messages
listlengths 2
200
| images
listlengths 1
3
| custom_id
int64 0
66.2k
⌀ |
|---|---|---|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Paris<image>Thought: Click GOOGLE to access a search engine and find hotel options in Paris.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [212, 1365]}<image>Thought: Click GOOGLE to access the search engine for finding hotels in Paris.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [210, 1366]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"hotels in Paris\" into the search box to initiate a hotel search.\nAction: {\"action\": \"TYPE\", \"value\": \"hotels in Paris\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14426125814766810011_1.png",
"training_data/AITW/aitw_images/general/14426125814766810011_2.png",
"training_data/AITW/aitw_images/general/14426125814766810011_3.png"
] | 1,100
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Paris<image>Thought: Click GOOGLE to access the search engine for finding hotels in Paris.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [210, 1366]}<image>Thought: Type \"hotels in Paris\" into the search box to initiate a hotel search.\nAction: {\"action\": \"TYPE\", \"value\": \"hotels in Paris\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the suggested search results to explore hotel options in Paris.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [216, 182]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14426125814766810011_2.png",
"training_data/AITW/aitw_images/general/14426125814766810011_3.png",
"training_data/AITW/aitw_images/general/14426125814766810011_4.png"
] | 1,101
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Paris<image>Thought: Type \"hotels in Paris\" into the search box to initiate a hotel search.\nAction: {\"action\": \"TYPE\", \"value\": \"hotels in Paris\", \"position\": null}<image>Thought: Click on the suggested search results to explore hotel options in Paris.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [216, 182]}<image>",
"role": "user"
},
{
"content": "Thought: The search for hotels in Paris was successful, indicated by the returned results on the Google screen. Various hotel options and ads are displayed, confirming that relevant information is available.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14426125814766810011_3.png",
"training_data/AITW/aitw_images/general/14426125814766810011_4.png",
"training_data/AITW/aitw_images/general/14426125814766810011_5.png"
] | 1,102
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen, allowing access to settings for notification controls.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_0.png"
] | 1,103
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Press home to return to the main screen, allowing access to settings for notification controls.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to access the settings menu and find notification options for Google Maps.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_0.png",
"training_data/AITW/aitw_images/general/9037664210789009586_1.png"
] | 1,104
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Press home to return to the main screen, allowing access to settings for notification controls.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Scroll down to access the settings menu and find notification options for Google Maps.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on LOCATION to access app settings and manage notifications specifically for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [483, 554]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_0.png",
"training_data/AITW/aitw_images/general/9037664210789009586_1.png",
"training_data/AITW/aitw_images/general/9037664210789009586_2.png"
] | 1,105
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Scroll down to access the settings menu and find notification options for Google Maps.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click on LOCATION to access app settings and manage notifications specifically for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [483, 554]}<image>",
"role": "user"
},
{
"content": "Thought: Click the THREE BARS to open the menu options for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [46, 83]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_1.png",
"training_data/AITW/aitw_images/general/9037664210789009586_2.png",
"training_data/AITW/aitw_images/general/9037664210789009586_3.png"
] | 1,106
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on LOCATION to access app settings and manage notifications specifically for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [483, 554]}<image>Thought: Click the THREE BARS to open the menu options for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [46, 83]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to reach the settings where notifications can be managed.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_2.png",
"training_data/AITW/aitw_images/general/9037664210789009586_3.png",
"training_data/AITW/aitw_images/general/9037664210789009586_4.png"
] | 1,107
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click the THREE BARS to open the menu options for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [46, 83]}<image>Thought: Scroll down to reach the settings where notifications can be managed.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click Settings to access the menu where notification options for apps are typically found.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [84, 848]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_3.png",
"training_data/AITW/aitw_images/general/9037664210789009586_4.png",
"training_data/AITW/aitw_images/general/9037664210789009586_5.png"
] | 1,108
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Scroll down to reach the settings where notifications can be managed.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click Settings to access the menu where notification options for apps are typically found.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [84, 848]}<image>",
"role": "user"
},
{
"content": "Thought: Click on Notifications to access the settings for managing Google Maps notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [60, 780]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_4.png",
"training_data/AITW/aitw_images/general/9037664210789009586_5.png",
"training_data/AITW/aitw_images/general/9037664210789009586_6.png"
] | 1,109
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click Settings to access the menu where notification options for apps are typically found.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [84, 848]}<image>Thought: Click on Notifications to access the settings for managing Google Maps notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [60, 780]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Getting around\" to access notification settings for the Google Maps app.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [200, 145]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_5.png",
"training_data/AITW/aitw_images/general/9037664210789009586_6.png",
"training_data/AITW/aitw_images/general/9037664210789009586_7.png"
] | 1,110
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on Notifications to access the settings for managing Google Maps notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [60, 780]}<image>Thought: Click on \"Getting around\" to access notification settings for the Google Maps app.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [200, 145]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the closures setting to access its toggle for turning off notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [501, 170]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_6.png",
"training_data/AITW/aitw_images/general/9037664210789009586_7.png",
"training_data/AITW/aitw_images/general/9037664210789009586_8.png"
] | 1,111
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on \"Getting around\" to access notification settings for the Google Maps app.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [200, 145]}<image>Thought: Click on the closures setting to access its toggle for turning off notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [501, 170]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the closures option to access notification settings for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [497, 291]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_7.png",
"training_data/AITW/aitw_images/general/9037664210789009586_8.png",
"training_data/AITW/aitw_images/general/9037664210789009586_9.png"
] | 1,112
