id
stringlengths 24
37
| image
imagewidth (px) 270
720
| conversations
listlengths 2
2
|
|---|---|---|
aitw_14527044287448665684_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's the US dollar exchange rateagainst the Swiss Franc?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Swipe [UP]\nStep 2: Tap [GOOGLE Icon]\nStep 3: Tap [good restaurant in miami florida]\nStep 4: Tap [X Icon]\n\nCurrent screenshot is labeled with 36 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"What's the US dollar exchange rateagainst the Swiss Franc?\"}"
}
] |
|
aitw_14527044287448665684_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's the US dollar exchange rateagainst the Swiss Franc?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Swipe [UP]\nStep 2: Tap [GOOGLE Icon]\nStep 3: Tap [good restaurant in miami florida]\nStep 4: Tap [X Icon]\nStep 5: Type [What's the US dollar exchange rateagainst the Swiss Franc?]\n\nCurrent screenshot is labeled with 16 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 10}}"
}
] |
|
aitw_14527044287448665684_7
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's the US dollar exchange rateagainst the Swiss Franc?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Swipe [UP]\nStep 2: Tap [GOOGLE Icon]\nStep 3: Tap [good restaurant in miami florida]\nStep 4: Tap [X Icon]\nStep 5: Type [What's the US dollar exchange rateagainst the Swiss Franc?]\nStep 6: Tap [...erateagainst the franc today swiss]\n\nCurrent screenshot is labeled with 48 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 7) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_6520755420117064583_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How do I get to the nearest Apple Store?\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 10 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.7, 0.79]}}"
}
] |
|
aitw_6520755420117064583_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How do I get to the nearest Apple Store?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 27 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 6}}"
}
] |
|
aitw_6520755420117064583_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How do I get to the nearest Apple Store?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/flights?c]\n\nCurrent screenshot is labeled with 9 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"nearest Apple Store\"}"
}
] |
|
aitw_6520755420117064583_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How do I get to the nearest Apple Store?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/flights?c]\nStep 2: Type [nearest Apple Store]\n\nCurrent screenshot is labeled with 5 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_6520755420117064583_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How do I get to the nearest Apple Store?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/flights?c]\nStep 2: Type [nearest Apple Store]\nStep 3: Tap [nearest Apple Store]\n\nCurrent screenshot is labeled with 32 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_11638207615492812576_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 11 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"PRESS\": \"HOME\"}"
}
] |
|
aitw_11638207615492812576_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\n\nCurrent screenshot is labeled with 9 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.67, 0.78]}}"
}
] |
|
aitw_11638207615492812576_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 12 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.68, 0.79]}}"
}
] |
|
aitw_11638207615492812576_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 12 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.68, 0.79]}}"
}
] |
|
aitw_11638207615492812576_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 15 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_11638207615492812576_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [google.com/search?]\n\nCurrent screenshot is labeled with 16 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 0}}"
}
] |
|
aitw_11638207615492812576_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [google.com/search?]\nStep 5: Tap [X Icon]\n\nCurrent screenshot is labeled with 13 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"good restaurant in Austin\"}"
}
] |
|
aitw_11638207615492812576_7
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [google.com/search?]\nStep 5: Tap [X Icon]\nStep 6: Type [good restaurant in Austin]\n\nCurrent screenshot is labeled with 5 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 7) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_11638207615492812576_8
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [google.com/search?]\nStep 5: Tap [X Icon]\nStep 6: Type [good restaurant in Austin]\nStep 7: Tap [good in Austin restaurant]\n\nCurrent screenshot is labeled with 31 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 8) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"UP\"}"
}
] |
|
aitw_11638207615492812576_9
