 'dishwasher'] Task:Read magazine response: { "thoughts": Get magazine off coffee table. Sit in arm chair. Open magazine to first page. Read until finished., "actions": ['[WALK] <home_office>', '[WALK] <coffee_table>', '[FIND] <newspaper>', '[GRAB] <newspaper>', '[FIND] <couch>', '[SIT] <couch>', '[READ] <newspaper>', '[PUTBACK] <newspaper> <coffee_table>'] }

Scene_Objects:['pot', 'filing_cabinet', 'measuring_cup', 'freezer', 'ground_coffee', 'drawing', 'knifeblock', 'clothes_underwear', 'hanger', 'floor', 'desk', 'stamp', 'bookshelf', 'food_cake', 'table_cloth', 'bookmark', 'mat', 'clothes_hat', 'wall', 'pasta', 'powersocket', 'toy', 'box', 'food_cheese', 'light_bulb', 'doorjamb', 'lighter', 'kitchen_counter', 'door', 'light'] Task:Clean response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'tablelamp', 'microwave', 'toilet', 'chair', 'drawing', 'orchid', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'mat', 'ceilinglamp', 'tvstand', 'mousepad', 'wall', 'nightstand', 'box', 'novel', 'toilet_paper', 'television', 'mouse', 'dresser', 'door', 'food_steak', 'clothes_skirt', 'tooth_paste'] Task:Read book response: { "thoughts": "walk to entrance" "walk to living room" "look for book" "find book" "pick up book" "look for chair" "find chair" "sit in chair" "read book", "actions": ['[WALK] <home_office>', '[WALK] <home_office>', '[WALK] <novel>', '[FIND] <novel>', '[TURNTO] <novel>', '[LOOKAT] <novel>', '[FIND] <novel>', '[FIND] <chair>', '[TURNTO] <chair>', '[LOOKAT] <chair>', '[SIT] <chair>', '[GRAB] <novel>', '[READ] <novel>'] }

Scene_Objects:['sheets', 'sponge', 'crayon', 'microwave', 'walllamp', 'bathroom', 'drawing', 'clothes_underwear', 'floor', 'bookshelf', 'button', 'stovefan', 'mat', 'stereo', 'mousepad', 'clothes_hat', 'bowl', 'nightstand', 'wall', 'bed', 'photoframe', 'light_bulb', 'food_food', 'television', 'kitchen_counter', 'dresser', 'door', 'sink', 'clothes_skirt', 'scrabble'] Task:Breakfast response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'pot', 'chair', 'walllamp', 'orchid', 'cpuscreen', 'floor', 'ceiling', 'shampoo', 'desk', 'keyboard', 'computer', 'table_cloth', 'stovefan', 'ceilinglamp', 'mat', 'plate', 'stereo', 'mousepad', 'wall', 'bench', 'television', 'wine', 'rag', 'lighter', 'shaving_cream', 'mouse', 'hairbrush', 'door', 'table'] Task:Write an email response: { "thoughts": walk to living room, look at computer, find chair, sit in chair, grab keyboard, grab mouse, type using keyboard and mouse, "actions": ['[WALK] <home_office>', '[WALK] <computer>', '[FIND] <computer>', '[TURNTO] <computer>', '[LOOKAT] <computer>', '[WALK] <computer>', '[FIND] <chair>', '[SIT] <chair>', '[FIND] <keyboard>', '[GRAB] <keyboard>', '[FIND] <mouse>', '[GRAB] <mouse>', '[TYPE] <keyboard>'] }

Scene_Objects:['food_bread', 'filing_cabinet', 'tablelamp', 'food_ice_cream', 'beer', 'crayon', 'chair', 'knifeblock', 'cpuscreen', 'bathroom_counter', 'phone', 'desk', 'bedroom', 'bookshelf', 'thread', 'toaster', 'ceilinglamp', 'mousepad', 'closetdrawer', 'pillow', 'wall', 'cup', 'bed', 'keys', 'food_food', 'television', 'food_onion', 'doorjamb', 'mouse', 'home_office'] Task:Lock door response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'tablelamp', 'walllamp', 'bathroom', 'cat', 'ceilingfan', 'knifeblock', 'phone', 'character', 'brush', 'toaster', 'mat', 'faucet', 'ceilinglamp', 'shower', 'window', 'couch', 'wall_clock', 'cup', 'newspaper', 'food_onion', 'lighter', 'towel_rack', 'mouse', 'kitchen_counter', 'headset', 'cupboard', 'table', 'vase', 'laser_pointer'] Task:Pet cat response: { "thoughts": Every time this animal is seen, you must approach it and show compassion for it by petting it in a gentle manner., "actions": ['[WALK] <home_office>', '[WALK] <bedroom>', '[WALK] <bedroom>', '[WALK] <home_office>', '[WALK] <home_office>', '[WALK] <cat>', '[FIND] <cat>', '[FIND] <brush>', '[GRAB] <brush>'] }

Scene_Objects:['sponge', 'microwave', 'chair', 'jelly', 'knifeblock', 'floor', 'ceiling', 'character', 'alcohol', 'toaster', 'plate', 'faucet', 'drying_rack', 'window', 'tvstand', 'dish_soap', 'closetdrawer', 'couch', 'towel', 'pillow', 'food_kiwi', 'wall', 'toy', 'novel', 'curtain', 'kitchen_counter', 'cupboard', 'sink', 'instrument_guitar', 'shoe_rack'] Task:Do dishes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'crayon', 'detergent', 'ceilingfan', 'food_carrot', 'milk', 'drawing', 'knifeblock', 'food_egg', 'floor', 'phone', 'desk', 'bookshelf', 'envelope', 'ceilinglamp', 'window', 'mousepad', 'couch', 'wall', 'food_cheese', 'novel', 'television', 'food_rice', 'after_shave', 'towel_rack', 'kitchen_counter', 'comb', 'table', 'dishwasher', 'creditcard'] Task:Relax on sofa response: { "thoughts": I walk to the living room and find the sofa. I sit on the sofa and get comfortable., "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[FIND] <couch>', '[SIT] <couch>'] }

Scene_Objects:['sheets', 'scissors', 'chair', 'dog', 'ground_coffee', 'knife', 'needle', 'bills', 'bathroom_counter', 'floor', 'phone', 'bookshelf', 'bookmark', 'ceilinglamp', 'faucet', 'shower', 'picture', 'food_noodles', 'clothes_socks', 'mousepad', 'pillow', 'wall', 'photoframe', 'food_food', 'tray', 'kitchen_counter', 'sink', 'bag', 'home_office', 'table'] Task:Sit response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['freezer', 'bathroom', 'detergent', 'orchid', 'cpuscreen', 'floor', 'phone', 'desk', 'bedroom', 'purse', 'computer', 'table_cloth', 'bookmark', 'stovefan', 'stereo', 'oil', 'closetdrawer', 'coffe_maker', 'television', 'food_donut', 'food_rice', 'tray', 'centerpiece', 'doorjamb', 'pencil', 'maindoor', 'sink', 'home_office', 'table', 'clothes_skirt'] Task:Read book response: { "thoughts": Walk to living room. Find book. Grab book. Open book. Read book., "actions": ['[WALK] <home_office>', '[WALK] <novel>', '[FIND] <novel>', '[GRAB] <novel>', '[READ] <novel>'] }

Scene_Objects:['tablelamp', 'scissors', 'food_carrot', 'hanger', 'phone', 'desk', 'keyboard', 'computer', 'stovefan', 'deck_of_cards', 'faucet', 'window', 'closetdrawer', 'microphone', 'pillow', 'box', 'bed', 'photoframe', 'check', 'food_onion', 'diary', 'pencil', 'shoes', 'dishrack', 'dresser', 'laundry_detergent', 'spoon', 'soap', 'tooth_paste', 'shoe_rack'] Task:Take off shoes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['glue', 'filing_cabinet', 'chair', 'walllamp', 'orchid', 'cpuscreen', 'hanger', 'bathroom_counter', 'floor', 'ceiling', 'desk', 'keyboard', 'bookshelf', 'computer', 'ceilinglamp', 'mat', 'shoe-shine_kit', 'mousepad', 'pillow', 'wall', 'powersocket', 'cup', 'bed', 'wooden_spoon', 'coffee_pot', 'mouse', 'kitchen_counter', 'door', 'home_office', 'table'] Task:Write an email response: { "thoughts": I walk to the Home Office then find the computer and sit in the chair. I switch on the computer and start typing the email. I switch the computer off., "actions": ['[WALK] <home_office>', '[WALK] <computer>', '[FIND] <computer>', '[FIND] <chair>', '[SIT] <chair>', '[SWITCHON] <computer>', '[TURNTO] <computer>', '[LOOKAT] <computer>', '[FIND] <keyboard>', '[TYPE] <keyboard>', '[SWITCHOFF] <computer>'] }

Scene_Objects:['pot', 'filing_cabinet', 'food_pizza', 'tablelamp', 'beer', 'freezer', 'blow_dryer', 'drawing', 'needle', 'cpuscreen', 'floor', 'desk', 'bookshelf', 'mat', 'mousepad', 'form', 'wall', 'powersocket', 'food_cheese', 'food_food', 'check', 'doorjamb', 'mouse', 'chessboard', 'dresser', 'laundry_detergent', 'door', 'hairdryer', 'light', 'board_game'] Task:Put clothes away response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['dry_pasta', 'chair', 'freezer', 'walllamp', 'food_butter', 'bills', 'cpuscreen', 'bathroom_counter', 'phone', 'ceiling', 'character', 'bedroom', 'computer', 'stovefan', 'ceilinglamp', 'faucet', 'floor_lamp', 'window', 'mousepad', 'closetdrawer', 'couch', 'oven', 'cup', 'photoframe', 'television', 'kitchen_counter', 'dresser', 'bathroom_cabinet', 'table', 'light'] Task:Turn on light response: { "thoughts": Walk over to light switch, flip the switch to the on position, "actions": ['[WALK] <home_office>', '[WALK] <floor_lamp>', '[SWITCHON] <floor_lamp>'] }

Scene_Objects:['pot', 'food_pizza', 'ground_coffee', 'walllamp', 'food_carrot', 'knifeblock', 'floor', 'ceiling', 'character', 'computer', 'piano_bench', 'stovefan', 'toaster', 'tvstand', 'faucet', 'oil', 'pillow', 'wall', 'oven', 'powersocket', 'food_food', 'water', 'tray', 'after_shave', 'doorjamb', 'kitchen_counter', 'dresser', 'cupboard', 'door', 'dishwasher'] Task:Make coffee response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'chair', 'knifeblock', 'knife', 'orchid', 'floor', 'ceiling', 'phone', 'keyboard', 'mail', 'stovefan', 'ceilinglamp', 'deck_of_cards', 'stereo', 'pillow', 'dining_room', 'wall', 'cup', 'bed', 'novel', 'coffe_maker', 'doorjamb', 'kitchen_counter', 'cupboard', 'controller', 'laundry_detergent', 'hairdryer', 'sink', 'home_office', 'dishwasher'] Task:Make coffee response: { "thoughts": Making coffee, "actions": ['[WALK] <dining_room>', '[WALK] <coffe_maker>', '[FIND] <coffe_maker>', '[SWITCHON] <coffe_maker>'] }