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on the closures setting to access its toggle for turning off notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [501, 170]}<image>Thought: Click on the closures option to access notification settings for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [497, 291]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the toggle for \"Parking location\" to turn off notifications for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [501, 383]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_8.png",
"training_data/AITW/aitw_images/general/9037664210789009586_9.png",
"training_data/AITW/aitw_images/general/9037664210789009586_10.png"
] | 1,113
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on the closures option to access notification settings for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [497, 291]}<image>Thought: Click on the toggle for \"Parking location\" to turn off notifications for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [501, 383]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the \"Time to leave\" toggle to access its settings for notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [495, 497]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_9.png",
"training_data/AITW/aitw_images/general/9037664210789009586_10.png",
"training_data/AITW/aitw_images/general/9037664210789009586_11.png"
] | 1,114
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on the toggle for \"Parking location\" to turn off notifications for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [501, 383]}<image>Thought: Click on the \"Time to leave\" toggle to access its settings for notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [495, 497]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"nearby\" to access notification settings related to Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [495, 606]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_10.png",
"training_data/AITW/aitw_images/general/9037664210789009586_11.png",
"training_data/AITW/aitw_images/general/9037664210789009586_12.png"
] | 1,115
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on the \"Time to leave\" toggle to access its settings for notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [495, 497]}<image>Thought: Click on \"nearby\" to access notification settings related to Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [495, 606]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to find the Google Maps notification settings related to traffic updates. Click on \"Your commute\" to access its settings. Toggle off the option to stop receiving notifications for commute updates from Google Maps. Status task complete indicates that notifications are now turned off successfully.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [488, 712]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_11.png",
"training_data/AITW/aitw_images/general/9037664210789009586_12.png",
"training_data/AITW/aitw_images/general/9037664210789009586_13.png"
] | 1,116
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on \"nearby\" to access notification settings related to Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [495, 606]}<image>Thought: Scroll down to find the Google Maps notification settings related to traffic updates. Click on \"Your commute\" to access its settings. Toggle off the option to stop receiving notifications for commute updates from Google Maps. Status task complete indicates that notifications are now turned off successfully.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [488, 712]}<image>",
"role": "user"
},
{
"content": "Thought: Click on nearby to access settings for Google Maps notifications related to traffic and events.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 703]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_12.png",
"training_data/AITW/aitw_images/general/9037664210789009586_13.png",
"training_data/AITW/aitw_images/general/9037664210789009586_14.png"
] | 1,117
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Scroll down to find the Google Maps notification settings related to traffic updates. Click on \"Your commute\" to access its settings. Toggle off the option to stop receiving notifications for commute updates from Google Maps. Status task complete indicates that notifications are now turned off successfully.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [488, 712]}<image>Thought: Click on nearby to access settings for Google Maps notifications related to traffic and events.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 703]}<image>",
"role": "user"
},
{
"content": "Thought: Navigate back to access the app notification settings for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [43, 62]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_13.png",
"training_data/AITW/aitw_images/general/9037664210789009586_14.png",
"training_data/AITW/aitw_images/general/9037664210789009586_15.png"
] | 1,118
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Click on nearby to access settings for Google Maps notifications related to traffic and events.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 703]}<image>Thought: Navigate back to access the app notification settings for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [43, 62]}<image>",
"role": "user"
},
{
"content": "Thought: Click the arrow backward to navigate back to the previous menu where app notifications can be adjusted.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [46, 64]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_14.png",
"training_data/AITW/aitw_images/general/9037664210789009586_15.png",
"training_data/AITW/aitw_images/general/9037664210789009586_16.png"
] | 1,119
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Turn off notifications for the Google Maps app<image>Thought: Navigate back to access the app notification settings for Google Maps.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [43, 62]}<image>Thought: Click the arrow backward to navigate back to the previous menu where app notifications can be adjusted.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [46, 64]}<image>",
"role": "user"
},
{
"content": "Thought: You would have accessed the Settings menu to find notification options. You likely scrolled down to locate the Notifications section. By selecting Notifications, you can manage Google Maps app notifications effectively. Once completed, you indicate that the task is complete by confirming changes made to notifications.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9037664210789009586_15.png",
"training_data/AITW/aitw_images/general/9037664210789009586_16.png",
"training_data/AITW/aitw_images/general/9037664210789009586_17.png"
] | 1,120
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Sweden?<image>",
"role": "user"
},
{
"content": "Thought: Press home to exit the email composition interface and return to the main screen for further navigation.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/2578257002013705924_0.png"
] | 1,121
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Sweden?<image>Thought: Press home to exit the email composition interface and return to the main screen for further navigation.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click PLAY to engage a voice assistant or an app that can provide information about Sweden's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 1153]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/2578257002013705924_0.png",
"training_data/AITW/aitw_images/general/2578257002013705924_1.png"
] | 1,122
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Sweden?<image>Thought: Press home to exit the email composition interface and return to the main screen for further navigation.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click PLAY to engage a voice assistant or an app that can provide information about Sweden's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 1153]}<image>",
"role": "user"
},
{
"content": "Thought: The user clicked on google.com to search for information, but appears to have typed only \"fl\" instead of the full query. The search results displayed are unrelated to the capital of Sweden and focus on flight schedules. Therefore, it seems that the user's initial input may have been an incomplete thought or a mistake in querying for \"capital of Sweden.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [365, 98]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/2578257002013705924_0.png",
"training_data/AITW/aitw_images/general/2578257002013705924_1.png",
"training_data/AITW/aitw_images/general/2578257002013705924_2.png"
] | 1,123
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Sweden?<image>Thought: Click PLAY to engage a voice assistant or an app that can provide information about Sweden's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 1153]}<image>Thought: The user clicked on google.com to search for information, but appears to have typed only \"fl\" instead of the full query. The search results displayed are unrelated to the capital of Sweden and focus on flight schedules. Therefore, it seems that the user's initial input may have been an incomplete thought or a mistake in querying for \"capital of Sweden.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [365, 98]}<image>",