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [google.com/search?]\nStep 5: Tap [X Icon]\nStep 6: Type [good restaurant in Austin]\nStep 7: Tap [good in Austin restaurant]\nStep 8: Swipe [UP]\n\nCurrent screenshot is labeled with 19 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 9) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 15}}"
}
] |
|
aitw_11638207615492812576_10
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a goodrestaurant in Austin?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [google.com/search?]\nStep 5: Tap [X Icon]\nStep 6: Type [good restaurant in Austin]\nStep 7: Tap [good in Austin restaurant]\nStep 8: Swipe [UP]\nStep 9: Tap [Olamaie 4.5 ***t $$$ (576) Southern 1610 San Antoni Closed Opens 5PM Modern Southern menu in a polished space]\n\nCurrent screenshot is labeled with 19 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 10) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_11550999498863875080_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 18 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"PRESS\": \"HOME\"}"
}
] |
|
aitw_11550999498863875080_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\n\nCurrent screenshot is labeled with 8 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 7}}"
}
] |
|
aitw_11550999498863875080_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\n\nCurrent screenshot is labeled with 12 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 11}}"
}
] |
|
aitw_11550999498863875080_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\n\nCurrent screenshot is labeled with 23 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 5}}"
}
] |
|
aitw_11550999498863875080_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\n\nCurrent screenshot is labeled with 9 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"new foundation (skincare) product\"}"
}
] |
|
aitw_11550999498863875080_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\n\nCurrent screenshot is labeled with 5 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_11550999498863875080_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\n\nCurrent screenshot is labeled with 25 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"UP\"}"
}
] |
|
aitw_11550999498863875080_7
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\n\nCurrent screenshot is labeled with 21 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 7) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"DOWN\"}"
}
] |
|
aitw_11550999498863875080_8
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\nStep 7: Swipe [DOWN]\n\nCurrent screenshot is labeled with 24 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 8) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 18}}"
}
] |
|
aitw_11550999498863875080_9
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\nStep 7: Swipe [DOWN]\nStep 8: Tap [Foundation Skincare | Dermatologist Formulated Hair Growth]\n\nCurrent screenshot is labeled with 18 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 9) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"UP\"}"
}
] |
|
aitw_11550999498863875080_10
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\nStep 7: Swipe [DOWN]\nStep 8: Tap [Foundation Skincare | Dermatologist Formulated Hair Growth]\nStep 9: Swipe [UP]\n\nCurrent screenshot is labeled with 14 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 10) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"UP\"}"
}
] |
|
aitw_11550999498863875080_11
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\nStep 7: Swipe [DOWN]\nStep 8: Tap [Foundation Skincare | Dermatologist Formulated Hair Growth]\nStep 9: Swipe [UP]\nStep 10: Swipe [UP]\n\nCurrent screenshot is labeled with 18 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 11) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 0}}"
}
] |
|
aitw_11550999498863875080_12
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\nStep 7: Swipe [DOWN]\nStep 8: Tap [Foundation Skincare | Dermatologist Formulated Hair Growth]\nStep 9: Swipe [UP]\nStep 10: Swipe [UP]\nStep 11: Tap [X Icon]\n\nCurrent screenshot is labeled with 22 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 12) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 9}}"
}
] |
|
aitw_11550999498863875080_13
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new foundation (skincare) product\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [90]\nStep 2: Tap [90]\nStep 3: Tap [amazon.com/EGO-P]\nStep 4: Type [new foundation (skincare) product]\nStep 5: Tap [(skincare) new foundation product]\nStep 6: Swipe [UP]\nStep 7: Swipe [DOWN]\nStep 8: Tap [Foundation Skincare | Dermatologist Formulated Hair Growth]\nStep 9: Swipe [UP]\nStep 10: Swipe [UP]\nStep 11: Tap [X Icon]\nStep 12: Tap [help? we]\n\nCurrent screenshot is labeled with 21 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 13) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_16429259049941881261_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for top rated burger restaurants on Maps\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 11 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"UP\"}"
}
] |
|