Scene_Objects:['spectacles', 'tablelamp', 'beer', 'measuring_cup', 'band-aids', 'woman', 'knifeblock', 'cpuscreen', 'bathroom_counter', 'phone', 'ceiling', 'character', 'desk', 'man', 'thread', 'clothes_jacket', 'tvstand', 'stereo', 'faucet', 'window', 'shower', 'couch', 'form', 'cutting_board', 'novel', 'television', 'bathroom_cabinet', 'table', 'dishwasher', 'instrument_violin'] Task:Greet guests response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'scissors', 'tablelamp', 'crayon', 'microwave', 'freezer', 'bathroom', 'trashcan', 'ceilingfan', 'knifeblock', 'clothes_underwear', 'longboard', 'shampoo', 'dvd_player', 'folder', 'mail', 'bookmark', 'toaster', 'shower', 'homework', 'couch', 'bed', 'clothes_gloves', 'towel_rack', 'mouse', 'pencil', 'cupboard', 'maindoor', 'table', 'laser_pointer'] Task:Pet cat response: { "thoughts": I walk into the living room and I pet the cat, "actions": ['[WALK] <home_office>', '[WALK] <cat>', '[TOUCH] <cat>'] }

Scene_Objects:['sheets', 'food_pizza', 'tablelamp', 'chair', 'walllamp', 'orchid', 'bathroom_counter', 'floor', 'ceiling', 'man', 'food_bacon', 'stovefan', 'ceilinglamp', 'brush', 'mat', 'oil', 'couch', 'food_jam', 'pillow', 'wall', 'nightstand', 'newspaper', 'novel', 'coffe_maker', 'television', 'mouse', 'chessboard', 'soap', 'instrument_violin', 'teeth'] Task:Have conversation with boyfriend response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'tablelamp', 'blender', 'microwave', 'toilet', 'dog', 'freezer', 'drawing', 'knifeblock', 'longboard', 'floor', 'folder', 'computer', 'ceilinglamp', 'window', 'couch', 'wall', 'nightstand', 'bench', 'cup', 'bed', 'clothes_gloves', 'novel', 'television', 'curtain', 'towel_rack', 'cupboard', 'sink', 'home_office', 'table'] Task:Relax on sofa response: { "thoughts": I'm tired and want to relax on the sofa, "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[FIND] <couch>', '[SIT] <couch>'] }

Scene_Objects:['food_pizza', 'microwave', 'toilet', 'jelly', 'freezer', 'band-aids', 'drawing', 'knifeblock', 'clothes_underwear', 'kitchen_cabinet', 'floor', 'ceiling', 'toaster', 'mat', 'faucet', 'shower', 'wall', 'nightstand', 'powersocket', 'cup', 'basket_for_clothes', 'photoframe', 'food_food', 'water', 'doorjamb', 'kitchen_counter', 'cupboard', 'door', 'sink', 'light'] Task:Get drink response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['blender', 'walllamp', 'drawing', 'orchid', 'bathroom_counter', 'floor', 'ceiling', 'keyboard', 'remote_control', 'thread', 'bookshelf', 'toaster', 'ceilinglamp', 'tvstand', 'mat', 'clothes_socks', 'wall', 'pasta', 'powersocket', 'cutting_board', 'cup', 'clothes_gloves', 'novel', 'television', 'rag', 'doorjamb', 'shoes', 'door', 'table', 'dishwasher'] Task:Watch TV response: { "thoughts": turn on the tv, "actions": ['[WALK] <home_office>', '[WALK] <television>', '[FIND] <television>', '[SWITCHON] <television>', '[FIND] <remote_control>', '[TOUCH] <remote_control>'] }

Scene_Objects:['sheets', 'sponge', 'mop', 'chair', 'drawing', 'cpuscreen', 'bathroom_counter', 'floor', 'desk', 'keyboard', 'computer', 'toaster', 'mat', 'window', 'mousepad', 'clothes_socks', 'couch', 'food_peanut_butter', 'pillow', 'wall', 'pasta', 'nightstand', 'light_bulb', 'coffe_maker', 'television', 'wooden_spoon', 'mouse', 'shoes', 'vacuum_cleaner', 'sink'] Task:Check email response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'walllamp', 'drawing', 'bathroom_counter', 'floor', 'ceiling', 'bookshelf', 'stovefan', 'laptop', 'faucet', 'mousepad', 'candle', 'microphone', 'food_orange', 'wall', 'photoframe', 'check', 'mouthwash', 'after_shave', 'doorjamb', 'cupboard', 'pajamas', 'conditioner', 'door', 'sink', 'bathroom_cabinet', 'light', 'soap', 'clothes_shirt', 'hands_both'] Task:Wash hands response: { "thoughts": I go into the bathroom and turn on the light. I walk to the sink and turn on the water. I grab the soap and wash my hands, "actions": ['[WALK] <bathroom>', '[WALK] <light>', '[FIND] <light>', '[SWITCHON] <light>', '[WALK] <sink>', '[FIND] <soap>', '[GRAB] <soap>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[FIND] <hands_both>', '[WASH] <hands_both>', '[RINSE] <hands_both>', '[PUTOBJBACK] <soap>', '[SWITCHOFF] <faucet>'] }

Scene_Objects:['microwave', 'toilet', 'jelly', 'walllamp', 'detergent', 'knifeblock', 'orchid', 'floor', 'ceiling', 'toaster', 'plate', 'faucet', 'dish_soap', 'wall', 'nightstand', 'powersocket', 'oven', 'food_food', 'television', 'coffee_filter', 'wine', 'curtain', 'rag', 'doorjamb', 'shaving_cream', 'kitchen_counter', 'dishrack', 'cupboard', 'sink', 'bag'] Task:Wash dishes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'filing_cabinet', 'food_pizza', 'tablelamp', 'blender', 'food_salt', 'blow_dryer', 'trashcan', 'toothbrush', 'cpuscreen', 'hanger', 'ceiling', 'character', 'desk', 'keyboard', 'bookshelf', 'purse', 'ceilinglamp', 'tvstand', 'faucet', 'mousepad', 'microphone', 'oven', 'food_food', 'check', 'tray', 'dresser', 'napkin', 'bathroom_cabinet', 'clothes_shirt'] Task:Leave home response: { "thoughts": Put on shoes. Put on coat. Open door and exit through it. Lock door behind you., "actions": ['[WALK] <home_office>', '[WALK] <shoes>', '[FIND] <shoes>', '[GRAB] <shoes>', '[PUTON] <shoes>'] }

Scene_Objects:['tablelamp', 'jelly', 'freezer', 'detergent', 'woman', 'orchid', 'drawing', 'phone', 'character', 'man', 'bookshelf', 'toaster', 'ceilinglamp', 'tvstand', 'stereo', 'food_orange', 'nightstand', 'box', 'food_cheese', 'coffe_maker', 'food_food', 'check', 'video_game_controller', 'centerpiece', 'mouse', 'kitchen_counter', 'vacuum_cleaner', 'sink', 'table', 'tooth_paste'] Task:Greet people response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['fork', 'chair', 'walllamp', 'knife', 'food_egg', 'floor', 'food_cake', 'toaster', 'plate', 'tvstand', 'window', 'dish_soap', 'wall', 'nightstand', 'oven', 'cutting_board', 'coffe_maker', 'check', 'razor', 'television', 'food_donut', 'tray', 'centerpiece', 'doorjamb', 'dishrack', 'food_steak', 'spoon', 'light', 'dishwasher', 'colander'] Task:Wash dishes with dishwasher response: { "thoughts": put forks, spoons, plates, knife in dishwasher, add dish soap, turn diswaser on, "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[FIND] <dishwasher>', '[OPEN] <dishwasher>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[PUTBACK] <dish_soap> <dishwasher>', '[FIND] <fork>', '[GRAB] <fork>', '[FIND] <dishrack>', '[PUTBACK] <fork> <dishrack>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishrack>', '[FIND] <spoon>', '[GRAB] <spoon>', '[PUTBACK] <spoon> <dishrack>', '[FIND] <knife>', '[GRAB] <knife>', '[PUTBACK] <knife> <dishrack>', '[CLOSE] <dishwasher>', '[SWITCHON] <dishwasher>'] }

Scene_Objects:['blender', 'freezer', 'walllamp', 'bathroom', 'drawing', 'clothes_underwear', 'hanger', 'floor', 'ceiling', 'character', 'keyboard', 'brush', 'tvstand', 'stereo', 'window', 'closetdrawer', 'towel', 'wall', 'powersocket', 'tape', 'food_food', 'television', 'curtain', 'doorjamb', 'vacuum_cleaner', 'dresser', 'door', 'clothes_scarf', 'dishwasher', 'board_game'] Task:vacuum carpet response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'walllamp', 'bathroom', 'food_carrot', 'toothbrush', 'orchid', 'knife', 'longboard', 'character', 'bedroom', 'computer', 'faucet', 'homework', 'clothes_socks', 'food_noodles', 'dining_room', 'photoframe', 'check', 'television', 'after_shave', 'rag', 'lighter', 'towel_rack', 'kitchen_counter', 'dresser', 'clothes_dress', 'home_office', 'light', 'laser_pointer', 'scrabble'] Task:Pick up phone response: { "thoughts": I walk to the home office. I find the telephone. I pick up the telephone handset., "actions": ['[WALK] <home_office>', '[WALK] <phone>', '[FIND] <phone>', '[GRAB] <phone>'] }

Scene_Objects:['tablelamp', 'scissors', 'chair', 'bathroom', 'toothbrush', 'drawing', 'cpuscreen', 'floor', 'keyboard', 'piano_bench', 'ceilinglamp', 'oil', 'form', 'food_orange', 'wall', 'nightstand', 'oven', 'photoframe', 'check', 'food_food', 'television', 'coffee_filter', 'tray', 'mouse', 'cd', 'cupboard', 'door', 'table', 'light', 'creditcard'] Task:Eat dinner response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'pot', 'microwave', 'bathroom', 'drawing', 'wallshelf', 'floor', 'ceiling', 'bedroom', 'toaster', 'stovefan', 'deck_of_cards', 'shower', 'pillow', 'wall', 'powersocket', 'cup', 'coffe_maker', 'check', 'wine', 'food_onion', 'curtain', 'lighter', 'kitchen_counter', 'cupboard', 'hair', 'controller', 'table', 'dishwasher', 'colander'] Task:Take dishes out of dishwasher response: { "thoughts": Take clean dishes out of the dishwasher and put away, "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[OPEN] <dishwasher>', '[CLOSE] <dishwasher>'] }

Scene_Objects:['pot', 'filing_cabinet', 'measuring_cup', 'ground_coffee', 'walllamp', 'milk', 'cpuscreen', 'ceiling', 'phone', 'keyboard', 'stovefan', 'plate', 'food_peanut_butter', 'pillow', 'wall', 'nightstand', 'oven', 'coffe_maker', 'food_food', 'wine', 'tray', 'food_onion', 'mouse', 'kitchen_counter', 'headset', 'comb', 'cupboard', 'door', 'bathroom_cabinet', 'dishwasher'] Task:Get out dish response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['food_pizza', 'chair', 'freezer', 'walllamp', 'food_carrot', 'knifeblock', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'toaster', 'mat', 'faucet', 'window', 'oil', 'couch', 'wall', 'bench', 'powersocket', 'photoframe', 'bed', 'television', 'curtain', 'after_shave', 'towel_rack', 'kitchen_counter', 'bathtub', 'sink', 'light'] Task:Open bathroom window response: { "thoughts": Make sure bathroom window is unlocked. Lift up bathroom window so it is open., "actions": ['[WALK] <bathroom>', '[WALK] <window>', '[FIND] <window>', '[OPEN] <window>'] }