"role": "user"
},
{
"content": "Thought: Type in the search bar to ask the question about Sweden's capital.\nAction: {\"action\": \"TYPE\", \"value\": \"What is the capital of Sweden?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/2578257002013705924_1.png",
"training_data/AITW/aitw_images/general/2578257002013705924_2.png",
"training_data/AITW/aitw_images/general/2578257002013705924_3.png"
] | 1,124
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Sweden?<image>Thought: The user clicked on google.com to search for information, but appears to have typed only \"fl\" instead of the full query. The search results displayed are unrelated to the capital of Sweden and focus on flight schedules. Therefore, it seems that the user's initial input may have been an incomplete thought or a mistake in querying for \"capital of Sweden.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [365, 98]}<image>Thought: Type in the search bar to ask the question about Sweden's capital.\nAction: {\"action\": \"TYPE\", \"value\": \"What is the capital of Sweden?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"is\" to complete the question in the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [222, 197]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/2578257002013705924_2.png",
"training_data/AITW/aitw_images/general/2578257002013705924_3.png",
"training_data/AITW/aitw_images/general/2578257002013705924_4.png"
] | 1,125
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Sweden?<image>Thought: Type in the search bar to ask the question about Sweden's capital.\nAction: {\"action\": \"TYPE\", \"value\": \"What is the capital of Sweden?\", \"position\": null}<image>Thought: Click on \"is\" to complete the question in the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [222, 197]}<image>",
"role": "user"
},
{
"content": "Thought: You complete the task by confirming that the capital of Sweden is displayed as \"Stockholm\" on the screen.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/2578257002013705924_3.png",
"training_data/AITW/aitw_images/general/2578257002013705924_4.png",
"training_data/AITW/aitw_images/general/2578257002013705924_5.png"
] | 1,126
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Mexico city<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to reveal more options for searching hotels.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8356723694870216517_0.png"
] | 1,127
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Mexico city<image>Thought: Scroll down to reveal more options for searching hotels.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click GOOGLE to open the search engine for finding hotels.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [279, 628]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8356723694870216517_0.png",
"training_data/AITW/aitw_images/general/8356723694870216517_1.png"
] | 1,128
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Mexico city<image>Thought: Scroll down to reveal more options for searching hotels.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click GOOGLE to open the search engine for finding hotels.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [279, 628]}<image>",
"role": "user"
},
{
"content": "Thought: Click Search to initiate the query for hotels in Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [218, 315]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8356723694870216517_0.png",
"training_data/AITW/aitw_images/general/8356723694870216517_1.png",
"training_data/AITW/aitw_images/general/8356723694870216517_2.png"
] | 1,129
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Mexico city<image>Thought: Click GOOGLE to open the search engine for finding hotels.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [279, 628]}<image>Thought: Click Search to initiate the query for hotels in Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [218, 315]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"Search for hotels in Mexico city\" in the search bar to initiate a hotel search.\nAction: {\"action\": \"TYPE\", \"value\": \"Search for hotels in Mexico city\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8356723694870216517_1.png",
"training_data/AITW/aitw_images/general/8356723694870216517_2.png",
"training_data/AITW/aitw_images/general/8356723694870216517_3.png"
] | 1,130
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Mexico city<image>Thought: Click Search to initiate the query for hotels in Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [218, 315]}<image>Thought: Type \"Search for hotels in Mexico city\" in the search bar to initiate a hotel search.\nAction: {\"action\": \"TYPE\", \"value\": \"Search for hotels in Mexico city\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the suggestion \"...ity mexico\" to finalize the search for hotels in Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [227, 207]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8356723694870216517_2.png",
"training_data/AITW/aitw_images/general/8356723694870216517_3.png",
"training_data/AITW/aitw_images/general/8356723694870216517_4.png"
] | 1,131
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for hotels in Mexico city<image>Thought: Type \"Search for hotels in Mexico city\" in the search bar to initiate a hotel search.\nAction: {\"action\": \"TYPE\", \"value\": \"Search for hotels in Mexico city\", \"position\": null}<image>Thought: Click on the suggestion \"...ity mexico\" to finalize the search for hotels in Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [227, 207]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully completed your search for hotels in Mexico City using the Google search bar.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8356723694870216517_3.png",
"training_data/AITW/aitw_images/general/8356723694870216517_4.png",
"training_data/AITW/aitw_images/general/8356723694870216517_5.png"
] | 1,132
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>",
"role": "user"
},
{
"content": "Thought: Press home to access the main screen and open the app for watching the latest BBC video.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_0.png"
] | 1,133
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Press home to access the main screen and open the app for watching the latest BBC video.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the CHAT icon to access the latest video content from BBC.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [197, 608]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_0.png",
"training_data/AITW/aitw_images/general/16736038483341010783_1.png"
] | 1,134
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Press home to access the main screen and open the app for watching the latest BBC video.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click on the CHAT icon to access the latest video content from BBC.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [197, 608]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"Play the latest video from the BBC\" to search for the content in the Google search bar.\nAction: {\"action\": \"TYPE\", \"value\": \"Play the latest video from the BBC\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_0.png",
"training_data/AITW/aitw_images/general/16736038483341010783_1.png",
"training_data/AITW/aitw_images/general/16736038483341010783_2.png"
] | 1,135
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Click on the CHAT icon to access the latest video content from BBC.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [197, 608]}<image>Thought: Type \"Play the latest video from the BBC\" to search for the content in the Google search bar.\nAction: {\"action\": \"TYPE\", \"value\": \"Play the latest video from the BBC\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to submit the search query for the latest video from BBC.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_1.png",
"training_data/AITW/aitw_images/general/16736038483341010783_2.png",
"training_data/AITW/aitw_images/general/16736038483341010783_3.png"
] | 1,136
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Type \"Play the latest video from the BBC\" to search for the content in the Google search bar.\nAction: {\"action\": \"TYPE\", \"value\": \"Play the latest video from the BBC\", \"position\": null}<image>Thought: Press enter to submit the search query for the latest video from BBC.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more content and find the latest BBC video among the search results.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_2.png",