aitw_16429259049941881261_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for top rated burger restaurants on Maps\n\nPrevious actions:\nStep 0: Swipe [UP]\n\nCurrent screenshot is labeled with 34 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 26}}"
}
] |
|
aitw_16429259049941881261_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for top rated burger restaurants on Maps\n\nPrevious actions:\nStep 0: Swipe [UP]\nStep 1: Tap [°OO Maps Messages Phone]\n\nCurrent screenshot is labeled with 13 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_16429259049941881261_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for top rated burger restaurants on Maps\n\nPrevious actions:\nStep 0: Swipe [UP]\nStep 1: Tap [°OO Maps Messages Phone]\nStep 2: Tap [Search here]\n\nCurrent screenshot is labeled with 26 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"top rated burger restaurants\"}"
}
] |
|
aitw_16429259049941881261_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for top rated burger restaurants on Maps\n\nPrevious actions:\nStep 0: Swipe [UP]\nStep 1: Tap [°OO Maps Messages Phone]\nStep 2: Tap [Search here]\nStep 3: Type [top rated burger restaurants]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_16429259049941881261_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for top rated burger restaurants on Maps\n\nPrevious actions:\nStep 0: Swipe [UP]\nStep 1: Tap [°OO Maps Messages Phone]\nStep 2: Tap [Search here]\nStep 3: Type [top rated burger restaurants]\nStep 4: Tap [top rated burger restaurants near me K]\n\nCurrent screenshot is labeled with 12 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_7956947281322954257_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for good Chinese restaurants\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 11 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.63, 0.83]}}"
}
] |
|
aitw_7956947281322954257_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for good Chinese restaurants\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 14 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 3}}"
}
] |
|
aitw_7956947281322954257_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for good Chinese restaurants\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [fiveguys.co.uk/menu/]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"good Chinese restaurants\"}"
}
] |
|
aitw_7956947281322954257_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for good Chinese restaurants\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [fiveguys.co.uk/menu/]\nStep 2: Type [good Chinese restaurants]\n\nCurrent screenshot is labeled with 20 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 12}}"
}
] |
|
aitw_7956947281322954257_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for good Chinese restaurants\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [fiveguys.co.uk/menu/]\nStep 2: Type [good Chinese restaurants]\nStep 3: Tap [good Chinese restaurants]\n\nCurrent screenshot is labeled with 19 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_450729698699239895_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a good restaurant in Atlanta?\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 13 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"PRESS\": \"HOME\"}"
}
] |
|
aitw_450729698699239895_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a good restaurant in Atlanta?\n\nPrevious actions:\nStep 0: Press [HOME]\n\nCurrent screenshot is labeled with 10 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.63, 0.83]}}"
}
] |
|
aitw_450729698699239895_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a good restaurant in Atlanta?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 2}}"
}
] |
|
aitw_450729698699239895_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a good restaurant in Atlanta?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [bloomberg.cominews/articl]\n\nCurrent screenshot is labeled with 4 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"a good restaurant in Atlanta\"}"
}
] |
|
aitw_450729698699239895_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a good restaurant in Atlanta?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [bloomberg.cominews/articl]\nStep 3: Type [a good restaurant in Atlanta]\n\nCurrent screenshot is labeled with 5 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_450729698699239895_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's a good restaurant in Atlanta?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [bloomberg.cominews/articl]\nStep 3: Type [a good restaurant in Atlanta]\nStep 4: Tap [a good restaurant in Atlanta]\n\nCurrent screenshot is labeled with 22 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_8083769447948585687_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What is the price of a 12' ladder at Home Depot?\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 10 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.64, 0.83]}}"
}
] |
|
aitw_8083769447948585687_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What is the price of a 12' ladder at Home Depot?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 3 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 2}}"