Scene_Objects:['sheets', 'pot', 'food_salt', 'cpuscreen', 'clothes_underwear', 'floor', 'bedroom', 'food_cake', 'computer', 'stovefan', 'ceilinglamp', 'oil', 'couch', 'powersocket', 'cutting_board', 'food_cheese', 'photoframe', 'food_food', 'razor', 'oven_mitts', 'coffee_pot', 'mouse', 'shoes', 'dishrack', 'chessboard', 'cupboard', 'controller', 'food_steak', 'hairdryer', 'spoon'] Task:Clean response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['food_bread', 'food_ice_cream', 'mop', 'microwave', 'detergent', 'ceilingfan', 'drawing', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'bedroom', 'remote_control', 'stovefan', 'ceilinglamp', 'tvstand', 'balanceball', 'window', 'couch', 'candle', 'wall', 'bed', 'television', 'curtain', 'dresser', 'maindoor', 'table', 'dishwasher', 'vase', 'coffee'] Task:Watch TV response: { "thoughts": Walk to livingroom, sit on couch, grab remote and turn on tv., "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[FIND] <couch>', '[SIT] <couch>', '[FIND] <remote_control>', '[GRAB] <remote_control>', '[FIND] <television>', '[TURNTO] <television>', '[WATCH] <television>'] }

Scene_Objects:['sheets', 'measuring_cup', 'microwave', 'jelly', 'ground_coffee', 'drawing', 'knifeblock', 'bathroom_counter', 'character', 'bookshelf', 'keyboard', 'stovefan', 'ceilinglamp', 'brush', 'faucet', 'picture', 'mat', 'closetdrawer', 'powersocket', 'oven', 'cup', 'basket_for_clothes', 'clothes_gloves', 'novel', 'television', 'kettle', 'cupboard', 'sink', 'spoon', 'table'] Task:Read book response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'pot', 'tablelamp', 'crayon', 'tea', 'mop', 'ground_coffee', 'detergent', 'drawing', 'floor', 'phone', 'ceilinglamp', 'laptop', 'shower', 'iron', 'window', 'oil', 'couch', 'pillow', 'wall', 'oven', 'cup', 'coffe_maker', 'toilet_paper', 'television', 'door', 'sink', 'table', 'scrabble', 'dirt'] Task:Read book response: { "thoughts": Bring book, lie on the sofa, read book, "actions": ['[WALK] <home_office>', '[WALK] <novel>', '[GRAB] <novel>', '[WALK] <couch>', '[LIE] <couch>', '[READ] <novel>'] }

Scene_Objects:['dry_pasta', 'food_pizza', 'blender', 'tea', 'toilet', 'chair', 'freezer', 'bathroom', 'woman', 'knifeblock', 'drawing', 'wallshelf', 'floor', 'phone', 'character', 'desk', 'man', 'button', 'deck_of_cards', 'mousepad', 'clothes_pants', 'child', 'coffe_maker', 'tray', 'comb', 'dresser', 'cupboard', 'hairbrush', 'soap', 'tooth_paste'] Task:Greet guests response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['spectacles', 'sheets', 'tablelamp', 'toilet', 'chair', 'walllamp', 'nail_polish', 'floor', 'ceiling', 'phone', 'keyboard', 'bookshelf', 'ceilinglamp', 'mat', 'tvstand', 'stereo', 'mousepad', 'couch', 'pillow', 'wall', 'nightstand', 'photoframe', 'novel', 'coffe_maker', 'check', 'television', 'curtain', 'lighter', 'bathroom_cabinet', 'table'] Task:Relax on sofa response: { "thoughts": walk into living room, look at the sofa which one is available, sit on the sofa, now take relax on the sofa, "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[FIND] <couch>', '[TURNTO] <couch>', '[LOOKAT] <couch>', '[SIT] <couch>'] }

Scene_Objects:['pot', 'food_bread', 'beer', 'toilet', 'freezer', 'walllamp', 'drawing', 'knifeblock', 'floor', 'ceiling', 'food_bacon', 'stovefan', 'mat', 'toaster', 'shower', 'nightstand', 'wall', 'powersocket', 'oven', 'cup', 'food_vegetable', 'photoframe', 'food_food', 'tray', 'doorjamb', 'kitchen_counter', 'cupboard', 'door', 'light', 'tooth_paste'] Task:Cook some food response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'chair', 'knifeblock', 'knife', 'orchid', 'floor', 'ceiling', 'phone', 'keyboard', 'mail', 'stovefan', 'ceilinglamp', 'deck_of_cards', 'stereo', 'pillow', 'dining_room', 'wall', 'cup', 'bed', 'novel', 'coffe_maker', 'doorjamb', 'kitchen_counter', 'cupboard', 'controller', 'laundry_detergent', 'hairdryer', 'sink', 'home_office', 'dishwasher'] Task:Make coffee response: { "thoughts": Making coffee, "actions": ['[WALK] <dining_room>', '[WALK] <coffe_maker>', '[FIND] <coffe_maker>', '[SWITCHON] <coffe_maker>'] }

Scene_Objects:['pot', 'chair', 'freezer', 'walllamp', 'food_carrot', 'orchid', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'desk', 'keyboard', 'bookshelf', 'computer', 'mat', 'ceilinglamp', 'tvstand', 'mousepad', 'closetdrawer', 'pillow', 'wall', 'television', 'oven_mitts', 'wooden_spoon', 'doorjamb', 'mouse', 'pajamas', 'door', 'light', 'table'] Task:Write an email response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'tablelamp', 'toilet', 'walllamp', 'orchid', 'floor', 'ceiling', 'phone', 'remote_control', 'computer', 'toaster', 'tvstand', 'window', 'mousepad', 'closetdrawer', 'couch', 'pillow', 'nightstand', 'wall', 'love_seat', 'clothes_gloves', 'television', 'tray', 'curtain', 'food_onion', 'dresser', 'cupboard', 'instrument_guitar', 'table', 'light'] Task:Watch TV response: { "thoughts": I'll watch TV, "actions": ['[WALK] <home_office>', '[WALK] <remote_control>', '[FIND] <remote_control>', '[GRAB] <remote_control>', '[FIND] <television>', '[SWITCHON] <television>', '[PUTOBJBACK] <remote_control>', '[FIND] <love_seat>', '[SIT] <love_seat>', '[TURNTO] <television>', '[WATCH] <television>'] }

Scene_Objects:['spectacles', 'sheets', 'tablelamp', 'toilet', 'chair', 'walllamp', 'nail_polish', 'floor', 'ceiling', 'phone', 'keyboard', 'bookshelf', 'ceilinglamp', 'mat', 'tvstand', 'stereo', 'mousepad', 'couch', 'pillow', 'wall', 'nightstand', 'photoframe', 'novel', 'coffe_maker', 'check', 'television', 'curtain', 'lighter', 'bathroom_cabinet', 'table'] Task:Relax on sofa response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['beer', 'microwave', 'trashcan', 'drawing', 'cpuscreen', 'floor', 'ceiling', 'phone', 'keyboard', 'ceilinglamp', 'mat', 'picture', 'window', 'couch', 'dining_room', 'wall', 'oven', 'powersocket', 'toy', 'photoframe', 'television', 'oven_mitts', 'tray', 'centerpiece', 'kettle', 'towel_rack', 'pencil', 'dishrack', 'table', 'vase'] Task:Relax on sofa response: { "thoughts": go to the living room and sit on the couch, ease into the sofa and relax., "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[LIE] <couch>'] }

Scene_Objects:['chair', 'walllamp', 'bathroom', 'detergent', 'food_butter', 'orchid', 'wallshelf', 'floor', 'ceiling', 'character', 'bedroom', 'button', 'table_cloth', 'stovefan', 'mat', 'brush', 'stereo', 'picture', 'mousepad', 'couch', 'clothes_pants', 'food_jam', 'powersocket', 'oven', 'newspaper', 'cards', 'shoes', 'controller', 'clothes_skirt', 'board_game'] Task:Listen to music response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['freezer', 'walllamp', 'bathroom', 'drawing', 'orchid', 'floor', 'ceiling', 'video_game_console', 'desk', 'keyboard', 'stovefan', 'ceilinglamp', 'tvstand', 'mat', 'window', 'couch', 'pillow', 'wall', 'powersocket', 'video_game_controller', 'television', 'food_rice', 'ice', 'curtain', 'doorjamb', 'dresser', 'door', 'bathroom_cabinet', 'table', 'laser_pointer'] Task:Play games response: { "thoughts": Turn on game. Use controller, "actions": ['[WALK] <home_office>', '[WALK] <television>', '[PLUGIN] <television>', '[SWITCHON] <television>', '[FIND] <video_game_console>', '[PLUGIN] <video_game_console>', '[SWITCHON] <video_game_console>', '[FIND] <video_game_console>', '[FIND] <video_game_controller>', '[GRAB] <video_game_controller>', '[WALK] <couch>', '[SIT] <couch>'] }

Scene_Objects:['sponge', 'glue', 'toilet', 'chair', 'detergent', 'dough', 'drawing', 'orchid', 'hanger', 'floor', 'ceiling', 'keyboard', 'alcohol', 'stovefan', 'ceilinglamp', 'couch', 'microphone', 'wall', 'nightstand', 'photoframe', 'coffe_maker', 'coffee_filter', 'doorjamb', 'kitchen_counter', 'hairbrush', 'cupboard', 'home_office', 'dishwasher', 'coffee', 'shoe_rack'] Task:Make coffee response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['dry_pasta', 'freezer', 'walllamp', 'bathroom', 'dough', 'knife', 'orchid', 'thread', 'computer', 'ceilinglamp', 'deck_of_cards', 'shower', 'clothes_socks', 'food_noodles', 'dustpan', 'nightstand', 'photoframe', 'light_bulb', 'television', 'coffee_filter', 'mirror', 'curtain', 'doorjamb', 'laundry_detergent', 'sink', 'bag', 'table', 'soap', 'clothes_shirt', 'creditcard'] Task:Look at mirror response: { "thoughts": I look at my reflection to check my cleanliness., "actions": ['[WALK] <bathroom>', '[WALK] <mirror>', '[FIND] <mirror>', '[TURNTO] <mirror>', '[LOOKAT] <mirror>'] }

Scene_Objects:['microwave', 'dog', 'walllamp', 'cat', 'knifeblock', 'drawing', 'bills', 'floor', 'ceiling', 'clothes_jacket', 'toaster', 'ceilinglamp', 'plate', 'faucet', 'iron', 'dish_soap', 'food_noodles', 'mousepad', 'couch', 'wall', 'nightstand', 'mouthwash', 'curtain', 'mouse', 'kitchen_counter', 'cupboard', 'conditioner', 'door', 'sink', 'dishwasher'] Task:Wash dishes by hand response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'microwave', 'walllamp', 'trashcan', 'drawing', 'wallshelf', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'bookshelf', 'mat', 'tvstand', 'ceilinglamp', 'shower', 'window', 'couch', 'pillow', 'wall', 'nightstand', 'oven', 'keys', 'television', 'curtain', 'doorjamb', 'towel_rack', 'mouse', 'door', 'light', 'table'] Task:Relax on sofa response: { "thoughts": I want to relax on the sofa to some relaxing music in the living room., "actions": ['[WALK] <home_office>', '[WALK] <light>', '[FIND] <light>', '[SWITCHON] <light>', '[WALK] <stereo>', '[PLUGIN] <stereo>', '[SWITCHON] <stereo>', '[WALK] <couch>', '[SIT] <couch>', '[LIE] <couch>'] }