"training_data/AITW/aitw_images/general/16736038483341010783_3.png",
"training_data/AITW/aitw_images/general/16736038483341010783_4.png"
] | 1,137
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Press enter to submit the search query for the latest video from BBC.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>Thought: Scroll down to view more content and find the latest BBC video among the search results.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click PLAY to start watching the latest video from the BBC displayed in the YouTube app.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [183, 170]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_3.png",
"training_data/AITW/aitw_images/general/16736038483341010783_4.png",
"training_data/AITW/aitw_images/general/16736038483341010783_5.png"
] | 1,138
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Scroll down to view more content and find the latest BBC video among the search results.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click PLAY to start watching the latest video from the BBC displayed in the YouTube app.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [183, 170]}<image>",
"role": "user"
},
{
"content": "Thought: Click the UPDATE button to ensure the app is up to date for playing the latest video from BBC.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 577]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_4.png",
"training_data/AITW/aitw_images/general/16736038483341010783_5.png",
"training_data/AITW/aitw_images/general/16736038483341010783_6.png"
] | 1,139
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Click PLAY to start watching the latest video from the BBC displayed in the YouTube app.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [183, 170]}<image>Thought: Click the UPDATE button to ensure the app is up to date for playing the latest video from BBC.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 577]}<image>",
"role": "user"
},
{
"content": "Thought: Choose Google Play Store from the options to access the latest video content.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [162, 488]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_5.png",
"training_data/AITW/aitw_images/general/16736038483341010783_6.png",
"training_data/AITW/aitw_images/general/16736038483341010783_7.png"
] | 1,140
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Click the UPDATE button to ensure the app is up to date for playing the latest video from BBC.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 577]}<image>Thought: Choose Google Play Store from the options to access the latest video content.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [162, 488]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Just once\" to select how to open the video link for viewing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [213, 635]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_6.png",
"training_data/AITW/aitw_images/general/16736038483341010783_7.png",
"training_data/AITW/aitw_images/general/16736038483341010783_8.png"
] | 1,141
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Choose Google Play Store from the options to access the latest video content.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [162, 488]}<image>Thought: Click on \"Just once\" to select how to open the video link for viewing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [213, 635]}<image>",
"role": "user"
},
{
"content": "Thought: Click the Update button to ensure YouTube has the latest features for viewing videos.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [146, 285]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_7.png",
"training_data/AITW/aitw_images/general/16736038483341010783_8.png",
"training_data/AITW/aitw_images/general/16736038483341010783_9.png"
] | 1,142
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Click on \"Just once\" to select how to open the video link for viewing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [213, 635]}<image>Thought: Click the Update button to ensure YouTube has the latest features for viewing videos.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [146, 285]}<image>",
"role": "user"
},
{
"content": "Thought: Click Open to launch the YouTube app and access the latest video.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [303, 274]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_8.png",
"training_data/AITW/aitw_images/general/16736038483341010783_9.png",
"training_data/AITW/aitw_images/general/16736038483341010783_10.png"
] | 1,143
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Play the latest video from the BBC<image>Thought: Click the Update button to ensure YouTube has the latest features for viewing videos.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [146, 285]}<image>Thought: Click Open to launch the YouTube app and access the latest video.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [303, 274]}<image>",
"role": "user"
},
{
"content": "Thought: The user successfully completed their task by finding and playing the latest BBC video on YouTube.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/16736038483341010783_9.png",
"training_data/AITW/aitw_images/general/16736038483341010783_10.png",
"training_data/AITW/aitw_images/general/16736038483341010783_11.png"
] | 1,144
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Google the capital of Argentina<image>",
"role": "user"
},
{
"content": "Thought: Click the \"Search or type web address\" bar to begin the Google search for Argentina's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [369, 303]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9846444028928426439_0.png"
] | 1,145
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Google the capital of Argentina<image>Thought: Click the \"Search or type web address\" bar to begin the Google search for Argentina's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [369, 303]}<image>",
"role": "user"
},
{
"content": "Thought: Type the search query into the search bar to initiate a Google search for information.\nAction: {\"action\": \"TYPE\", \"value\": \"Google the capital of Argentina\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9846444028928426439_0.png",
"training_data/AITW/aitw_images/general/9846444028928426439_1.png"
] | 1,146
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Google the capital of Argentina<image>Thought: Click the \"Search or type web address\" bar to begin the Google search for Argentina's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [369, 303]}<image>Thought: Type the search query into the search bar to initiate a Google search for information.\nAction: {\"action\": \"TYPE\", \"value\": \"Google the capital of Argentina\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on 'Argentina' to select this search suggestion and proceed with the query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [249, 134]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9846444028928426439_0.png",
"training_data/AITW/aitw_images/general/9846444028928426439_1.png",
"training_data/AITW/aitw_images/general/9846444028928426439_2.png"
] | 1,147
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Google the capital of Argentina<image>Thought: Type the search query into the search bar to initiate a Google search for information.\nAction: {\"action\": \"TYPE\", \"value\": \"Google the capital of Argentina\", \"position\": null}<image>Thought: Click on 'Argentina' to select this search suggestion and proceed with the query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [249, 134]}<image>",
"role": "user"
},
{
"content": "Thought: You conducted a search for the capital of Argentina using Google. Upon receiving the search results, you confirmed that Buenos Aires is indeed the capital.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9846444028928426439_1.png",
"training_data/AITW/aitw_images/general/9846444028928426439_2.png",
"training_data/AITW/aitw_images/general/9846444028928426439_3.png"
] | 1,148
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen and access the necessary apps for searching apartment rentals.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_0.png"
] | 1,149
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Press home to return to the main screen and access the necessary apps for searching apartment rentals.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to find more information, likely related to apartment rentals in Los Angeles.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_0.png",
"training_data/AITW/aitw_images/general/1233143297051599737_1.png"
] | 1,150