}
] |
|
aitw_8083769447948585687_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What is the price of a 12' ladder at Home Depot?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [ic-cricket.com/news/2848!]\n\nCurrent screenshot is labeled with 4 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"price of a 12' ladder at Home Depot\"}"
}
] |
|
aitw_8083769447948585687_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What is the price of a 12' ladder at Home Depot?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [ic-cricket.com/news/2848!]\nStep 2: Type [price of a 12' ladder at Home Depot]\n\nCurrent screenshot is labeled with 5 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 4}}"
}
] |
|
aitw_8083769447948585687_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What is the price of a 12' ladder at Home Depot?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [ic-cricket.com/news/2848!]\nStep 2: Type [price of a 12' ladder at Home Depot]\nStep 3: Tap [price of a 12' ladder at Home Depot]\n\nCurrent screenshot is labeled with 14 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_675110900048817819_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's on the menu at Panera?\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 10 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"PRESS\": \"HOME\"}"
}
] |
|
aitw_675110900048817819_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's on the menu at Panera?\n\nPrevious actions:\nStep 0: Press [HOME]\n\nCurrent screenshot is labeled with 10 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.62, 0.83]}}"
}
] |
|
aitw_675110900048817819_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's on the menu at Panera?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 8 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 2}}"
}
] |
|
aitw_675110900048817819_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's on the menu at Panera?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [google.con/search?q=atga KruTContomporary VJapanese Cuisine Safeway Canonast Sacramentb Fae NEWTON BOOTH Man data 82n22 Gge Cafė Dantorels 4.4 k**** (1.3K) $$ Ad Restaurant-2700 24th St Closed Opens 8AM Wed Takecut No delvery Dine-in]\n\nCurrent screenshot is labeled with 4 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"Panera menu\"}"
}
] |
|
aitw_675110900048817819_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's on the menu at Panera?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [google.con/search?q=atga KruTContomporary VJapanese Cuisine Safeway Canonast Sacramentb Fae NEWTON BOOTH Man data 82n22 Gge Cafė Dantorels 4.4 k**** (1.3K) $$ Ad Restaurant-2700 24th St Closed Opens 8AM Wed Takecut No delvery Dine-in]\nStep 3: Type [Panera menu]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 5}}"
}
] |
|
aitw_675110900048817819_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: What's on the menu at Panera?\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [google.con/search?q=atga KruTContomporary VJapanese Cuisine Safeway Canonast Sacramentb Fae NEWTON BOOTH Man data 82n22 Gge Cafė Dantorels 4.4 k**** (1.3K) $$ Ad Restaurant-2700 24th St Closed Opens 8AM Wed Takecut No delvery Dine-in]\nStep 3: Type [Panera menu]\nStep 4: Tap [Panera menu]\n\nCurrent screenshot is labeled with 20 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_9462289622230282202_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 8 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.69, 0.79]}}"
}
] |
|
aitw_9462289622230282202_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 19 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 8}}"
}
] |
|
aitw_9462289622230282202_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/search?q=gooc]\n\nCurrent screenshot is labeled with 18 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 0}}"
}
] |
|
aitw_9462289622230282202_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/search?q=gooc]\nStep 2: Tap [X Icon]\n\nCurrent screenshot is labeled with 17 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 5}}"
}
] |
|
aitw_9462289622230282202_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/search?q=gooc]\nStep 2: Tap [X Icon]\nStep 3: Tap [Search or type web address]\n\nCurrent screenshot is labeled with 16 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"How much does the new iPad cost\"}"
}
] |
|
aitw_9462289622230282202_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/search?q=gooc]\nStep 2: Tap [X Icon]\nStep 3: Tap [Search or type web address]\nStep 4: Type [How much does the new iPad cost]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 6}}"
}
] |
|
aitw_9462289622230282202_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: How much does the new iPad cost?\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [google.com/search?q=gooc]\nStep 2: Tap [X Icon]\nStep 3: Tap [Search or type web address]\nStep 4: Type [How much does the new iPad cost]\nStep 5: Tap [How much does the new iPad cost]\n\nCurrent screenshot is labeled with 39 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_3773553433338569519_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 10 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.7, 0.8]}}"