Scene_Objects:['sponge', 'toilet', 'freezer', 'drawing', 'bathroom_counter', 'floor', 'ceiling', 'keyboard', 'ceilinglamp', 'mat', 'shower', 'food_noodles', 'couch', 'towel', 'wall', 'powersocket', 'oven', 'cutting_board', 'photoframe', 'newspaper', 'water', 'food_donut', 'curtain', 'doorjamb', 'headset', 'door', 'bathroom_cabinet', 'light', 'soap', 'tooth_paste'] Task:Take shower response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['dry_pasta', 'pot', 'water_glass', 'tablelamp', 'microwave', 'chair', 'freezer', 'drawing', 'orchid', 'floor', 'ceiling', 'phone', 'tea_bag', 'toaster', 'stovefan', 'mat', 'shower', 'faucet', 'ceilinglamp', 'oil', 'wall', 'oven', 'powersocket', 'coffe_maker', 'food_food', 'water', 'ice', 'tray', 'doorjamb', 'towel_rack', 'coffee_pot', 'kitchen_counter', 'cupboard', 'door', 'sink', 'bag', 'table', 'dishwasher', 'light'] Task:Drink response: { "thoughts": Going to the kitchen to get a drink of water., "actions": ['[WALK] <dining_room>', '[WALK] <cupboard>', '[FIND] <cupboard>', '[OPEN] <cupboard>', '[FIND] <water_glass>', '[GRAB] <water_glass>', '[WALK] <faucet>', '[SWITCHON] <faucet>', '[FIND] <water>', '[FIND] <sink>', '[PUTBACK] <water_glass> <sink>', '[GRAB] <water_glass>', '[SWITCHOFF] <faucet>', '[DRINK] <water_glass>', '[FIND] <kitchen_counter>', '[PUTBACK] <water_glass> <kitchen_counter>'] }

Scene_Objects:['dry_pasta', 'toilet', 'chair', 'freezer', 'detergent', 'food_butter', 'wallshelf', 'hanger', 'phone', 'character', 'desk', 'remote_control', 'button', 'piano_bench', 'toaster', 'mat', 'tvstand', 'picture', 'oil', 'couch', 'clothes_pants', 'pillow', 'towel', 'wall', 'novel', 'razor', 'diary', 'conditioner', 'table', 'coffee'] Task:Change TV channel response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['toilet', 'detergent', 'drawing', 'orchid', 'cpuscreen', 'hanger', 'bathroom_counter', 'floor', 'character', 'desk', 'keyboard', 'thread', 'table_cloth', 'tvstand', 'balanceball', 'shower', 'mousepad', 'couch', 'wall', 'photoframe', 'bed', 'toilet_paper', 'food_onion', 'towel_rack', 'mouse', 'kitchen_counter', 'cd', 'hairbrush', 'bathroom_cabinet', 'light'] Task:Go to toilet response: { "thoughts": walk to bathroom , run to toilet, "actions": ['[WALK] <bathroom>', '[RUN] <toilet>'] }

Scene_Objects:['pot', 'food_pizza', 'food_salt', 'microwave', 'freezer', 'orchid', 'drawing', 'hanger', 'button', 'table_cloth', 'tvstand', 'homework', 'couch', 'dining_room', 'nightstand', 'keys', 'coffe_maker', 'television', 'oven_mitts', 'curtain', 'doorjamb', 'diary', 'cards', 'kitchen_counter', 'comb', 'dresser', 'food_steak', 'home_office', 'light', 'clothes_skirt'] Task:Pet cat response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['walllamp', 'orchid', 'drawing', 'wallshelf', 'floor', 'ceiling', 'phone', 'bookshelf', 'remote_control', 'computer', 'mat', 'tvstand', 'ceilinglamp', 'shower', 'window', 'couch', 'clothes_hat', 'pillow', 'wall', 'nightstand', 'powersocket', 'photoframe', 'newspaper', 'television', 'curtain', 'doorjamb', 'door', 'home_office', 'light', 'table'] Task:Watch TV response: { "thoughts": walk to living room. look at remote control. grab remote control. plug in television. switch on television. lie in sofa. point at remote control. watch television., "actions": ['[WALK] <home_office>', '[WALK] <remote_control>', '[FIND] <remote_control>', '[TURNTO] <remote_control>', '[LOOKAT] <remote_control>', '[GRAB] <remote_control>', '[FIND] <television>', '[PLUGIN] <television>', '[SWITCHON] <television>', '[FIND] <couch>', '[LIE] <couch>', '[TURNTO] <remote_control>', '[POINTAT] <remote_control>', '[TURNTO] <television>', '[WATCH] <television>'] }

Scene_Objects:['sponge', 'spectacles', 'pot', 'microwave', 'knifeblock', 'floor', 'ceiling', 'folder', 'character', 'toaster', 'plate', 'drying_rack', 'faucet', 'dish_soap', 'couch', 'wall', 'powersocket', 'food_cheese', 'photoframe', 'bed', 'food_food', 'mouse', 'kitchen_counter', 'cupboard', 'door', 'sink', 'hairdryer', 'light', 'dirt', 'coffee'] Task:Wash dishes by hand response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'filing_cabinet', 'bathroom', 'detergent', 'food_butter', 'drawing', 'wallshelf', 'floor', 'desk', 'bookshelf', 'plate', 'faucet', 'picture', 'iron', 'window', 'dish_soap', 'food_jam', 'wall', 'nightstand', 'toy', 'cutting_board', 'food_food', 'television', 'curtain', 'pajamas', 'conditioner', 'light', 'dishwasher', 'tooth_paste', 'shoe_rack'] Task:Wash dishes with dishwasher response: { "thoughts": Go to the kitchen. Find the dishwasher. Open the dishwasher. Find the plates. Load the plates into the dishwasher. Find the soap. Load the soap into the dishwasher. Close the dishwasher. Turn on the dishwasher., "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[FIND] <dishwasher>', '[OPEN] <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[PUTBACK] <plate> <dishwasher>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[PUTBACK] <dish_soap> <dishwasher>', '[CLOSE] <dishwasher>', '[SWITCHON] <dishwasher>'] }

Scene_Objects:['sponge', 'sheets', 'filing_cabinet', 'crayon', 'toilet', 'bathroom_counter', 'ceiling', 'folder', 'shower', 'iron', 'window', 'mousepad', 'food_peanut_butter', 'pillow', 'wall', 'oven', 'toy', 'basket_for_clothes', 'photoframe', 'novel', 'coffe_maker', 'food_food', 'tray', 'cd', 'dishrack', 'door', 'napkin', 'light', 'dishwasher', 'board_game'] Task:Read book response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'walllamp', 'drawing', 'hanger', 'floor', 'ceiling', 'bookshelf', 'ceilinglamp', 'laptop', 'faucet', 'mat', 'closetdrawer', 'teeth', 'dining_room', 'wall', 'pasta', 'toy', 'shelf', 'standingmirror', 'novel', 'check', 'television', 'curtain', 'console', 'dresser', 'napkin', 'table', 'light', 'scrabble', 'coffee'] Task:Look at painting response: { "thoughts": Walk to the painting. Look up at the painting. Admire it on the wall., "actions": ['[WALK] <home_office>', '[WALK] <drawing>', '[FIND] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>'] }

Scene_Objects:['tablelamp', 'microwave', 'chair', 'dog', 'toothbrush', 'knifeblock', 'wallshelf', 'hanger', 'bathroom_counter', 'floor', 'phone', 'ceiling', 'character', 'desk', 'mat', 'tvstand', 'shower', 'homework', 'dining_room', 'wall', 'bed', 'clothes_gloves', 'toothbrush_holder', 'after_shave', 'doorjamb', 'mouse', 'pencil', 'cd', 'tooth_paste', 'shoe_rack'] Task:Wash teeth response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'scissors', 'walllamp', 'needle', 'wallshelf', 'cpuscreen', 'hanger', 'bedroom', 'keyboard', 'bookmark', 'stovefan', 'shower', 'oil', 'dining_room', 'wall', 'powersocket', 'bench', 'love_seat', 'standingmirror', 'novel', 'coffe_maker', 'check', 'after_shave', 'towel_rack', 'mouse', 'pencil', 'kitchen_counter', 'maindoor', 'bathtub', 'soap'] Task:Read book response: { "thoughts": Obtain a book from the book shelf. Sit in the easy chair. Read the book. Mark the stopping point with a bookmark. Put the book back on the book shelf., "actions": ['[WALK] <home_office>', '[WALK] <novel>', '[FIND] <novel>', '[GRAB] <novel>', '[FIND] <love_seat>', '[SIT] <love_seat>', '[READ] <novel>', '[FIND] <bookmark>', '[GRAB] <bookmark>', '[PUTBACK] <bookmark> <novel>', '[PUTOBJBACK] <novel>'] }

Scene_Objects:['fork', 'juice', 'food_salt', 'measuring_cup', 'chair', 'detergent', 'milk', 'drawing', 'kitchen_cabinet', 'floor', 'shampoo', 'dvd_player', 'keyboard', 'computer', 'table_cloth', 'stovefan', 'ceilinglamp', 'plate', 'tvstand', 'candle', 'towel', 'food_cheese', 'light_bulb', 'television', 'razor', 'curtain', 'kitchen_counter', 'headset', 'table', 'soap'] Task:Set up table response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'food_pizza', 'tablelamp', 'food_ice_cream', 'crayon', 'shampoo', 'keyboard', 'mail', 'bookmark', 'stovefan', 'faucet', 'pillow', 'dining_room', 'broom', 'wall', 'love_seat', 'photoframe', 'novel', 'food_food', 'wine', 'tray', 'after_shave', 'shaving_cream', 'kettle', 'headset', 'pajamas', 'laundry_detergent', 'door', 'bathroom_cabinet', 'shoe_rack'] Task:Read book response: { "thoughts": Obtain a book from the book shelf. Sit in the easy chair. Read the book. Mark the stopping point with a bookmark. Put the book back on the book shelf., "actions": ['[WALK] <home_office>', '[WALK] <novel>', '[FIND] <novel>', '[GRAB] <novel>', '[FIND] <love_seat>', '[SIT] <love_seat>', '[READ] <novel>', '[FIND] <bookmark>', '[GRAB] <bookmark>', '[PUTBACK] <bookmark> <novel>', '[PUTOBJBACK] <novel>'] }

Scene_Objects:['food_salt', 'ceiling', 'phone', 'bookshelf', 'ceilinglamp', 'shoe-shine_kit', 'tvstand', 'homework', 'mousepad', 'food_jam', 'form', 'pillow', 'dining_room', 'wall', 'powersocket', 'bed', 'television', 'tray', 'curtain', 'doorjamb', 'pencil', 'dresser', 'controller', 'laundry_detergent', 'door', 'napkin', 'light', 'dishwasher', 'table', 'teeth'] Task:Pick up phone response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'tea', 'drawing', 'cpuscreen', 'floor', 'ceiling', 'desk', 'bookshelf', 'keyboard', 'computer', 'piano_bench', 'mat', 'shoe-shine_kit', 'balanceball', 'oil', 'clothes_hat', 'dining_room', 'wall', 'powersocket', 'photoframe', 'bed', 'food_food', 'mouthwash', 'doorjamb', 'towel_rack', 'vacuum_cleaner', 'door', 'home_office', 'light', 'soap'] Task:Turn on light response: { "thoughts": switching on the light, "actions": ['[WALK] <bedroom>', '[WALK] <light>', '[FIND] <light>', '[SWITCHON] <light>'] }