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Press home to return to the main screen and access the necessary apps for searching apartment rentals.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Scroll down to find more information, likely related to apartment rentals in Los Angeles.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the Google app to start searching for apartment rental prices in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 744]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_0.png",
"training_data/AITW/aitw_images/general/1233143297051599737_1.png",
"training_data/AITW/aitw_images/general/1233143297051599737_2.png"
] | 1,151
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Scroll down to find more information, likely related to apartment rentals in Los Angeles.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click the Google app to start searching for apartment rental prices in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 744]}<image>",
"role": "user"
},
{
"content": "Thought: Click GOOGLE to access the search engine for information on apartment rents in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [496, 698]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_1.png",
"training_data/AITW/aitw_images/general/1233143297051599737_2.png",
"training_data/AITW/aitw_images/general/1233143297051599737_3.png"
] | 1,152
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Click the Google app to start searching for apartment rental prices in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 744]}<image>Thought: Click GOOGLE to access the search engine for information on apartment rents in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [496, 698]}<image>",
"role": "user"
},
{
"content": "Thought: Status task impossible, as the current search for blush is unrelated to renting apartments in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [509, 196]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_2.png",
"training_data/AITW/aitw_images/general/1233143297051599737_3.png",
"training_data/AITW/aitw_images/general/1233143297051599737_4.png"
] | 1,153
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Click GOOGLE to access the search engine for information on apartment rents in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [496, 698]}<image>Thought: Status task impossible, as the current search for blush is unrelated to renting apartments in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [509, 196]}<image>",
"role": "user"
},
{
"content": "Thought: The user clicks the X to clear the previous search query for a new task.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [644, 114]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_3.png",
"training_data/AITW/aitw_images/general/1233143297051599737_4.png",
"training_data/AITW/aitw_images/general/1233143297051599737_5.png"
] | 1,154
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Status task impossible, as the current search for blush is unrelated to renting apartments in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [509, 196]}<image>Thought: The user clicks the X to clear the previous search query for a new task.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [644, 114]}<image>",
"role": "user"
},
{
"content": "Thought: Type the question in the search bar to find information about 2 bedroom apartment rents.\nAction: {\"action\": \"TYPE\", \"value\": \"How much does a 2 bedroom apartment rent for in Los Angeles?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_4.png",
"training_data/AITW/aitw_images/general/1233143297051599737_5.png",
"training_data/AITW/aitw_images/general/1233143297051599737_6.png"
] | 1,155
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: The user clicks the X to clear the previous search query for a new task.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [644, 114]}<image>Thought: Type the question in the search bar to find information about 2 bedroom apartment rents.\nAction: {\"action\": \"TYPE\", \"value\": \"How much does a 2 bedroom apartment rent for in Los Angeles?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You clicked \"Search...\" to initiate a query about two-bedroom apartment rents in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [232, 96]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_5.png",
"training_data/AITW/aitw_images/general/1233143297051599737_6.png",
"training_data/AITW/aitw_images/general/1233143297051599737_7.png"
] | 1,156
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Type the question in the search bar to find information about 2 bedroom apartment rents.\nAction: {\"action\": \"TYPE\", \"value\": \"How much does a 2 bedroom apartment rent for in Los Angeles?\", \"position\": null}<image>Thought: You clicked \"Search...\" to initiate a query about two-bedroom apartment rents in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [232, 96]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"loes\" to enter the query for apartment rent in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [176, 95]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_6.png",
"training_data/AITW/aitw_images/general/1233143297051599737_7.png",
"training_data/AITW/aitw_images/general/1233143297051599737_8.png"
] | 1,157
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: You clicked \"Search...\" to initiate a query about two-bedroom apartment rents in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [232, 96]}<image>Thought: Click on \"loes\" to enter the query for apartment rent in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [176, 95]}<image>",
"role": "user"
},
{
"content": "Thought: You would press enter to submit your search query for the 2 bedroom apartment rent in Los Angeles.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_7.png",
"training_data/AITW/aitw_images/general/1233143297051599737_8.png",
"training_data/AITW/aitw_images/general/1233143297051599737_9.png"
] | 1,158
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: How much does a 2 bedroom apartment rent for in Los Angeles?<image>Thought: Click on \"loes\" to enter the query for apartment rent in Los Angeles.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [176, 95]}<image>Thought: You would press enter to submit your search query for the 2 bedroom apartment rent in Los Angeles.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: The search query shows relevant information about rental costs for a 2 bedroom apartment, confirming the task is complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1233143297051599737_8.png",
"training_data/AITW/aitw_images/general/1233143297051599737_9.png",
"training_data/AITW/aitw_images/general/1233143297051599737_10.png"
] | 1,159
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Open a new incognito window in Chrome<image>",
"role": "user"
},
{
"content": "Thought: Click the Chrome icon to launch the browser for incognito browsing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [362, 836]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/6814581428117316762_0.png"
] | 1,160
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Open a new incognito window in Chrome<image>Thought: Click the Chrome icon to launch the browser for incognito browsing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [362, 836]}<image>",
"role": "user"
},
{
"content": "Thought: Click the THREE DOTS to access the Chrome menu for incognito options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [496, 71]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/6814581428117316762_0.png",
"training_data/AITW/aitw_images/general/6814581428117316762_1.png"
] | 1,161
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Open a new incognito window in Chrome<image>Thought: Click the Chrome icon to launch the browser for incognito browsing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [362, 836]}<image>Thought: Click the THREE DOTS to access the Chrome menu for incognito options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [496, 71]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"New incognito tab\" to open a private browsing session.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [384, 198]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/6814581428117316762_0.png",
"training_data/AITW/aitw_images/general/6814581428117316762_1.png",
"training_data/AITW/aitw_images/general/6814581428117316762_2.png"
] | 1,162
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Open a new incognito window in Chrome<image>Thought: Click the THREE DOTS to access the Chrome menu for incognito options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [496, 71]}<image>Thought: Click on \"New incognito tab\" to open a private browsing session.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [384, 198]}<image>",
"role": "user"
},
{