}
] |
|
aitw_3773553433338569519_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 33 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 8}}"
}
] |
|
aitw_3773553433338569519_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [crs4rec.com/shelter/?gclid:]\n\nCurrent screenshot is labeled with 32 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 0}}"
}
] |
|
aitw_3773553433338569519_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [crs4rec.com/shelter/?gclid:]\nStep 2: Tap [X Icon]\n\nCurrent screenshot is labeled with 31 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"flights from boston to zurich\"}"
}
] |
|
aitw_3773553433338569519_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [crs4rec.com/shelter/?gclid:]\nStep 2: Tap [X Icon]\nStep 3: Type [flights from boston to zurich]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"\"}"
}
] |
|
aitw_3773553433338569519_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [crs4rec.com/shelter/?gclid:]\nStep 2: Tap [X Icon]\nStep 3: Type [flights from boston to zurich]\nStep 4: Type []\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 6}}"
}
] |
|
aitw_3773553433338569519_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [crs4rec.com/shelter/?gclid:]\nStep 2: Tap [X Icon]\nStep 3: Type [flights from boston to zurich]\nStep 4: Type []\nStep 5: Tap [flights from boston to zurich]\n\nCurrent screenshot is labeled with 33 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 22}}"
}
] |
|
aitw_3773553433338569519_7
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from Boston to Zurich\n\nPrevious actions:\nStep 0: Tap [somewhere on the screen]\nStep 1: Tap [crs4rec.com/shelter/?gclid:]\nStep 2: Tap [X Icon]\nStep 3: Type [flights from boston to zurich]\nStep 4: Type []\nStep 5: Tap [flights from boston to zurich]\nStep 6: Tap [Flights to Zurich - 70% Off On Business Class - Find The Best Offer Business]\n\nCurrent screenshot is labeled with 40 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 7) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_11419303056851211144_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 43 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"PRESS\": \"HOME\"}"
}
] |
|
aitw_11419303056851211144_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\n\nCurrent screenshot is labeled with 9 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.68, 0.8]}}"
}
] |
|
aitw_11419303056851211144_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 14 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.68, 0.8]}}"
}
] |
|
aitw_11419303056851211144_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 41 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 16}}"
}
] |
|
aitw_11419303056851211144_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [airbusinessclass.com/popu]\n\nCurrent screenshot is labeled with 34 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 0}}"
}
] |
|
aitw_11419303056851211144_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [airbusinessclass.com/popu]\nStep 4: Tap [X Icon]\n\nCurrent screenshot is labeled with 33 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"flights from nyc to san francisco\"}"
}
] |
|
aitw_11419303056851211144_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [airbusinessclass.com/popu]\nStep 4: Tap [X Icon]\nStep 5: Type [flights from nyc to san francisco]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 6}}"
}
] |
|
aitw_11419303056851211144_7
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [airbusinessclass.com/popu]\nStep 4: Tap [X Icon]\nStep 5: Type [flights from nyc to san francisco]\nStep 6: Tap [flights from san f nyc to]\n\nCurrent screenshot is labeled with 33 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 7) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 28}}"
}
] |
|
aitw_11419303056851211144_8
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for flights from NYC to San Francisco\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [airbusinessclass.com/popu]\nStep 4: Tap [X Icon]\nStep 5: Type [flights from nyc to san francisco]\nStep 6: Tap [flights from san f nyc to]\nStep 7: Tap [Cheap Flights from New York to San Francisco from $82 | (NYC - SFO)]\n\nCurrent screenshot is labeled with 43 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 8) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_9547779115067061183_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 15 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"PRESS\": \"HOME\"}"
}
] |
|
aitw_9547779115067061183_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\n\nCurrent screenshot is labeled with 9 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.72, 0.79]}}"
}
] |
|
aitw_9547779115067061183_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 13 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.72, 0.79]}}"
}
] |
|
aitw_9547779115067061183_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 13 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"NORMALIZED_COORDINATE\": [0.72, 0.79]}}"
}
] |
|
aitw_9547779115067061183_4
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\n\nCurrent screenshot is labeled with 23 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 4) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 8}}"
}
] |
|
aitw_9547779115067061183_5
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [cowboyworldcorp.com/?gcl]\n\nCurrent screenshot is labeled with 19 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 5) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 0}}"
}
] |
|
aitw_9547779115067061183_6
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [cowboyworldcorp.com/?gcl]\nStep 5: Tap [X Icon]\n\nCurrent screenshot is labeled with 18 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 6) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"new lipstick on sephora\"}"
}
] |
|
aitw_9547779115067061183_7
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [cowboyworldcorp.com/?gcl]\nStep 5: Tap [X Icon]\nStep 6: Type [new lipstick on sephora]\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 7) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"\"}"
}
] |
|
aitw_9547779115067061183_8
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [cowboyworldcorp.com/?gcl]\nStep 5: Tap [X Icon]\nStep 6: Type [new lipstick on sephora]\nStep 7: Type []\n\nCurrent screenshot is labeled with 7 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 8) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 6}}"
}
] |
|
aitw_9547779115067061183_9
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [cowboyworldcorp.com/?gcl]\nStep 5: Tap [X Icon]\nStep 6: Type [new lipstick on sephora]\nStep 7: Type []\nStep 8: Tap [lipstick on sephora new]\n\nCurrent screenshot is labeled with 39 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 9) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 29}}"
}
] |
|
aitw_9547779115067061183_10
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Search for a new lipstick on Sephora\n\nPrevious actions:\nStep 0: Press [HOME]\nStep 1: Tap [somewhere on the screen]\nStep 2: Tap [somewhere on the screen]\nStep 3: Tap [somewhere on the screen]\nStep 4: Tap [cowboyworldcorp.com/?gcl]\nStep 5: Tap [X Icon]\nStep 6: Type [new lipstick on sephora]\nStep 7: Type []\nStep 8: Tap [lipstick on sephora new]\nStep 9: Tap [Ford Beauty Lip Stick - Best Lip Color Tom Collections - Shop Color Now For Your Lip]\n\nCurrent screenshot is labeled with 22 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 10) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SET_TASK_STATUS\": \"COMPLETE\"}"
}
] |
|
aitw_1684888717715898644_0
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Searchfor good French restaurants\n\nPrevious actions:\nNone\n\nCurrent screenshot is labeled with 12 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 0) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"SWIPE\": \"UP\"}"
}
] |
|
aitw_1684888717715898644_1
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Searchfor good French restaurants\n\nPrevious actions:\nStep 0: Swipe [UP]\n\nCurrent screenshot is labeled with 35 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 1) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 2}}"
}
] |
|
aitw_1684888717715898644_2
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Searchfor good French restaurants\n\nPrevious actions:\nStep 0: Swipe [UP]\nStep 1: Tap [LOCATION Icon]\n\nCurrent screenshot is labeled with 15 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 2) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TAP\": {\"MARK\": 2}}"
}
] |
|
aitw_1684888717715898644_3
|
[
{
"from": "user",
"value": "<image>\nImagine you are interacting with an Android device step by step. You can see the screen and know the previous actions through recorded history. The screenshots are labeled with bounding boxes and numeric marks to guide your interactions. Your task is to determine the next action to perform. The possible actions include tapping, swiping, navigating, typing, or setting the task status.\n\nUse the following format for your output, selecting the appropriate key-value pair:\n```json\n{\n \"TAP\": dict, // Use {\"MARK\": mark_index} where `mark_index` is an integer corresponding to the bounding box on the screen, or {\"NORMALIZED_COORDINATE\": [x, y]} where `x` and `y` are float numbers between 0 and 1 representing the normalized coordinates on the screen.\n \"SWIPE\": str, // Choose \"UP\", \"DOWN\", \"LEFT\", or \"RIGHT\"\n \"NAVIGATE\": str, // Choose \"BACK\", \"HOME\", or \"ENTER\"\n \"TYPE\": str, // The text to be typed\n \"SET_TASK_STATUS\": str, // Choose \"COMPLETE\" or \"IMPOSSIBLE\"\n}\n```\n\nCurrent task: Searchfor good French restaurants\n\nPrevious actions:\nStep 0: Swipe [UP]\nStep 1: Tap [LOCATION Icon]\nStep 2: Tap [X Icon]\n\nCurrent screenshot is labeled with 16 bounding boxes and numeric marks to guide your interactions. Determine your next action (Step 3) and provide the output in the specified dictionary format."
},
{
"from": "assistant",
"value": "{\"TYPE\": \"good French restaurants\"}"
}
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.