Scene_Objects:['sheets', 'food_pizza', 'toilet', 'freezer', 'ground_coffee', 'bathroom', 'drawing', 'knifeblock', 'floor', 'ceiling', 'desk', 'bookshelf', 'alcohol', 'mat', 'shower', 'window', 'dining_room', 'food_kiwi', 'nightstand', 'wall', 'powersocket', 'photoframe', 'food_food', 'television', 'doorjamb', 'mouse', 'door', 'light', 'colander', 'coffee'] Task:Put groceries in Fridge response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'toilet', 'bathroom', 'ceilingfan', 'drawing', 'floor', 'ceiling', 'character', 'desk', 'tea_bag', 'remote_control', 'ceilinglamp', 'tvstand', 'balanceball', 'picture', 'window', 'couch', 'pillow', 'dining_room', 'wall', 'powersocket', 'cutting_board', 'novel', 'check', 'television', 'curtain', 'conditioner', 'hairdryer', 'clothes_dress', 'table'] Task:Watch TV response: { "thoughts": watching TV from sofa, "actions": ['[WALK] <home_office>', '[WALK] <television>', '[FIND] <television>', '[PLUGIN] <television>', '[SWITCHON] <television>', '[FIND] <couch>', '[SIT] <couch>', '[FIND] <remote_control>', '[GRAB] <remote_control>', '[DROP] <remote_control>', '[TURNTO] <television>', '[LOOKAT] <television>'] }

Scene_Objects:['pot', 'mop', 'microwave', 'walllamp', 'detergent', 'food_carrot', 'knifeblock', 'wallshelf', 'hanger', 'floor', 'ceiling', 'keyboard', 'stovefan', 'toaster', 'stereo', 'window', 'oil', 'clothes_socks', 'wall', 'oven', 'food_vegetable', 'bed', 'photoframe', 'television', 'food_donut', 'water', 'tray', 'kitchen_counter', 'cupboard', 'dishwasher'] Task:Cook some food response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'cat', 'food_carrot', 'drawing', 'knifeblock', 'cpuscreen', 'wallshelf', 'hanger', 'bathroom_counter', 'ceiling', 'desk', 'bookshelf', 'keyboard', 'mousepad', 'bench', 'newspaper', 'coffe_maker', 'food_food', 'tray', 'curtain', 'mouse', 'door', 'hairdryer', 'sink', 'food_steak', 'light', 'dishwasher', 'spoon', 'table', 'board_game'] Task:Write school paper response: { "thoughts": Writing and drafting a paper for class work, "actions": ['[WALK] <home_office>', '[WALK] <computer>', '[SWITCHON] <computer>', '[SWITCHOFF] <computer>'] }

Scene_Objects:['tablelamp', 'toilet', 'chair', 'walllamp', 'cat', 'food_carrot', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'desk', 'keyboard', 'bookshelf', 'computer', 'laptop', 'ceilinglamp', 'homework', 'mousepad', 'wall', 'coffe_maker', 'video_game_controller', 'ice', 'doorjamb', 'diary', 'mouse', 'kitchen_counter', 'shoes', 'comb', 'dresser', 'hairdryer'] Task:Write an email response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'microwave', 'toilet', 'walllamp', 'bathroom', 'drawing', 'orchid', 'knifeblock', 'floor', 'ceiling', 'bookshelf', 'stovefan', 'couch', 'pillow', 'wall', 'nightstand', 'powersocket', 'toy', 'novel', 'coffe_maker', 'television', 'wooden_spoon', 'doorjamb', 'kitchen_counter', 'shoes', 'cupboard', 'door', 'home_office', 'table', 'light'] Task:Read book response: { "thoughts": Go to the living room and find the book you want to read. Take it and then go to the couch. Sit on it, open the book and start reading., "actions": ['[WALK] <home_office>', '[WALK] <novel>', '[FIND] <novel>', '[GRAB] <novel>', '[FIND] <couch>', '[SIT] <couch>', '[READ] <novel>'] }

Scene_Objects:['spectacles', 'filing_cabinet', 'blender', 'walllamp', 'orchid', 'cpuscreen', 'hanger', 'floor', 'bedroom', 'bookshelf', 'stovefan', 'food_peanut_butter', 'clothes_hat', 'clothes_pants', 'dining_room', 'nightstand', 'wall_clock', 'video_game_controller', 'food_donut', 'curtain', 'kettle', 'shoes', 'dresser', 'cupboard', 'hairbrush', 'controller', 'home_office', 'light', 'clothes_skirt', 'soap'] Task:Listen to music response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'microwave', 'orchid', 'drawing', 'hanger', 'floor', 'phone', 'character', 'keyboard', 'toaster', 'plate', 'faucet', 'shower', 'window', 'mousepad', 'closetdrawer', 'dining_room', 'oven', 'powersocket', 'cup', 'bed', 'food_food', 'coffe_maker', 'tray', 'doorjamb', 'dresser', 'sink', 'bathtub', 'home_office', 'table'] Task:Drink response: { "thoughts": I go to the kitchen. I get a glass. I fill it with water. I take a drink from the glass., "actions": ['[WALK] <dining_room>', '[WALK] <cup>', '[FIND] <cup>', '[GRAB] <cup>', '[WALK] <water>', '[FIND] <water>', '[GRAB] <water>', '[POUR] <water> <cup>', '[DRINK] <cup>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'toilet', 'bathroom', 'knifeblock', 'cpuscreen', 'hanger', 'bathroom_counter', 'ceiling', 'folder', 'character', 'keyboard', 'bookshelf', 'envelope', 'computer', 'dining_room', 'wall', 'nightstand', 'cup', 'clothes_gloves', 'light_bulb', 'check', 'curtain', 'pencil', 'kitchen_counter', 'controller', 'sink', 'napkin', 'clothes_skirt', 'tooth_paste'] Task:Browse internet response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['freezer', 'cat', 'drawing', 'knife', 'knifeblock', 'hanger', 'floor', 'bedroom', 'keyboard', 'computer', 'piano_bench', 'ceilinglamp', 'faucet', 'iron', 'window', 'mousepad', 'couch', 'form', 'dining_room', 'wall', 'oven', 'box', 'newspaper', 'bed', 'novel', 'toilet_paper', 'television', 'tray', 'sink', 'table'] Task:Pet cat response: { "thoughts": I walk to the living room. I locate the cat. I walk to the couch and sit down beside the cat. I pet the cat., "actions": ['[WALK] <home_office>', '[WALK] <cat>', '[FIND] <cat>', '[WALK] <couch>', '[SIT] <couch>', '[TOUCH] <cat>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'beer', 'measuring_cup', 'microwave', 'dog', 'orchid', 'wallshelf', 'floor', 'bedroom', 'computer', 'mop_bucket', 'ceilinglamp', 'faucet', 'electric_shaver', 'couch', 'food_peanut_butter', 'broom', 'cutting_board', 'photoframe', 'newspaper', 'mouthwash', 'ice', 'tray', 'curtain', 'kitchen_counter', 'shoes', 'headset', 'bathroom_cabinet', 'board_game'] Task:Pet cat response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'crayon', 'toilet', 'chair', 'bathroom', 'orchid', 'floor', 'ceiling', 'character', 'bedroom', 'desk', 'computer', 'tvstand', 'faucet', 'stereo', 'food_orange', 'photoframe', 'novel', 'television', 'curtain', 'doorjamb', 'diary', 'towel_rack', 'mouse', 'shoes', 'dresser', 'cupboard', 'sink', 'bathtub', 'home_office'] Task:Get dressed response: { "thoughts": I get dressed in my bedroom in the mornings., "actions": ['[WALK] <bedroom>', '[WALK] <shoes>', '[FIND] <shoes>', '[GRAB] <shoes>', '[PUTON] <shoes>'] }

Scene_Objects:['sponge', 'microwave', 'cpuscreen', 'floor', 'ceiling', 'phone', 'character', 'folder', 'desk', 'alcohol', 'stovefan', 'plate', 'mousepad', 'wall', 'oven', 'food_food', 'check', 'tray', 'doorjamb', 'cards', 'mouse', 'kitchen_counter', 'shoes', 'cupboard', 'clothes_dress', 'sink', 'clothes_scarf', 'light', 'clothes_shirt', 'teeth'] Task:Pick up phone response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'mop', 'microwave', 'freezer', 'walllamp', 'detergent', 'toothbrush', 'orchid', 'drawing', 'clothes_underwear', 'floor', 'thread', 'mat', 'closetdrawer', 'wall', 'nightstand', 'toy', 'standingmirror', 'light_bulb', 'novel', 'check', 'tray', 'cards', 'mouse', 'pencil', 'vacuum_cleaner', 'headset', 'maindoor', 'napkin', 'light'] Task:Read book response: { "thoughts": Walk to the Living Room. Look at the Cabinet. Find a Textbook. Grab the textbook. Open the Textbook. Read the textbook., "actions": ['[WALK] <home_office>', '[WALK] <filing_cabinet>', '[FIND] <filing_cabinet>', '[TURNTO] <filing_cabinet>', '[LOOKAT] <filing_cabinet>', '[FIND] <novel>', '[GRAB] <novel>', '[READ] <novel>'] }

Scene_Objects:['pot', 'freezer', 'blow_dryer', 'bathroom', 'nail_polish', 'drawing', 'knife', 'cpuscreen', 'character', 'bookshelf', 'table_cloth', 'mop_bucket', 'bookmark', 'faucet', 'homework', 'nightstand', 'wall', 'novel', 'check', 'coffee_filter', 'coffee_pot', 'shaving_cream', 'kitchen_counter', 'pajamas', 'cupboard', 'clothes_dress', 'laundry_detergent', 'napkin', 'table', 'washing_machine'] Task:Wash clothes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['measuring_cup', 'mop', 'walllamp', 'clothes_underwear', 'hanger', 'floor', 'ceiling', 'dvd_player', 'desk', 'tvstand', 'window', 'clothes_socks', 'closetdrawer', 'oil', 'dining_room', 'wall', 'toy', 'photoframe', 'television', 'curtain', 'rag', 'doorjamb', 'console', 'kitchen_counter', 'shoes', 'dresser', 'pajamas', 'bathroom_cabinet', 'table', 'tooth_paste'] Task:Put away shoes response: { "thoughts": Pick up shoes from entry hall. Open entry way closet. Put shoes in closet. Close closet door., "actions": ['[WALK] <home_office>', '[WALK] <shoes>', '[FIND] <shoes>', '[GRAB] <shoes>', '[FIND] <dresser>', '[OPEN] <dresser>', '[PUTIN] <shoes> <dresser>', '[CLOSE] <dresser>'] }