"content": "Thought: You would open a new incognito window in Chrome to browse privately. You would expect to see confirmation that you've gone incognito after completing the action. The initial GUI indicates successful entry into incognito mode with specific information displayed.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/6814581428117316762_1.png",
"training_data/AITW/aitw_images/general/6814581428117316762_2.png",
"training_data/AITW/aitw_images/general/6814581428117316762_3.png"
] | 1,163
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for a new blush on Sephora<image>",
"role": "user"
},
{
"content": "Thought: Click on the Google Chrome icon to start browsing for Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [435, 1246]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/18046781092504945893_0.png"
] | 1,164
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for a new blush on Sephora<image>Thought: Click on the Google Chrome icon to start browsing for Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [435, 1246]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the News section to check for any recent updates or promotions related to blush on Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 266]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/18046781092504945893_0.png",
"training_data/AITW/aitw_images/general/18046781092504945893_1.png"
] | 1,165
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for a new blush on Sephora<image>Thought: Click on the Google Chrome icon to start browsing for Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [435, 1246]}<image>Thought: Click on the News section to check for any recent updates or promotions related to blush on Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 266]}<image>",
"role": "user"
},
{
"content": "Thought: Click the X to clear the previous search query about rockets.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [654, 182]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/18046781092504945893_0.png",
"training_data/AITW/aitw_images/general/18046781092504945893_1.png",
"training_data/AITW/aitw_images/general/18046781092504945893_2.png"
] | 1,166
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for a new blush on Sephora<image>Thought: Click on the News section to check for any recent updates or promotions related to blush on Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [492, 266]}<image>Thought: Click the X to clear the previous search query about rockets.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [654, 182]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"new blush on Sephora\" in the search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"new blush on Sephora\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/18046781092504945893_1.png",
"training_data/AITW/aitw_images/general/18046781092504945893_2.png",
"training_data/AITW/aitw_images/general/18046781092504945893_3.png"
] | 1,167
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for a new blush on Sephora<image>Thought: Click the X to clear the previous search query about rockets.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [654, 182]}<image>Thought: Type \"new blush on Sephora\" in the search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"new blush on Sephora\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You want to see the search results for \"new blush on Sephora,\" so you press enter.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/18046781092504945893_2.png",
"training_data/AITW/aitw_images/general/18046781092504945893_3.png",
"training_data/AITW/aitw_images/general/18046781092504945893_4.png"
] | 1,168
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for a new blush on Sephora<image>Thought: Type \"new blush on Sephora\" in the search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"new blush on Sephora\", \"position\": null}<image>Thought: You want to see the search results for \"new blush on Sephora,\" so you press enter.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You searched for \"new blush on Sephora\" in the Google search bar, and the results indicated that you successfully completed your task.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/18046781092504945893_3.png",
"training_data/AITW/aitw_images/general/18046781092504945893_4.png",
"training_data/AITW/aitw_images/general/18046781092504945893_5.png"
] | 1,169
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen and access apps for news about space exploration.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_0.png"
] | 1,170
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>Thought: Press home to return to the main screen and access apps for news about space exploration.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the NAV BAR CIRCLE to access the main menu and find news apps for space exploration.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [335, 1039]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_0.png",
"training_data/AITW/aitw_images/general/17116418152226965165_1.png"
] | 1,171
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>Thought: Press home to return to the main screen and access apps for news about space exploration.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click the NAV BAR CIRCLE to access the main menu and find news apps for space exploration.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [335, 1039]}<image>",
"role": "user"
},
{
"content": "Thought: Type in the search bar to query the latest news in space exploration.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the latest news in space exploration\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_0.png",
"training_data/AITW/aitw_images/general/17116418152226965165_1.png",
"training_data/AITW/aitw_images/general/17116418152226965165_2.png"
] | 1,172
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>Thought: Click the NAV BAR CIRCLE to access the main menu and find news apps for space exploration.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [335, 1039]}<image>Thought: Type in the search bar to query the latest news in space exploration.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the latest news in space exploration\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to submit the search query in the Google search bar for space exploration news.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_1.png",
"training_data/AITW/aitw_images/general/17116418152226965165_2.png",
"training_data/AITW/aitw_images/general/17116418152226965165_3.png"
] | 1,173
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>Thought: Type in the search bar to query the latest news in space exploration.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the latest news in space exploration\", \"position\": null}<image>Thought: Press enter to submit the search query in the Google search bar for space exploration news.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Exploration\" to access relevant news about space exploration directly from the search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [335, 322]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_2.png",
"training_data/AITW/aitw_images/general/17116418152226965165_3.png",
"training_data/AITW/aitw_images/general/17116418152226965165_4.png"
] | 1,174
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>Thought: Press enter to submit the search query in the Google search bar for space exploration news.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>Thought: Click on \"Exploration\" to access relevant news about space exploration directly from the search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [335, 322]}<image>",
"role": "user"
},
{
"content": "Thought: Click the OPTIONS button to explore additional preferences regarding personal data processing before accessing news.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [470, 847]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_3.png",
"training_data/AITW/aitw_images/general/17116418152226965165_4.png",
"training_data/AITW/aitw_images/general/17116418152226965165_5.png"
] | 1,175
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What's the latest news in space exploration?<image>Thought: Click on \"Exploration\" to access relevant news about space exploration directly from the search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [335, 322]}<image>Thought: Click the OPTIONS button to explore additional preferences regarding personal data processing before accessing news.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [470, 847]}<image>",
"role": "user"
},
{
"content": "Thought: You found the latest news articles on space exploration in the 'LATEST NEWS' section of the website.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17116418152226965165_4.png",
"training_data/AITW/aitw_images/general/17116418152226965165_5.png",
"training_data/AITW/aitw_images/general/17116418152226965165_6.png"
] | 1,176
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Italy?<image>",
"role": "user"
},
{
"content": "Thought: Click LOCATION to access a map or search feature that can provide information about the capital of Italy.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [170, 473]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/7408959087407856867_0.png"
] | 1,177
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Italy?<image>Thought: Click LOCATION to access a map or search feature that can provide information about the capital of Italy.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [170, 473]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the Google search bar to enter a query about the capital of Italy.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 29]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/7408959087407856867_0.png",
"training_data/AITW/aitw_images/general/7408959087407856867_1.png"
] | 1,178
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Italy?<image>Thought: Click LOCATION to access a map or search feature that can provide information about the capital of Italy.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [170, 473]}<image>Thought: Click on the Google search bar to enter a query about the capital of Italy.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 29]}<image>",
"role": "user"
},
{
"content": "Thought: Type in the search bar to find information about the capital of Italy.\nAction: {\"action\": \"TYPE\", \"value\": \"What is the capital of Italy?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/7408959087407856867_0.png",
"training_data/AITW/aitw_images/general/7408959087407856867_1.png",
"training_data/AITW/aitw_images/general/7408959087407856867_2.png"
] | 1,179
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Italy?<image>Thought: Click on the Google search bar to enter a query about the capital of Italy.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 29]}<image>Thought: Type in the search bar to find information about the capital of Italy.\nAction: {\"action\": \"TYPE\", \"value\": \"What is the capital of Italy?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the search bar to enter the query for finding Italy's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 67]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/7408959087407856867_1.png",
"training_data/AITW/aitw_images/general/7408959087407856867_2.png",
"training_data/AITW/aitw_images/general/7408959087407856867_3.png"
] | 1,180
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: What is the capital of Italy?<image>Thought: Type in the search bar to find information about the capital of Italy.\nAction: {\"action\": \"TYPE\", \"value\": \"What is the capital of Italy?\", \"position\": null}<image>Thought: Click the search bar to enter the query for finding Italy's capital.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 67]}<image>",
"role": "user"
},
{
"content": "Thought: Recognize the displayed information indicating that Rome is the capital of Italy. Conclude that I have obtained the required answer.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/7408959087407856867_2.png",
"training_data/AITW/aitw_images/general/7408959087407856867_3.png",
"training_data/AITW/aitw_images/general/7408959087407856867_4.png"
] | 1,181
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>",
"role": "user"
},
{
"content": "Thought: Click the ENVELOPE to access email for flight information and deals.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [471, 1188]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_0.png"
] | 1,182
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the ENVELOPE to access email for flight information and deals.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [471, 1188]}<image>",
"role": "user"
},
{
"content": "Thought: Press back to return from the current screen, as no flights were found yet.\nAction: {\"action\": \"PRESS BACK\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_0.png",
"training_data/AITW/aitw_images/general/10716603362934555615_1.png"
] | 1,183
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the ENVELOPE to access email for flight information and deals.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [471, 1188]}<image>Thought: Press back to return from the current screen, as no flights were found yet.\nAction: {\"action\": \"PRESS BACK\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen and access the browser for flight searching.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_0.png",
"training_data/AITW/aitw_images/general/10716603362934555615_1.png",
"training_data/AITW/aitw_images/general/10716603362934555615_2.png"
] | 1,184
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Press back to return from the current screen, as no flights were found yet.\nAction: {\"action\": \"PRESS BACK\", \"value\": null, \"position\": null}<image>Thought: Press home to return to the main screen and access the browser for flight searching.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the ENVELOPE icon to access my email, possibly looking for flight confirmations or notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 1199]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_1.png",
"training_data/AITW/aitw_images/general/10716603362934555615_2.png",
"training_data/AITW/aitw_images/general/10716603362934555615_3.png"
] | 1,185
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Press home to return to the main screen and access the browser for flight searching.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click the ENVELOPE icon to access my email, possibly looking for flight confirmations or notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 1199]}<image>",
"role": "user"
},
{
"content": "Thought: Click the Chrome icon to open the browser for searching flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [536, 1137]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_2.png",
"training_data/AITW/aitw_images/general/10716603362934555615_3.png",
"training_data/AITW/aitw_images/general/10716603362934555615_4.png"
] | 1,186
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the ENVELOPE icon to access my email, possibly looking for flight confirmations or notifications.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 1199]}<image>Thought: Click the Chrome icon to open the browser for searching flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [536, 1137]}<image>",
"role": "user"
},
{
"content": "Thought: Press back to return to the previous screen and possibly explore other options for flight search.\nAction: {\"action\": \"PRESS BACK\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_3.png",
"training_data/AITW/aitw_images/general/10716603362934555615_4.png",
"training_data/AITW/aitw_images/general/10716603362934555615_5.png"
] | 1,187
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the Chrome icon to open the browser for searching flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [536, 1137]}<image>Thought: Press back to return to the previous screen and possibly explore other options for flight search.\nAction: {\"action\": \"PRESS BACK\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You want to go back to the home screen to start a new task.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_4.png",
"training_data/AITW/aitw_images/general/10716603362934555615_5.png",
"training_data/AITW/aitw_images/general/10716603362934555615_6.png"
] | 1,188
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Press back to return to the previous screen and possibly explore other options for flight search.\nAction: {\"action\": \"PRESS BACK\", \"value\": null, \"position\": null}<image>Thought: You want to go back to the home screen to start a new task.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the ENVELOPE icon to access the email app for flight confirmations or inquiries.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 1175]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_5.png",
"training_data/AITW/aitw_images/general/10716603362934555615_6.png",
"training_data/AITW/aitw_images/general/10716603362934555615_7.png"
] | 1,189
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: You want to go back to the home screen to start a new task.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click the ENVELOPE icon to access the email app for flight confirmations or inquiries.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 1175]}<image>",
"role": "user"
},
{
"content": "Thought: Click the “New tab” option to open a browser for flight searches.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [480, 1163]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_6.png",
"training_data/AITW/aitw_images/general/10716603362934555615_7.png",
"training_data/AITW/aitw_images/general/10716603362934555615_8.png"
] | 1,190
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the ENVELOPE icon to access the email app for flight confirmations or inquiries.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 1175]}<image>Thought: Click the “New tab” option to open a browser for flight searches.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [480, 1163]}<image>",
"role": "user"
},
{
"content": "Thought: Click on the Shopping option to access flight deals and information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [408, 311]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_7.png",
"training_data/AITW/aitw_images/general/10716603362934555615_8.png",
"training_data/AITW/aitw_images/general/10716603362934555615_9.png"
] | 1,191
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the “New tab” option to open a browser for flight searches.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [480, 1163]}<image>Thought: Click on the Shopping option to access flight deals and information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [408, 311]}<image>",
"role": "user"
},
{
"content": "Thought: Click the X to clear the current search input for a new flight query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [661, 188]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_8.png",
"training_data/AITW/aitw_images/general/10716603362934555615_9.png",
"training_data/AITW/aitw_images/general/10716603362934555615_10.png"
] | 1,192
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click on the Shopping option to access flight deals and information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [408, 311]}<image>Thought: Click the X to clear the current search input for a new flight query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [661, 188]}<image>",
"role": "user"
},
{
"content": "Thought: Type the flight search query in the search bar to find relevant flights from NYC to San Diego.\nAction: {\"action\": \"TYPE\", \"value\": \"flight from nyc to san diego\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_9.png",
"training_data/AITW/aitw_images/general/10716603362934555615_10.png",
"training_data/AITW/aitw_images/general/10716603362934555615_11.png"
] | 1,193
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the X to clear the current search input for a new flight query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [661, 188]}<image>Thought: Type the flight search query in the search bar to find relevant flights from NYC to San Diego.\nAction: {\"action\": \"TYPE\", \"value\": \"flight from nyc to san diego\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the suggestion \"flight from nyc to san diego\" to perform the flight search. This selection directly leads to relevant flight options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [413, 280]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_10.png",
"training_data/AITW/aitw_images/general/10716603362934555615_11.png",
"training_data/AITW/aitw_images/general/10716603362934555615_12.png"
] | 1,194
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Type the flight search query in the search bar to find relevant flights from NYC to San Diego.\nAction: {\"action\": \"TYPE\", \"value\": \"flight from nyc to san diego\", \"position\": null}<image>Thought: Click the suggestion \"flight from nyc to san diego\" to perform the flight search. This selection directly leads to relevant flight options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [413, 280]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Flights from\" to access specific flight options for NYC to San Diego.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [188, 537]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_11.png",
"training_data/AITW/aitw_images/general/10716603362934555615_12.png",
"training_data/AITW/aitw_images/general/10716603362934555615_13.png"
] | 1,195
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the suggestion \"flight from nyc to san diego\" to perform the flight search. This selection directly leads to relevant flight options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [413, 280]}<image>Thought: Click on \"Flights from\" to access specific flight options for NYC to San Diego.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [188, 537]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more flight options and details for the NYC to San Diego search.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_12.png",
"training_data/AITW/aitw_images/general/10716603362934555615_13.png",
"training_data/AITW/aitw_images/general/10716603362934555615_14.png"
] | 1,196
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click on \"Flights from\" to access specific flight options for NYC to San Diego.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [188, 537]}<image>Thought: Scroll down to view more flight options and details for the NYC to San Diego search.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the \"Additional terms apply\" to view important flight information and conditions before finalizing the search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [427, 1131]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_13.png",
"training_data/AITW/aitw_images/general/10716603362934555615_14.png",
"training_data/AITW/aitw_images/general/10716603362934555615_15.png"
] | 1,197
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Scroll down to view more flight options and details for the NYC to San Diego search.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click the \"Additional terms apply\" to view important flight information and conditions before finalizing the search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [427, 1131]}<image>",
"role": "user"
},
{
"content": "Thought: Click Search to initiate the flight search process based on the inputted details for NYC to San Diego.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [443, 842]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_14.png",
"training_data/AITW/aitw_images/general/10716603362934555615_15.png",
"training_data/AITW/aitw_images/general/10716603362934555615_16.png"
] | 1,198
|
[
{
"content": "You are an assistant trained to navigate the mobile phone. \nGiven a task instruction, a screen observation, and an action history sequence, \noutput the next action and wait for the next observation. \nHere is the action space:\n1. `CLICK`: Click on an element, value is not applicable and the position [x,y] is required. \n2. `TYPE`: Type a string into an element, value is a string to type and the position is not applicable.\n3. `SELECT`: Select a value for an element, value is the value to select and the position is not applicable.\n4. `SCROLL UP`: Scroll up for the screen.\n5. `SCROLL DOWN`: Scroll down for the screen.\n6. `SCROLL LEFT`: Scroll left for the screen.\n7. `SCROLL RIGHT`: Scroll right for the screen.\n8. `PRESS BACK`: Press for returning to the previous step, value and position are not applicable.\n9. `PRESS HOME`: Press for returning to the home screen, value and position are not applicable.\n10. `PRESS ENTER`: Press for submitting the input content, value and position are not applicable.\n11. `STATUS TASK COMPLETE`: Indicate the task is completed, value and position are not applicable.\n12. `STATUS TASK IMPOSSIBLE `: Indicate the task is impossible to complete, value and position are not applicable.\n\nFormat your response as\nThought: <your reasoning process>\nAction: <the next action>\n\nFormat the action as a JSON object with the following keys:\n{\"action\": \"ACTION_TYPE\", \"value\": \"element\", \"position\": [x,y]}\n\nIf value or position is not applicable, set it as `None`.\nPosition represents the absolute pixel coordinates on the screenshot.\nTask: Search for flights from NYC to San Diego<image>Thought: Click the \"Additional terms apply\" to view important flight information and conditions before finalizing the search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [427, 1131]}<image>Thought: Click Search to initiate the flight search process based on the inputted details for NYC to San Diego.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [443, 842]}<image>",
"role": "user"
},
{
"content": "Thought: The task is marked complete after successfully searching for flights from NYC to San Diego. The user likely found the desired flight information displayed on the screen, allowing them to consider their options.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/10716603362934555615_15.png",
"training_data/AITW/aitw_images/general/10716603362934555615_16.png",
"training_data/AITW/aitw_images/general/10716603362934555615_17.png"
] | 1,199
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.