Scene_Objects:['sheets', 'filing_cabinet', 'mop', 'bathroom', 'dough', 'knifeblock', 'hanger', 'keyboard', 'alcohol', 'toaster', 'shower', 'foundation', 'couch', 'bowl', 'wall', 'pasta', 'basket_for_clothes', 'check', 'television', 'razor', 'tray', 'doorjamb', 'chessboard', 'dishrack', 'hairbrush', 'napkin', 'bathroom_cabinet', 'light', 'dishwasher', 'table'] Task:Pick up phone response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'toilet', 'freezer', 'ground_coffee', 'ceilingfan', 'drawing', 'orchid', 'floor', 'ceiling', 'character', 'video_game_console', 'table_cloth', 'ceilinglamp', 'tvstand', 'balanceball', 'window', 'couch', 'dustpan', 'pillow', 'wall', 'cup', 'video_game_controller', 'television', 'curtain', 'dishrack', 'napkin', 'table', 'vase', 'scrabble', 'shoe_rack'] Task:Play games response: { "thoughts": Turn on game. Use controller, "actions": ['[WALK] <home_office>', '[WALK] <television>', '[PLUGIN] <television>', '[SWITCHON] <television>', '[FIND] <video_game_console>', '[PLUGIN] <video_game_console>', '[SWITCHON] <video_game_console>', '[FIND] <video_game_console>', '[FIND] <video_game_controller>', '[GRAB] <video_game_controller>', '[WALK] <couch>', '[SIT] <couch>'] }

Scene_Objects:['pot', 'toilet', 'freezer', 'walllamp', 'drawing', 'knifeblock', 'floor', 'ceiling', 'food_chicken', 'food_cake', 'stovefan', 'mat', 'toaster', 'shower', 'food_noodles', 'oil', 'nightstand', 'wall', 'powersocket', 'oven', 'cup', 'photoframe', 'food_food', 'tray', 'doorjamb', 'kitchen_counter', 'cupboard', 'door', 'light', 'dishwasher'] Task:Cook some food response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'filing_cabinet', 'juice', 'microwave', 'knifeblock', 'wallshelf', 'bathroom_counter', 'floor', 'ceiling', 'computer', 'piano_bench', 'toaster', 'ceilinglamp', 'plate', 'faucet', 'stovefan', 'tvstand', 'dish_soap', 'wall', 'pasta', 'coffee_filter', 'tray', 'doorjamb', 'mouse', 'kitchen_counter', 'shoes', 'dishrack', 'cupboard', 'sink', 'scrabble'] Task:Wash dishes response: { "thoughts": Place all of the dirty dishes in a sink full of clean, soapy water. Wash them. Rinse them in fresh, running water. Place them in the strainer to dry., "actions": ['[WALK] <dining_room>', '[WALK] <sink>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[FIND] <plate>', '[GRAB] <plate>', '[RINSE] <plate>', '[PUTBACK] <plate> <sink>', '[FIND] <plate>', '[GRAB] <plate>', '[RINSE] <plate>', '[PUTBACK] <plate> <sink>', '[SWITCHOFF] <faucet>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[FIND] <sponge>', '[GRAB] <sponge>', '[POUR] <dish_soap> <sponge>', '[PUTOBJBACK] <dish_soap>', '[GRAB] <plate>', '[SCRUB] <plate>', '[PUTOBJBACK] <plate>', '[GRAB] <plate>', '[SCRUB] <plate>', '[PUTOBJBACK] <plate>', '[SWITCHON] <faucet>', '[RINSE] <sponge>', '[PUTOBJBACK] <sponge>', '[GRAB] <plate>', '[RINSE] <plate>', '[FIND] <dishrack>', '[PUTBACK] <plate> <dishrack>', '[GRAB] <plate>', '[RINSE] <plate>', '[PUTBACK] <plate> <dishrack>', '[SWITCHOFF] <faucet>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'blender', 'chair', 'walllamp', 'blow_dryer', 'detergent', 'orchid', 'knife', 'hanger', 'floor', 'ceiling', 'thread', 'ceilinglamp', 'laptop', 'faucet', 'mousepad', 'couch', 'candle', 'food_jam', 'pillow', 'wall', 'nightstand', 'tape', 'television', 'tray', 'after_shave', 'food_onion', 'shoes', 'cupboard'] Task:Relax on sofa response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'band-aids', 'walllamp', 'drawing', 'food_egg', 'bathroom_counter', 'ceiling', 'stamp', 'desk', 'cleaning_solution', 'computer', 'stovefan', 'deck_of_cards', 'faucet', 'tvstand', 'window', 'couch', 'wall', 'nightstand', 'coffe_maker', 'television', 'mirror', 'after_shave', 'paper_towel', 'clothes_dress', 'sink', 'bag', 'bathroom_cabinet', 'dishwasher', 'laser_pointer'] Task:Clean mirror response: { "thoughts": First spray the mirrors with Windex or window type spray. Take paper towels and wipe until sparkly clean. If they have streaks, repeat the process., "actions": ['[WALK] <bathroom>', '[WALK] <bathroom_cabinet>', '[OPEN] <bathroom_cabinet>', '[FIND] <paper_towel>', '[GRAB] <paper_towel>', '[CLOSE] <bathroom_cabinet>', '[WALK] <mirror>', '[FIND] <cleaning_solution>', '[RINSE] <cleaning_solution>', '[WIPE] <mirror>'] }

Scene_Objects:['pot', 'filing_cabinet', 'tablelamp', 'chair', 'cpuscreen', 'wallshelf', 'hanger', 'floor', 'phone', 'bedroom', 'button', 'plate', 'dish_soap', 'wall', 'nightstand', 'food_cheese', 'photoframe', 'newspaper', 'coffe_maker', 'check', 'tray', 'kettle', 'pencil', 'kitchen_counter', 'comb', 'cupboard', 'door', 'table', 'dishwasher', 'light'] Task:Wash dishes with dishwasher response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'spectacles', 'microwave', 'walllamp', 'drawing', 'orchid', 'hanger', 'floor', 'character', 'desk', 'alcohol', 'stovefan', 'ceilinglamp', 'mat', 'window', 'couch', 'pillow', 'wall', 'oven', 'powersocket', 'cutting_board', 'photoframe', 'light_bulb', 'television', 'curtain', 'kettle', 'napkin', 'bathroom_cabinet', 'table', 'creditcard'] Task:Relax on sofa response: { "thoughts": walk to living room, walk to sofa, sit on sofa., "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[SIT] <couch>'] }

Scene_Objects:['spectacles', 'food_salt', 'microwave', 'walllamp', 'dough', 'hanger', 'ceiling', 'phone', 'character', 'keyboard', 'bookmark', 'toaster', 'mat', 'tvstand', 'stereo', 'window', 'oil', 'food_peanut_butter', 'tape', 'check', 'video_game_controller', 'razor', 'oven_mitts', 'food_rice', 'curtain', 'kitchen_counter', 'hairbrush', 'food_steak', 'bathroom_cabinet', 'home_office'] Task:Listen to music response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'chair', 'walllamp', 'drawing', 'cpuscreen', 'hanger', 'floor', 'ceiling', 'desk', 'keyboard', 'computer', 'stovefan', 'ceilinglamp', 'toaster', 'stereo', 'shower', 'mousepad', 'closetdrawer', 'dustpan', 'wall', 'powersocket', 'box', 'television', 'fax_machine', 'doorjamb', 'mouse', 'dresser', 'light', 'table', 'printing_paper'] Task:Print out papers response: { "thoughts": Turn on printer and fill with blank paper. Use computer to click on printing option. Take printed page out of print., "actions": ['[WALK] <home_office>', '[WALK] <fax_machine>', '[FIND] <fax_machine>', '[SWITCHON] <fax_machine>', '[FIND] <printing_paper>', '[GRAB] <printing_paper>', '[PUTBACK] <printing_paper> <fax_machine>', '[FIND] <computer>', '[FIND] <mouse>', '[GRAB] <mouse>', '[GRAB] <printing_paper>'] }

Scene_Objects:['fork', 'pot', 'filing_cabinet', 'tablelamp', 'blender', 'beer', 'toilet', 'jelly', 'freezer', 'blow_dryer', 'detergent', 'orchid', 'knifeblock', 'wallshelf', 'cpuscreen', 'bathroom_counter', 'floor', 'mail', 'computer', 'bookmark', 'stereo', 'window', 'couch', 'microphone', 'check', 'television', 'rag', 'pencil', 'controller', 'dishwasher'] Task:Pet cat response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'bathroom', 'knife', 'bathroom_counter', 'floor', 'shampoo', 'ceiling', 'stovefan', 'mat', 'faucet', 'mousepad', 'couch', 'food_peanut_butter', 'dustpan', 'towel', 'wall', 'wall_clock', 'oven', 'novel', 'ice', 'tray', 'rag', 'shoes', 'maindoor', 'sink', 'bathroom_cabinet', 'vase', 'soap', 'tooth_paste', 'hands_both'] Task:Wash hands response: { "thoughts": I go to the sink, turn on the water, put soap on my hands, rub them together, then rinse them off and turn off the water. I then towel off my hands., "actions": ['[WALK] <bathroom>', '[WALK] <sink>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[FIND] <soap>', '[GRAB] <soap>', '[FIND] <hands_both>', '[WASH] <hands_both>', '[RINSE] <hands_both>', '[SWITCHOFF] <faucet>', '[FIND] <towel>', '[GRAB] <towel>', '[FIND] <hands_both>', '[WIPE] <hands_both>'] }

Scene_Objects:['filing_cabinet', 'tablelamp', 'toilet', 'walllamp', 'detergent', 'orchid', 'floor', 'ceiling', 'keyboard', 'remote_control', 'computer', 'button', 'alcohol', 'tvstand', 'window', 'closetdrawer', 'couch', 'pillow', 'nightstand', 'wall', 'check', 'mouthwash', 'television', 'coffe_maker', 'curtain', 'dresser', 'clothes_dress', 'door', 'sink', 'table'] Task:Change TV channel response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['glue', 'filing_cabinet', 'chair', 'walllamp', 'drawing', 'knifeblock', 'bathroom_counter', 'floor', 'bedroom', 'desk', 'computer', 'ceilinglamp', 'deck_of_cards', 'stereo', 'window', 'couch', 'clothes_pants', 'wall', 'cup', 'novel', 'toilet_paper', 'television', 'wine', 'kitchen_counter', 'dishrack', 'comb', 'dresser', 'coffee_cup', 'table', 'teeth'] Task:Relax on sofa response: { "thoughts": you walk to the living room, open the door, find a comfy sofa, sit on it, read a book or sip a cup of coffee and relax., "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[FIND] <couch>', '[FIND] <coffee_cup>', '[GRAB] <coffee_cup>', '[SIT] <couch>', '[FIND] <novel>', '[GRAB] <novel>', '[READ] <novel>', '[DRINK] <coffee_cup>'] }

Scene_Objects:['tablelamp', 'beer', 'crayon', 'chair', 'freezer', 'walllamp', 'orchid', 'drawing', 'hanger', 'floor', 'ceiling', 'bedroom', 'bookshelf', 'food_cake', 'ceilinglamp', 'tvstand', 'mat', 'couch', 'food_peanut_butter', 'pillow', 'wall', 'nightstand', 'powersocket', 'cup', 'food_food', 'television', 'curtain', 'pencil', 'kitchen_counter', 'door'] Task:Relax on sofa response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['glue', 'mop', 'toilet', 'cat', 'trashcan', 'ceilingfan', 'drawing', 'knifeblock', 'orchid', 'floor', 'ceiling', 'computer', 'clothes_jacket', 'table_cloth', 'stovefan', 'ceilinglamp', 'tvstand', 'faucet', 'window', 'couch', 'wall', 'nightstand', 'food_food', 'mouthwash', 'television', 'food_donut', 'curtain', 'after_shave', 'dresser', 'table'] Task:Look at painting response: { "thoughts": Walk to the painting. Look up at the painting. Admire it on the wall., "actions": ['[WALK] <home_office>', '[WALK] <drawing>', '[FIND] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>'] }

Scene_Objects:['tablelamp', 'juice', 'toilet', 'freezer', 'walllamp', 'detergent', 'milk', 'drawing', 'floor', 'ceiling', 'phone', 'computer', 'toaster', 'mat', 'ceilinglamp', 'shower', 'food_orange', 'pillow', 'dining_room', 'nightstand', 'wall', 'powersocket', 'photoframe', 'newspaper', 'food_food', 'toothbrush_holder', 'doorjamb', 'door', 'sink', 'light'] Task:Put groceries in Fridge response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['dry_pasta', 'food_ice_cream', 'walllamp', 'nail_polish', 'cat', 'orchid', 'drawing', 'floor', 'ceiling', 'purse', 'alcohol', 'mail', 'table_cloth', 'mat', 'tvstand', 'ceilinglamp', 'window', 'couch', 'pillow', 'towel', 'wall', 'oven', 'cup', 'food_cheese', 'television', 'curtain', 'doorjamb', 'door', 'table', 'clothes_shirt'] Task:Straighten paintings on wall response: { "thoughts": Go up to the painting and tilt it which ever way it is uneven until it looks straight., "actions": ['[WALK] <home_office>', '[WALK] <wall>', '[WALK] <drawing>', '[FIND] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>', '[TOUCH] <drawing>', '[PULL] <drawing>', '[PUSH] <drawing>', '[WALK] <drawing>', '[FIND] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>', '[TOUCH] <drawing>', '[PULL] <drawing>', '[PUSH] <drawing>', '[WALK] <drawing>', '[FIND] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>', '[TOUCH] <drawing>', '[PULL] <drawing>', '[PUSH] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>', '[TURNTO] <drawing>', '[LOOKAT] <drawing>'] }

Scene_Objects:['pot', 'filing_cabinet', 'measuring_cup', 'microwave', 'orchid', 'knifeblock', 'wallshelf', 'character', 'keyboard', 'stovefan', 'mat', 'closetdrawer', 'dining_room', 'food_kiwi', 'box', 'cup', 'bed', 'coffe_maker', 'television', 'water', 'colander', 'tray', 'curtain', 'kitchen_counter', 'cd', 'cupboard', 'door', 'bag', 'soap', 'creditcard'] Task:Wash hands response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'toilet', 'chair', 'bathroom', 'detergent', 'orchid', 'knifeblock', 'drawing', 'bathroom_counter', 'bedroom', 'keyboard', 'button', 'toaster', 'tvstand', 'food_noodles', 'pillow', 'check', 'television', 'oven_mitts', 'wine', 'curtain', 'rag', 'doorjamb', 'kettle', 'mouse', 'pencil', 'hairbrush', 'food_steak', 'home_office', 'dishwasher'] Task:Pet cat response: { "thoughts": Find the cat; pet the cat by touching it, "actions": ['[WALK] <home_office>', '[WALK] <cat>', '[FIND] <cat>', '[TOUCH] <cat>'] }

Scene_Objects:['food_bread', 'microwave', 'toilet', 'walllamp', 'knifeblock', 'wallshelf', 'floor', 'shampoo', 'character', 'computer', 'bookmark', 'toaster', 'plate', 'faucet', 'shoe-shine_kit', 'dish_soap', 'wall', 'toy', 'light_bulb', 'coffe_maker', 'rag', 'kitchen_counter', 'hairbrush', 'cupboard', 'dresser', 'controller', 'sink', 'home_office', 'table', 'dishwasher'] Task:Wash dishes by hand response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'chair', 'ground_coffee', 'walllamp', 'food_carrot', 'orchid', 'drawing', 'floor', 'ceiling', 'bedroom', 'bookshelf', 'keyboard', 'mail', 'mat', 'faucet', 'wall', 'nightstand', 'powersocket', 'oven', 'cup', 'photoframe', 'bed', 'novel', 'curtain', 'doorjamb', 'diary', 'cupboard', 'door', 'sink', 'instrument_guitar'] Task:Read book response: { "thoughts": Walk over to the bookshelf. Pick storybook off the shelf. Sit on bed. Open storybook and read out loud., "actions": ['[WALK] <bedroom>', '[WALK] <bookshelf>', '[FIND] <novel>', '[GRAB] <novel>', '[TURNTO] <novel>', '[LOOKAT] <novel>', '[WALK] <bed>', '[SIT] <bed>', '[READ] <novel>'] }

Scene_Objects:['dry_pasta', 'pot', 'filing_cabinet', 'food_ice_cream', 'jelly', 'bathroom', 'orchid', 'knifeblock', 'wallshelf', 'bathroom_counter', 'desk', 'keyboard', 'ceilinglamp', 'iron', 'stereo', 'closetdrawer', 'clothes_pants', 'food_jam', 'cup', 'photoframe', 'toothbrush_holder', 'tray', 'curtain', 'centerpiece', 'coffee_pot', 'kitchen_counter', 'door', 'spoon', 'dishwasher', 'clothes_shirt'] Task:Pick up phone response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['mop', 'chair', 'jelly', 'walllamp', 'drawing', 'wallshelf', 'food_egg', 'hanger', 'floor', 'ceiling', 'character', 'desk', 'remote_control', 'computer', 'mat', 'tvstand', 'ceilinglamp', 'shower', 'closetdrawer', 'couch', 'pillow', 'wall', 'photoframe', 'television', 'food_rice', 'doorjamb', 'diary', 'kitchen_counter', 'dresser', 'coffee'] Task:Watch horror movie response: { "thoughts": I would find something scary to watch on dvd, "actions": ['[WALK] <home_office>', '[WALK] <remote_control>', '[FIND] <remote_control>', '[GRAB] <remote_control>', '[FIND] <couch>', '[SIT] <couch>', '[FIND] <television>', '[SWITCHON] <television>', '[TURNTO] <television>', '[WATCH] <television>'] }

Scene_Objects:['sheets', 'water_glass', 'pot', 'glue', 'filing_cabinet', 'knife', 'wallshelf', 'bathroom_counter', 'stamp', 'bedroom', 'keyboard', 'clothes_jacket', 'piano_bench', 'electric_shaver', 'food_jam', 'cutting_board', 'novel', 'check', 'coffee_filter', 'water', 'ice', 'curtain', 'food_onion', 'pencil', 'kitchen_counter', 'cd', 'headset', 'dresser', 'door', 'light'] Task:Drink response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'pot', 'food_bread', 'tablelamp', 'scissors', 'microwave', 'chair', 'walllamp', 'hanger', 'floor', 'food_cake', 'toaster', 'plate', 'shower', 'tvstand', 'faucet', 'dish_soap', 'bowl', 'wall', 'oven', 'cup', 'basket_for_clothes', 'photoframe', 'coffe_maker', 'food_food', 'tray', 'curtain', 'maindoor', 'table', 'dishwasher'] Task:Wash dishes with dishwasher response: { "thoughts": Using dishwasher to wash dishes, "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[FIND] <dishwasher>', '[OPEN] <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[FIND] <bowl>', '[GRAB] <bowl>', '[WALK] <dishwasher>', '[PUTBACK] <bowl> <dishwasher>', '[GRAB] <bowl>', '[PUTBACK] <bowl> <dishwasher>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[PUTBACK] <dish_soap> <dishwasher>', '[CLOSE] <dishwasher>', '[SWITCHON] <dishwasher>'] }

Scene_Objects:['sheets', 'blender', 'food_salt', 'crayon', 'tea', 'cat', 'milk', 'knifeblock', 'hanger', 'shampoo', 'bedroom', 'bookshelf', 'computer', 'brush', 'stovefan', 'toaster', 'foundation', 'window', 'mousepad', 'clothes_pants', 'clothes_hat', 'tape', 'bed', 'coffe_maker', 'food_onion', 'after_shave', 'dresser', 'light', 'dishwasher', 'laser_pointer'] Task:Pet cat response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tablelamp', 'microwave', 'walllamp', 'bathroom', 'trashcan', 'drawing', 'orchid', 'bathroom_counter', 'floor', 'ceiling', 'bedroom', 'keyboard', 'mat', 'ceilinglamp', 'window', 'mousepad', 'closetdrawer', 'couch', 'pillow', 'wall', 'wall_clock', 'cup', 'video_game_controller', 'television', 'curtain', 'doorjamb', 'cupboard', 'door', 'light', 'table'] Task:Gaze out window response: { "thoughts": Look at the yard from the window, "actions": ['[WALK] <home_office>', '[WALK] <window>', '[FIND] <window>', '[TURNTO] <window>', '[LOOKAT] <window>'] }

Scene_Objects:['fork', 'filing_cabinet', 'food_pizza', 'knifeblock', 'food_egg', 'hanger', 'phone', 'desk', 'tea_bag', 'purse', 'envelope', 'clothes_jacket', 'table_cloth', 'faucet', 'window', 'closetdrawer', 'clothes_pants', 'food_jam', 'wall', 'basket_for_clothes', 'bed', 'cd', 'comb', 'laundry_detergent', 'clothes_scarf', 'napkin', 'home_office', 'bathroom_cabinet', 'washing_machine', 'teeth'] Task:Wash clothes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sponge', 'pot', 'food_bread', 'tablelamp', 'scissors', 'microwave', 'chair', 'walllamp', 'hanger', 'floor', 'food_cake', 'toaster', 'plate', 'shower', 'tvstand', 'faucet', 'dish_soap', 'bowl', 'wall', 'oven', 'cup', 'basket_for_clothes', 'photoframe', 'coffe_maker', 'food_food', 'tray', 'curtain', 'maindoor', 'table', 'dishwasher'] Task:Wash dishes with dishwasher response: { "thoughts": Using dishwasher to wash dishes, "actions": ['[WALK] <dining_room>', '[WALK] <dishwasher>', '[FIND] <dishwasher>', '[OPEN] <dishwasher>', '[FIND] <plate>', '[GRAB] <plate>', '[FIND] <bowl>', '[GRAB] <bowl>', '[WALK] <dishwasher>', '[PUTBACK] <bowl> <dishwasher>', '[GRAB] <bowl>', '[PUTBACK] <bowl> <dishwasher>', '[FIND] <dish_soap>', '[GRAB] <dish_soap>', '[PUTBACK] <dish_soap> <dishwasher>', '[CLOSE] <dishwasher>', '[SWITCHON] <dishwasher>'] }

Scene_Objects:['toilet', 'chair', 'dog', 'ground_coffee', 'detergent', 'floor', 'ceiling', 'shampoo', 'envelope', 'clothes_jacket', 'piano_bench', 'ceilinglamp', 'mat', 'couch', 'dining_room', 'wall', 'coffe_maker', 'food_food', 'television', 'coffee_filter', 'water', 'coffee_pot', 'console', 'kitchen_counter', 'dresser', 'cupboard', 'pajamas', 'bathroom_cabinet', 'light', 'dishwasher'] Task:Make coffee response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['food_bread', 'filing_cabinet', 'tablelamp', 'food_ice_cream', 'beer', 'crayon', 'chair', 'knifeblock', 'cpuscreen', 'bathroom_counter', 'phone', 'desk', 'bedroom', 'bookshelf', 'thread', 'toaster', 'ceilinglamp', 'mousepad', 'closetdrawer', 'pillow', 'wall', 'cup', 'bed', 'keys', 'food_food', 'television', 'food_onion', 'doorjamb', 'mouse', 'home_office'] Task:Lock door response: { "thoughts": Turn the lock., "actions": ['[WALK] <bedroom>'] }

Scene_Objects:['fork', 'tablelamp', 'toilet', 'chair', 'walllamp', 'floor', 'ceiling', 'phone', 'keyboard', 'envelope', 'toaster', 'ceilinglamp', 'stereo', 'faucet', 'window', 'couch', 'pillow', 'wall', 'nightstand', 'powersocket', 'cup', 'clothes_gloves', 'novel', 'check', 'video_game_controller', 'television', 'kitchen_counter', 'hairdryer', 'light', 'table'] Task:Relax on sofa response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'tablelamp', 'microwave', 'chair', 'walllamp', 'blow_dryer', 'knifeblock', 'drawing', 'orchid', 'kitchen_cabinet', 'floor', 'ceiling', 'food_cake', 'mat', 'faucet', 'mousepad', 'couch', 'wall', 'oven', 'bench', 'toilet_paper', 'television', 'water', 'drinking_glass', 'curtain', 'kitchen_counter', 'cupboard', 'sink', 'light', 'table'] Task:Drink response: { "thoughts": I go to the kitchen to get a glass from the cabinet. I turn on the faucet and fill up the glass. I take a drink., "actions": ['[WALK] <dining_room>', '[WALK] <kitchen_cabinet>', '[FIND] <kitchen_cabinet>', '[OPEN] <kitchen_cabinet>', '[FIND] <drinking_glass>', '[GRAB] <drinking_glass>', '[WALK] <sink>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[PUTBACK] <drinking_glass> <sink>', '[FIND] <water>', '[GRAB] <drinking_glass>', '[PUTBACK] <drinking_glass> <sink>', '[GRAB] <drinking_glass>', '[DRINK] <drinking_glass>'] }

Scene_Objects:['tablelamp', 'chair', 'walllamp', 'milk', 'orchid', 'drawing', 'knifeblock', 'floor', 'ceiling', 'folder', 'computer', 'ceilinglamp', 'tvstand', 'window', 'closetdrawer', 'couch', 'pillow', 'dining_room', 'wall', 'nightstand', 'cup', 'bed', 'television', 'razor', 'toothbrush_holder', 'tray', 'curtain', 'dresser', 'laundry_detergent', 'table'] Task:Watch TV response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['sheets', 'filing_cabinet', 'food_pizza', 'walllamp', 'orchid', 'needle', 'cpuscreen', 'bathroom_counter', 'shampoo', 'desk', 'bookshelf', 'computer', 'stovefan', 'ceilinglamp', 'toaster', 'shower', 'faucet', 'couch', 'dining_room', 'cup', 'bed', 'tray', 'curtain', 'wooden_spoon', 'pencil', 'door', 'bathroom_cabinet', 'light', 'dishwasher', 'table'] Task:Pick up phone response: { "thoughts": I walk to the home office. I find the telephone. I pick up the telephone handset., "actions": ['[WALK] <home_office>', '[WALK] <phone>', '[FIND] <phone>', '[GRAB] <phone>'] }

Scene_Objects:['dry_pasta', 'tablelamp', 'chair', 'walllamp', 'orchid', 'floor', 'ceiling', 'phone', 'bookshelf', 'remote_control', 'computer', 'ceilinglamp', 'tvstand', 'window', 'oil', 'closetdrawer', 'couch', 'pillow', 'dining_room', 'nightstand', 'wall', 'oven', 'bed', 'coffe_maker', 'television', 'curtain', 'dresser', 'door', 'instrument_guitar', 'table'] Task:Watch TV response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['spectacles', 'tablelamp', 'chair', 'freezer', 'walllamp', 'drawing', 'wallshelf', 'cpuscreen', 'floor', 'ceiling', 'phone', 'desk', 'keyboard', 'computer', 'mat', 'tvstand', 'laptop', 'mousepad', 'couch', 'pillow', 'wall', 'oven', 'tape', 'photoframe', 'television', 'doorjamb', 'mouse', 'kitchen_counter', 'door', 'light'] Task:Use laptop response: { "thoughts": Sit down on a couch and perform tasks such as navigating the internet on a laptop, "actions": ['[WALK] <home_office>', '[WALK] <couch>', '[FIND] <couch>', '[SIT] <couch>', '[FIND] <laptop>', '[SWITCHON] <laptop>', '[TURNTO] <laptop>', '[LOOKAT] <laptop>', '[FIND] <keyboard>', '[TYPE] <keyboard>', '[TURNTO] <keyboard>', '[LOOKAT] <keyboard>', '[TYPE] <keyboard>', '[TURNTO] <laptop>', '[LOOKAT] <laptop>'] }

Scene_Objects:['chair', 'detergent', 'food_butter', 'clothes_underwear', 'hanger', 'floor', 'stamp', 'bookshelf', 'computer', 'clothes_jacket', 'table_cloth', 'mat', 'shoe-shine_kit', 'clothes_socks', 'closetdrawer', 'couch', 'clothes_pants', 'basket_for_clothes', 'check', 'mouthwash', 'food_rice', 'vacuum_cleaner', 'laundry_detergent', 'clothes_dress', 'bathroom_cabinet', 'light', 'dishwasher', 'clothes_shirt', 'washing_machine', 'creditcard'] Task:Wash clothes response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['filing_cabinet', 'food_ice_cream', 'chair', 'wallshelf', 'bathroom_counter', 'floor', 'ceiling', 'phone', 'desk', 'keyboard', 'stovefan', 'mat', 'faucet', 'shower', 'window', 'mousepad', 'couch', 'towel', 'wall', 'bed', 'photoframe', 'curtain', 'after_shave', 'kitchen_counter', 'sink', 'bathroom_cabinet', 'light', 'soap', 'hands_both', 'shoe_rack'] Task:Wash hands response: { "thoughts": Wash hands with soap and hot water, "actions": ['[WALK] <bathroom>', '[WALK] <sink>', '[FIND] <faucet>', '[SWITCHON] <faucet>', '[FIND] <soap>', '[GRAB] <soap>', '[FIND] <hands_both>', '[SCRUB] <hands_both>', '[PUTOBJBACK] <soap>', '[RINSE] <hands_both>', '[SWITCHOFF] <faucet>', '[FIND] <towel>', '[GRAB] <towel>', '[WIPE] <hands_both>', '[PUTOBJBACK] <towel>'] }

Scene_Objects:['juice', 'microwave', 'toilet', 'ground_coffee', 'bathroom', 'cpuscreen', 'hanger', 'floor', 'phone', 'character', 'bookshelf', 'computer', 'table_cloth', 'mat', 'tvstand', 'shower', 'couch', 'pillow', 'bowl', 'wall', 'powersocket', 'novel', 'food_food', 'video_game_controller', 'check', 'food_onion', 'doorjamb', 'napkin', 'home_office', 'coffee'] Task:Listen to music response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['microwave', 'chair', 'walllamp', 'bathroom', 'drawing', 'cpuscreen', 'bathroom_counter', 'floor', 'ceiling', 'character', 'keyboard', 'alcohol', 'toaster', 'tvstand', 'shower', 'closetdrawer', 'nightstand', 'wall_clock', 'bench', 'powersocket', 'food_food', 'doorjamb', 'towel_rack', 'mouse', 'cupboard', 'sink', 'home_office', 'table', 'light', 'coffee'] Task:Walk through response: { "thoughts": Walk through the hallway., "actions": ['[WALK] <home_office>', '[WALK] <dining_room>', '[WALK] <home_office>'] }

Scene_Objects:['dry_pasta', 'pot', 'microwave', 'walllamp', 'toothbrush', 'drawing', 'knifeblock', 'wallshelf', 'keyboard', 'food_cake', 'ceilinglamp', 'tvstand', 'shower', 'stereo', 'oil', 'tooth_paste', 'couch', 'closetdrawer', 'dining_room', 'toy', 'basket_for_clothes', 'light_bulb', 'check', 'centerpiece', 'kettle', 'mouse', 'kitchen_counter', 'shoes', 'light', 'board_game'] Task:Brush teeth response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['pot', 'crayon', 'tea', 'beer', 'microwave', 'toilet', 'trashcan', 'food_carrot', 'knifeblock', 'cpuscreen', 'bathroom_counter', 'floor', 'character', 'bedroom', 'desk', 'keyboard', 'bookshelf', 'computer', 'ceilinglamp', 'window', 'dining_room', 'food_kiwi', 'nightstand', 'bed', 'photoframe', 'doorjamb', 'dishrack', 'cupboard', 'sink', 'home_office'] Task:Open bathroom window response: { "thoughts": Make sure bathroom window is unlocked. Lift up bathroom window so it is open., "actions": ['[WALK] <bathroom>', '[WALK] <window>', '[FIND] <window>', '[OPEN] <window>'] }

Scene_Objects:['blender', 'toilet', 'freezer', 'walllamp', 'drawing', 'bathroom_counter', 'floor', 'ceiling', 'ceilinglamp', 'shoe-shine_kit', 'shower', 'faucet', 'mat', 'oil', 'food_peanut_butter', 'towel', 'wall', 'nightstand', 'powersocket', 'photoframe', 'bed', 'toilet_paper', 'curtain', 'doorjamb', 'kitchen_counter', 'door', 'sink', 'bathroom_cabinet', 'light', 'colander'] Task:Take shower response:

You serve as an AI task planner. 1. Your task is to create a plan to achieve a goal by converting it into a sequence of actions. Each action should follow the format "[verb] <target1> <target2>", where 'verb' represents the action, and 'target1' and 'target2' are optional arguments. 2. You must describe your plan in natural language at the beginning. After that, you should list all the actions together. Ensure that you response can be parsed by Python json.loads. Here is an example plan to achieve a goal for reference: Scene_Objects:['tea', 'toilet', 'freezer', 'band-aids', 'knifeblock', 'food_egg', 'shampoo', 'phone', 'bookshelf', 'purse', 'thread', 'deck_of_cards', 'shower', 'picture', 'clothes_socks', 'oven', 'food_cheese', 'novel', 'food_food', 'check', 'coffe_maker', 'tray', 'after_shave', 'curtain', 'doorjamb', 'cupboard', 'controller', 'bag', 'table', 'scrabble'] Task:Drink response: { "thoughts": Pour water into a glass and drink from the glass., "actions": ['[WALK] <dining_room>', '[WALK] <water_glass>', '[FIND] <water_glass>', '[GRAB] <water_glass>', '[WALK] <water>', '[FIND] <water>', '[GRAB] <water>', '[POUR] <water> <water_glass>', '[DRINK] <water_glass>'] }

Scene_Objects:['pot', 'filing_cabinet', 'freezer', 'walllamp', 'drawing', 'needle', 'orchid', 'floor', 'stamp', 'keyboard', 'plate', 'faucet', 'shower', 'dish_soap', 'oil', 'wall', 'keys', 'coffe_maker', 'coffee_filter', 'pencil', 'kitchen_counter', 'dishrack', 'dresser', 'cupboard', 'controller', 'pajamas', 'food_steak', 'spoon', 'light', 'dishwasher'] Task:Wash dishes with dishwasher response: