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: Open the gallery<image>Thought: Click on the Photos icon to access the gallery app directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [208, 169]}<image>Thought: You click \"Update now\" because the current version of Google Photos is not usable.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [281, 423]}<image>",
"role": "user"
},
{
"content": "Thought: Click 'Update' to ensure the Google Photos app is the latest version before opening the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [157, 318]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9668891882429128741_1.png",
"training_data/AITW/aitw_images/general/9668891882429128741_2.png",
"training_data/AITW/aitw_images/general/9668891882429128741_3.png"
] | 2,900
|
[
{
"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 the gallery<image>Thought: You click \"Update now\" because the current version of Google Photos is not usable.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [281, 423]}<image>Thought: Click 'Update' to ensure the Google Photos app is the latest version before opening the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [157, 318]}<image>",
"role": "user"
},
{
"content": "Thought: Click Open to launch the Google Photos app, which allows access to the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [311, 313]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9668891882429128741_2.png",
"training_data/AITW/aitw_images/general/9668891882429128741_3.png",
"training_data/AITW/aitw_images/general/9668891882429128741_4.png"
] | 2,901
|
[
{
"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 the gallery<image>Thought: Click 'Update' to ensure the Google Photos app is the latest version before opening the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [157, 318]}<image>Thought: Click Open to launch the Google Photos app, which allows access to the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [311, 313]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Do not back up\" to proceed without activating backup settings, allowing access to the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [146, 589]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9668891882429128741_3.png",
"training_data/AITW/aitw_images/general/9668891882429128741_4.png",
"training_data/AITW/aitw_images/general/9668891882429128741_5.png"
] | 2,902
|
[
{
"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 the gallery<image>Thought: Click Open to launch the Google Photos app, which allows access to the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [311, 313]}<image>Thought: Click on \"Do not back up\" to proceed without activating backup settings, allowing access to the gallery.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [146, 589]}<image>",
"role": "user"
},
{
"content": "Thought: You accessed Google Photos, which serves as a gallery app for viewing images and videos. The task of opening the gallery was achieved by successfully navigating to the Photos section.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/9668891882429128741_4.png",
"training_data/AITW/aitw_images/general/9668891882429128741_5.png",
"training_data/AITW/aitw_images/general/9668891882429128741_6.png"
] | 2,903
|
[
{
"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 news in Pakistan?<image>",
"role": "user"
},
{
"content": "Thought: Click the Chrome icon to navigate to the browser for checking news in Pakistan.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [367, 895]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/3135337687959494817_0.png"
] | 2,904
|
[
{
"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 news in Pakistan?<image>Thought: Click the Chrome icon to navigate to the browser for checking news in Pakistan.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [367, 895]}<image>",
"role": "user"
},
{
"content": "Thought: Realize the task is to find news in Pakistan, so first search for relevant content. Recognize that searching for hotels is not related to news. Notice the Google search bar is used incorrectly for this task. Adjust the query in the search bar to include 'news in Pakistan' instead of 'Search for hotels.'\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [219, 65]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/3135337687959494817_0.png",
"training_data/AITW/aitw_images/general/3135337687959494817_1.png"
] | 2,905
|
[
{
"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 news in Pakistan?<image>Thought: Click the Chrome icon to navigate to the browser for checking news in Pakistan.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [367, 895]}<image>Thought: Realize the task is to find news in Pakistan, so first search for relevant content. Recognize that searching for hotels is not related to news. Notice the Google search bar is used incorrectly for this task. Adjust the query in the search bar to include 'news in Pakistan' instead of 'Search for hotels.'\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [219, 65]}<image>",
"role": "user"
},
{
"content": "Thought: You click on the search bar to change the query for news in Pakistan.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [315, 66]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/3135337687959494817_0.png",
"training_data/AITW/aitw_images/general/3135337687959494817_1.png",
"training_data/AITW/aitw_images/general/3135337687959494817_2.png"
] | 2,906
|
[
{
"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 news in Pakistan?<image>Thought: Realize the task is to find news in Pakistan, so first search for relevant content. Recognize that searching for hotels is not related to news. Notice the Google search bar is used incorrectly for this task. Adjust the query in the search bar to include 'news in Pakistan' instead of 'Search for hotels.'\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [219, 65]}<image>Thought: You click on the search bar to change the query for news in Pakistan.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [315, 66]}<image>",
"role": "user"
},
{
"content": "Thought: Type the query into the search bar to find news related to Pakistan.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the news in Pakistan?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/3135337687959494817_1.png",
"training_data/AITW/aitw_images/general/3135337687959494817_2.png",
"training_data/AITW/aitw_images/general/3135337687959494817_3.png"
] | 2,907
|
[
{
"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 news in Pakistan?<image>Thought: You click on the search bar to change the query for news in Pakistan.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [315, 66]}<image>Thought: Type the query into the search bar to find news related to Pakistan.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the news in Pakistan?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"What's the news in Pakistan?\" to view relevant news articles related to that query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [315, 132]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/3135337687959494817_2.png",
"training_data/AITW/aitw_images/general/3135337687959494817_3.png",
"training_data/AITW/aitw_images/general/3135337687959494817_4.png"
] | 2,908
|
[
{
"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 news in Pakistan?<image>Thought: Type the query into the search bar to find news related to Pakistan.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the news in Pakistan?\", \"position\": null}<image>Thought: Click on \"What's the news in Pakistan?\" to view relevant news articles related to that query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [315, 132]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully searched for news in Pakistan using the Google search interface. The results displayed various news sources relevant to your query.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/3135337687959494817_3.png",
"training_data/AITW/aitw_images/general/3135337687959494817_4.png",
"training_data/AITW/aitw_images/general/3135337687959494817_5.png"
] | 2,909
|
[
{
"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 weather like in New York?<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view additional weather information related to New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_0.png"
] | 2,910
|
[
{
"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 weather like in New York?<image>Thought: Scroll down to view additional weather information related to New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click INFO to access detailed weather information for New York.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [318, 527]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_0.png",
"training_data/AITW/aitw_images/general/13100191455285533715_1.png"
] | 2,911
|
[
{
"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 weather like in New York?<image>Thought: Scroll down to view additional weather information related to New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click INFO to access detailed weather information for New York.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [318, 527]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to check for more weather information or apps that can provide the forecast for New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_0.png",
"training_data/AITW/aitw_images/general/13100191455285533715_1.png",
"training_data/AITW/aitw_images/general/13100191455285533715_2.png"
] | 2,912
|
[
{
"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 weather like in New York?<image>Thought: Click INFO to access detailed weather information for New York.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [318, 527]}<image>Thought: Scroll down to check for more weather information or apps that can provide the forecast for New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the Google app icon to access weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [360, 548]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_1.png",
"training_data/AITW/aitw_images/general/13100191455285533715_2.png",
"training_data/AITW/aitw_images/general/13100191455285533715_3.png"
] | 2,913
|
[
{
"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 weather like in New York?<image>Thought: Scroll down to check for more weather information or apps that can provide the forecast for New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click on the Google app icon to access weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [360, 548]}<image>",
"role": "user"
},
{
"content": "Thought: Click the GOOGLE app to access the internet and check the weather in New York.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [369, 536]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_2.png",
"training_data/AITW/aitw_images/general/13100191455285533715_3.png",
"training_data/AITW/aitw_images/general/13100191455285533715_4.png"
] | 2,914
|
[
{
"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 weather like in New York?<image>Thought: Click on the Google app icon to access weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [360, 548]}<image>Thought: Click the GOOGLE app to access the internet and check the weather in New York.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [369, 536]}<image>",
"role": "user"
},
{
"content": "Thought: Click \"get to the nearest\" because I want directions, not weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [333, 170]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_3.png",
"training_data/AITW/aitw_images/general/13100191455285533715_4.png",
"training_data/AITW/aitw_images/general/13100191455285533715_5.png"
] | 2,915
|
[
{
"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 weather like in New York?<image>Thought: Click the GOOGLE app to access the internet and check the weather in New York.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [369, 536]}<image>Thought: Click \"get to the nearest\" because I want directions, not weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [333, 170]}<image>",
"role": "user"
},
{
"content": "Thought: The user clicks the \"X\" to clear the current search text, indicating a need to start fresh.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [478, 83]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_4.png",
"training_data/AITW/aitw_images/general/13100191455285533715_5.png",
"training_data/AITW/aitw_images/general/13100191455285533715_6.png"
] | 2,916
|
[
{
"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 weather like in New York?<image>Thought: Click \"get to the nearest\" because I want directions, not weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [333, 170]}<image>Thought: The user clicks the \"X\" to clear the current search text, indicating a need to start fresh.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [478, 83]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more search results related to the weather. Click on a relevant result about New York's weather.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [279, 79]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_5.png",
"training_data/AITW/aitw_images/general/13100191455285533715_6.png",
"training_data/AITW/aitw_images/general/13100191455285533715_7.png"
] | 2,917
|
[
{
"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 weather like in New York?<image>Thought: The user clicks the \"X\" to clear the current search text, indicating a need to start fresh.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [478, 83]}<image>Thought: Scroll down to view more search results related to the weather. Click on a relevant result about New York's weather.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [279, 79]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"What's the weather like in New York?\" in the search bar to find relevant information.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the weather like in New York? \", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_6.png",
"training_data/AITW/aitw_images/general/13100191455285533715_7.png",
"training_data/AITW/aitw_images/general/13100191455285533715_8.png"
] | 2,918
|
[
{
"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 weather like in New York?<image>Thought: Scroll down to view more search results related to the weather. Click on a relevant result about New York's weather.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [279, 79]}<image>Thought: Type \"What's the weather like in New York?\" in the search bar to find relevant information.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the weather like in New York? \", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"november\" to find specific weather information for New York during that month.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [222, 179]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_7.png",
"training_data/AITW/aitw_images/general/13100191455285533715_8.png",
"training_data/AITW/aitw_images/general/13100191455285533715_9.png"
] | 2,919
|
[
{
"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 weather like in New York?<image>Thought: Type \"What's the weather like in New York?\" in the search bar to find relevant information.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the weather like in New York? \", \"position\": null}<image>Thought: Click on \"november\" to find specific weather information for New York during that month.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [222, 179]}<image>",
"role": "user"
},
{
"content": "Thought: You would have looked up the weather by using Google's search bar to get the current information.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13100191455285533715_8.png",
"training_data/AITW/aitw_images/general/13100191455285533715_9.png",
"training_data/AITW/aitw_images/general/13100191455285533715_10.png"
] | 2,920
|
[
{
"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 do I get to the nearest Verizon Store?<image>",
"role": "user"
},
{
"content": "Thought: Click the CHAT icon to seek assistance for finding the nearest Verizon Store.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [157, 564]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14943476856120983082_0.png"
] | 2,921
|
[
{
"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 do I get to the nearest Verizon Store?<image>Thought: Click the CHAT icon to seek assistance for finding the nearest Verizon Store.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [157, 564]}<image>",
"role": "user"
},
{
"content": "Thought: Type the query in the search bar to find information about the nearest Verizon Store.\nAction: {\"action\": \"TYPE\", \"value\": \"How do I get to the nearest Verizon Store?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14943476856120983082_0.png",
"training_data/AITW/aitw_images/general/14943476856120983082_1.png"
] | 2,922
|
[
{
"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 do I get to the nearest Verizon Store?<image>Thought: Click the CHAT icon to seek assistance for finding the nearest Verizon Store.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [157, 564]}<image>Thought: Type the query in the search bar to find information about the nearest Verizon Store.\nAction: {\"action\": \"TYPE\", \"value\": \"How do I get to the nearest Verizon Store?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the \"directions to the nearest verizon store\" option to obtain navigation guidance.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [111, 69]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14943476856120983082_0.png",
"training_data/AITW/aitw_images/general/14943476856120983082_1.png",
"training_data/AITW/aitw_images/general/14943476856120983082_2.png"
] | 2,923
|
[
{
"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 do I get to the nearest Verizon Store?<image>Thought: Type the query in the search bar to find information about the nearest Verizon Store.\nAction: {\"action\": \"TYPE\", \"value\": \"How do I get to the nearest Verizon Store?\", \"position\": null}<image>Thought: Click on the \"directions to the nearest verizon store\" option to obtain navigation guidance.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [111, 69]}<image>",
"role": "user"
},
{
"content": "Thought: You completed the task by obtaining directions to the nearest Verizon Store shown in search results.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14943476856120983082_1.png",
"training_data/AITW/aitw_images/general/14943476856120983082_2.png",
"training_data/AITW/aitw_images/general/14943476856120983082_3.png"
] | 2,924
|
[
{
"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: Where can I buy a nice beach towel?<image>",
"role": "user"
},
{
"content": "Thought: Go back to the previous screen to find the option for buying a beach towel.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [248, 1392]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/11651723313535563989_0.png"
] | 2,925
|
[
{
"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: Where can I buy a nice beach towel?<image>Thought: Go back to the previous screen to find the option for buying a beach towel.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [248, 1392]}<image>",
"role": "user"
},
{
"content": "Thought: Click on Search apps, web, and more to start looking for a beach towel.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [202, 90]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/11651723313535563989_0.png",
"training_data/AITW/aitw_images/general/11651723313535563989_1.png"
] | 2,926
|
[
{
"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: Where can I buy a nice beach towel?<image>Thought: Go back to the previous screen to find the option for buying a beach towel.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [248, 1392]}<image>Thought: Click on Search apps, web, and more to start looking for a beach towel.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [202, 90]}<image>",
"role": "user"
},
{
"content": "Thought: Type the question in the search bar to get results for buying a beach towel.\nAction: {\"action\": \"TYPE\", \"value\": \"Where can I buy a nice beach towel?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/11651723313535563989_0.png",
"training_data/AITW/aitw_images/general/11651723313535563989_1.png",
"training_data/AITW/aitw_images/general/11651723313535563989_2.png"
] | 2,927
|
[
{
"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: Where can I buy a nice beach towel?<image>Thought: Click on Search apps, web, and more to start looking for a beach towel.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [202, 90]}<image>Thought: Type the question in the search bar to get results for buying a beach towel.\nAction: {\"action\": \"TYPE\", \"value\": \"Where can I buy a nice beach towel?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"towel near me\" to find local stores that sell beach towels.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [221, 184]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/11651723313535563989_1.png",
"training_data/AITW/aitw_images/general/11651723313535563989_2.png",
"training_data/AITW/aitw_images/general/11651723313535563989_3.png"
] | 2,928
|
[
{
"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: Where can I buy a nice beach towel?<image>Thought: Type the question in the search bar to get results for buying a beach towel.\nAction: {\"action\": \"TYPE\", \"value\": \"Where can I buy a nice beach towel?\", \"position\": null}<image>Thought: Click on \"towel near me\" to find local stores that sell beach towels.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [221, 184]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully found options for buying a nice beach towel featured prominently in the Google search results.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/11651723313535563989_2.png",
"training_data/AITW/aitw_images/general/11651723313535563989_3.png",
"training_data/AITW/aitw_images/general/11651723313535563989_4.png"
] | 2,929
|
[
{
"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 price of the new iPhone on eBay?<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more content on the homepage and locate the eBay app or website for pricing information.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_0.png"
] | 2,930
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Scroll down to view more content on the homepage and locate the eBay app or website for pricing information.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click Chrome to open the web browser for searching information about the new iPhone on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [161, 208]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_0.png",
"training_data/AITW/aitw_images/general/13290084166510249058_1.png"
] | 2,931
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Scroll down to view more content on the homepage and locate the eBay app or website for pricing information.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click Chrome to open the web browser for searching information about the new iPhone on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [161, 208]}<image>",
"role": "user"
},
{
"content": "Thought: Status task impossible: You tried to click an unrelated element (weather) instead of searching eBay for the iPhone price.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [131, 100]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_0.png",
"training_data/AITW/aitw_images/general/13290084166510249058_1.png",
"training_data/AITW/aitw_images/general/13290084166510249058_2.png"
] | 2,932
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Click Chrome to open the web browser for searching information about the new iPhone on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [161, 208]}<image>Thought: Status task impossible: You tried to click an unrelated element (weather) instead of searching eBay for the iPhone price.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [131, 100]}<image>",
"role": "user"
},
{
"content": "Thought: Status task impossible: The operation is unrelated to checking the iPhone price on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 111]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_1.png",
"training_data/AITW/aitw_images/general/13290084166510249058_2.png",
"training_data/AITW/aitw_images/general/13290084166510249058_3.png"
] | 2,933
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Status task impossible: You tried to click an unrelated element (weather) instead of searching eBay for the iPhone price.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [131, 100]}<image>Thought: Status task impossible: The operation is unrelated to checking the iPhone price on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 111]}<image>",
"role": "user"
},
{
"content": "Thought: status task impossible: The user is focused on weather information instead of searching for iPhone prices on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 112]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_2.png",
"training_data/AITW/aitw_images/general/13290084166510249058_3.png",
"training_data/AITW/aitw_images/general/13290084166510249058_4.png"
] | 2,934
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Status task impossible: The operation is unrelated to checking the iPhone price on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 111]}<image>Thought: status task impossible: The user is focused on weather information instead of searching for iPhone prices on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 112]}<image>",
"role": "user"
},
{
"content": "Thought: You wanted to check the weather in Beijing, but you clicked on the wrong query instead of searching for iPhone prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 111]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_3.png",
"training_data/AITW/aitw_images/general/13290084166510249058_4.png",
"training_data/AITW/aitw_images/general/13290084166510249058_5.png"
] | 2,935
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: status task impossible: The user is focused on weather information instead of searching for iPhone prices on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 112]}<image>Thought: You wanted to check the weather in Beijing, but you clicked on the wrong query instead of searching for iPhone prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 111]}<image>",
"role": "user"
},
{
"content": "Thought: Click the search bar to enter a new query for the iPhone price.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 43]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_4.png",
"training_data/AITW/aitw_images/general/13290084166510249058_5.png",
"training_data/AITW/aitw_images/general/13290084166510249058_6.png"
] | 2,936
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: You wanted to check the weather in Beijing, but you clicked on the wrong query instead of searching for iPhone prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [129, 111]}<image>Thought: Click the search bar to enter a new query for the iPhone price.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 43]}<image>",
"role": "user"
},
{
"content": "Thought: Click in the search bar to start entering the query for iPhone price on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [156, 29]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_5.png",
"training_data/AITW/aitw_images/general/13290084166510249058_6.png",
"training_data/AITW/aitw_images/general/13290084166510249058_7.png"
] | 2,937
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Click the search bar to enter a new query for the iPhone price.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [150, 43]}<image>Thought: Click in the search bar to start entering the query for iPhone price on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [156, 29]}<image>",
"role": "user"
},
{
"content": "Thought: Type the query into the search bar to find the price information for the iPhone on eBay.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the price of the new iPhone on eBay?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_6.png",
"training_data/AITW/aitw_images/general/13290084166510249058_7.png",
"training_data/AITW/aitw_images/general/13290084166510249058_8.png"
] | 2,938
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Click in the search bar to start entering the query for iPhone price on eBay.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [156, 29]}<image>Thought: Type the query into the search bar to find the price information for the iPhone on eBay.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the price of the new iPhone on eBay?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You likely finished typing your search query in the search bar, so pressing enter will execute the search.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_7.png",
"training_data/AITW/aitw_images/general/13290084166510249058_8.png",
"training_data/AITW/aitw_images/general/13290084166510249058_9.png"
] | 2,939
|
[
{
"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 price of the new iPhone on eBay?<image>Thought: Type the query into the search bar to find the price information for the iPhone on eBay.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the price of the new iPhone on eBay?\", \"position\": null}<image>Thought: You likely finished typing your search query in the search bar, so pressing enter will execute the search.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You have completed the task because the price information for the new iPhone was successfully retrieved from eBay.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/13290084166510249058_8.png",
"training_data/AITW/aitw_images/general/13290084166510249058_9.png",
"training_data/AITW/aitw_images/general/13290084166510249058_10.png"
] | 2,940
|
[
{
"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 3 bedroom apartment rent for in New York?<image>",
"role": "user"
},
{
"content": "Thought: Press home to exit the current settings menu and return to the main screen for searching apartment prices.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_0.png"
] | 2,941
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Press home to exit the current settings menu and return to the main screen for searching apartment prices.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to see more search results or information related to apartment rents in New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_0.png",
"training_data/AITW/aitw_images/general/15844287487684771521_1.png"
] | 2,942
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Press home to exit the current settings menu and return to the main screen for searching apartment prices.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Scroll down to see more search results or information related to apartment rents in New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click GOOGLE to access the browser for searching apartment rental prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [469, 536]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_0.png",
"training_data/AITW/aitw_images/general/15844287487684771521_1.png",
"training_data/AITW/aitw_images/general/15844287487684771521_2.png"
] | 2,943
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Scroll down to see more search results or information related to apartment rents in New York.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click GOOGLE to access the browser for searching apartment rental prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [469, 536]}<image>",
"role": "user"
},
{
"content": "Thought: You likely intended to search for something related to wallpaper instead of apartment rent. Click on the suggested query \"show me some nice wallpapers\" to proceed with that interest.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [226, 366]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_1.png",
"training_data/AITW/aitw_images/general/15844287487684771521_2.png",
"training_data/AITW/aitw_images/general/15844287487684771521_3.png"
] | 2,944
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Click GOOGLE to access the browser for searching apartment rental prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [469, 536]}<image>Thought: You likely intended to search for something related to wallpaper instead of apartment rent. Click on the suggested query \"show me some nice wallpapers\" to proceed with that interest.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [226, 366]}<image>",
"role": "user"
},
{
"content": "Thought: Click the X to clear the search for relevant apartment rental information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [491, 67]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_2.png",
"training_data/AITW/aitw_images/general/15844287487684771521_3.png",
"training_data/AITW/aitw_images/general/15844287487684771521_4.png"
] | 2,945
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: You likely intended to search for something related to wallpaper instead of apartment rent. Click on the suggested query \"show me some nice wallpapers\" to proceed with that interest.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [226, 366]}<image>Thought: Click the X to clear the search for relevant apartment rental information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [491, 67]}<image>",
"role": "user"
},
{
"content": "Thought: You clicked on \"show me some nice wallpapers for my phone free\" likely due to an unrelated interest, diverting from the original inquiry about apartment rent.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [259, 82]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_3.png",
"training_data/AITW/aitw_images/general/15844287487684771521_4.png",
"training_data/AITW/aitw_images/general/15844287487684771521_5.png"
] | 2,946
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Click the X to clear the search for relevant apartment rental information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [491, 67]}<image>Thought: You clicked on \"show me some nice wallpapers for my phone free\" likely due to an unrelated interest, diverting from the original inquiry about apartment rent.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [259, 82]}<image>",
"role": "user"
},
{
"content": "Thought: Type the question in the search bar to find rental prices easily.\nAction: {\"action\": \"TYPE\", \"value\": \"How much does a 3 bedroom apartment rent for in New York \", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_4.png",
"training_data/AITW/aitw_images/general/15844287487684771521_5.png",
"training_data/AITW/aitw_images/general/15844287487684771521_6.png"
] | 2,947
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: You clicked on \"show me some nice wallpapers for my phone free\" likely due to an unrelated interest, diverting from the original inquiry about apartment rent.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [259, 82]}<image>Thought: Type the question in the search bar to find rental prices easily.\nAction: {\"action\": \"TYPE\", \"value\": \"How much does a 3 bedroom apartment rent for in New York \", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the city suggestion to refine the search for apartment rents specifically in New York City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [190, 241]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_5.png",
"training_data/AITW/aitw_images/general/15844287487684771521_6.png",
"training_data/AITW/aitw_images/general/15844287487684771521_7.png"
] | 2,948
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Type the question in the search bar to find rental prices easily.\nAction: {\"action\": \"TYPE\", \"value\": \"How much does a 3 bedroom apartment rent for in New York \", \"position\": null}<image>Thought: Click the city suggestion to refine the search for apartment rents specifically in New York City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [190, 241]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"3 Bdrm Apts In Lower Manhattan - A Manhattan Masterpiece\" link to find rental information for 3-bedroom apartments.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [271, 421]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_6.png",
"training_data/AITW/aitw_images/general/15844287487684771521_7.png",
"training_data/AITW/aitw_images/general/15844287487684771521_8.png"
] | 2,949
|
[
{
"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 3 bedroom apartment rent for in New York?<image>Thought: Click the city suggestion to refine the search for apartment rents specifically in New York City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [190, 241]}<image>Thought: Click on \"3 Bdrm Apts In Lower Manhattan - A Manhattan Masterpiece\" link to find rental information for 3-bedroom apartments.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [271, 421]}<image>",
"role": "user"
},
{
"content": "Thought: You must have accepted the cookie policy to access rental information on the website.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/15844287487684771521_7.png",
"training_data/AITW/aitw_images/general/15844287487684771521_8.png",
"training_data/AITW/aitw_images/general/15844287487684771521_9.png"
] | 2,950
|
[
{
"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 weather like in London?<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen and search for a weather app.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_0.png"
] | 2,951
|
[
{
"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 weather like in London?<image>Thought: Press home to return to the main screen and search for a weather app.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click BACKWARD to return to the previous screen where weather information is likely accessible.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [186, 1016]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_0.png",
"training_data/AITW/aitw_images/general/12735674421312260299_1.png"
] | 2,952
|
[
{
"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 weather like in London?<image>Thought: Press home to return to the main screen and search for a weather app.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click BACKWARD to return to the previous screen where weather information is likely accessible.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [186, 1016]}<image>",
"role": "user"
},
{
"content": "Thought: Click back to return to the previous screen where I can find weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [174, 1034]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_0.png",
"training_data/AITW/aitw_images/general/12735674421312260299_1.png",
"training_data/AITW/aitw_images/general/12735674421312260299_2.png"
] | 2,953
|
[
{
"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 weather like in London?<image>Thought: Click BACKWARD to return to the previous screen where weather information is likely accessible.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [186, 1016]}<image>Thought: Click back to return to the previous screen where I can find weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [174, 1034]}<image>",
"role": "user"
},
{
"content": "Thought: Click GOOGLE to open the search browser for checking the weather in London.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [178, 998]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_1.png",
"training_data/AITW/aitw_images/general/12735674421312260299_2.png",
"training_data/AITW/aitw_images/general/12735674421312260299_3.png"
] | 2,954
|
[
{
"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 weather like in London?<image>Thought: Click back to return to the previous screen where I can find weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [174, 1034]}<image>Thought: Click GOOGLE to open the search browser for checking the weather in London.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [178, 998]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Search apps, web, and more...\" to start looking for weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [147, 68]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_2.png",
"training_data/AITW/aitw_images/general/12735674421312260299_3.png",
"training_data/AITW/aitw_images/general/12735674421312260299_4.png"
] | 2,955
|
[
{
"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 weather like in London?<image>Thought: Click GOOGLE to open the search browser for checking the weather in London.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [178, 998]}<image>Thought: Click on \"Search apps, web, and more...\" to start looking for weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [147, 68]}<image>",
"role": "user"
},
{
"content": "Thought: Type the query in the search box to find information about London's weather.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the weather like in London?\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_3.png",
"training_data/AITW/aitw_images/general/12735674421312260299_4.png",
"training_data/AITW/aitw_images/general/12735674421312260299_5.png"
] | 2,956
|
[
{
"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 weather like in London?<image>Thought: Click on \"Search apps, web, and more...\" to start looking for weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [147, 68]}<image>Thought: Type the query in the search box to find information about London's weather.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the weather like in London?\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the search suggestion \"what's the weather like in london today\" to find current weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [280, 157]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_4.png",
"training_data/AITW/aitw_images/general/12735674421312260299_5.png",
"training_data/AITW/aitw_images/general/12735674421312260299_6.png"
] | 2,957
|
[
{
"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 weather like in London?<image>Thought: Type the query in the search box to find information about London's weather.\nAction: {\"action\": \"TYPE\", \"value\": \"What's the weather like in London?\", \"position\": null}<image>Thought: Click on the search suggestion \"what's the weather like in london today\" to find current weather information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [280, 157]}<image>",
"role": "user"
},
{
"content": "Thought: You completed the task by checking the weather information displayed for London, UK.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/12735674421312260299_5.png",
"training_data/AITW/aitw_images/general/12735674421312260299_6.png",
"training_data/AITW/aitw_images/general/12735674421312260299_7.png"
] | 2,958
|
[
{
"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 Seoul to Mexico city<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen and access the flight search app.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_0.png"
] | 2,959
|
[
{
"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 Seoul to Mexico city<image>Thought: Press home to return to the main screen and access the flight search app.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You would click CHAT to open a messaging app for assistance in searching for flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [164, 478]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_0.png",
"training_data/AITW/aitw_images/general/17212614498000101786_1.png"
] | 2,960
|
[
{
"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 Seoul to Mexico city<image>Thought: Press home to return to the main screen and access the flight search app.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: You would click CHAT to open a messaging app for assistance in searching for flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [164, 478]}<image>",
"role": "user"
},
{
"content": "Thought: You clicked X to dismiss an unrelated search result and refocus on the flight search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [250, 109]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_0.png",
"training_data/AITW/aitw_images/general/17212614498000101786_1.png",
"training_data/AITW/aitw_images/general/17212614498000101786_2.png"
] | 2,961
|
[
{
"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 Seoul to Mexico city<image>Thought: You would click CHAT to open a messaging app for assistance in searching for flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [164, 478]}<image>Thought: You clicked X to dismiss an unrelated search result and refocus on the flight search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [250, 109]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"flights from Seoul to Mexico city\" in the search bar to initiate the flight search.\nAction: {\"action\": \"TYPE\", \"value\": \"flights from Seoul to Mexico city\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_1.png",
"training_data/AITW/aitw_images/general/17212614498000101786_2.png",
"training_data/AITW/aitw_images/general/17212614498000101786_3.png"
] | 2,962
|
[
{
"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 Seoul to Mexico city<image>Thought: You clicked X to dismiss an unrelated search result and refocus on the flight search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [250, 109]}<image>Thought: Type \"flights from Seoul to Mexico city\" in the search bar to initiate the flight search.\nAction: {\"action\": \"TYPE\", \"value\": \"flights from Seoul to Mexico city\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You click on \"mexico city\" from the search suggestions to refine your flight search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [199, 101]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_2.png",
"training_data/AITW/aitw_images/general/17212614498000101786_3.png",
"training_data/AITW/aitw_images/general/17212614498000101786_4.png"
] | 2,963
|
[
{
"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 Seoul to Mexico city<image>Thought: Type \"flights from Seoul to Mexico city\" in the search bar to initiate the flight search.\nAction: {\"action\": \"TYPE\", \"value\": \"flights from Seoul to Mexico city\", \"position\": null}<image>Thought: You click on \"mexico city\" from the search suggestions to refine your flight search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [199, 101]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"All\" to see a comprehensive list of results for flights from Seoul to Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [15, 142]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_3.png",
"training_data/AITW/aitw_images/general/17212614498000101786_4.png",
"training_data/AITW/aitw_images/general/17212614498000101786_5.png"
] | 2,964
|
[
{
"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 Seoul to Mexico city<image>Thought: You click on \"mexico city\" from the search suggestions to refine your flight search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [199, 101]}<image>Thought: Click on \"All\" to see a comprehensive list of results for flights from Seoul to Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [15, 142]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more search results for flights from Seoul to Mexico City.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_4.png",
"training_data/AITW/aitw_images/general/17212614498000101786_5.png",
"training_data/AITW/aitw_images/general/17212614498000101786_6.png"
] | 2,965
|
[
{
"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 Seoul to Mexico city<image>Thought: Click on \"All\" to see a comprehensive list of results for flights from Seoul to Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [15, 142]}<image>Thought: Scroll down to view more search results for flights from Seoul to Mexico City.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more flight options or scheduling details for the search.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_5.png",
"training_data/AITW/aitw_images/general/17212614498000101786_6.png",
"training_data/AITW/aitw_images/general/17212614498000101786_7.png"
] | 2,966
|
[
{
"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 Seoul to Mexico city<image>Thought: Scroll down to view more search results for flights from Seoul to Mexico City.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Scroll down to view more flight options or scheduling details for the search.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the downward arrow to view more flight options listed below.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [108, 149]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_6.png",
"training_data/AITW/aitw_images/general/17212614498000101786_7.png",
"training_data/AITW/aitw_images/general/17212614498000101786_8.png"
] | 2,967
|
[
{
"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 Seoul to Mexico city<image>Thought: Scroll down to view more flight options or scheduling details for the search.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click the downward arrow to view more flight options listed below.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [108, 149]}<image>",
"role": "user"
},
{
"content": "Thought: Click the dropdown arrow to expand options for flight details.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [123, 162]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_7.png",
"training_data/AITW/aitw_images/general/17212614498000101786_8.png",
"training_data/AITW/aitw_images/general/17212614498000101786_9.png"
] | 2,968
|
[
{
"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 Seoul to Mexico city<image>Thought: Click the downward arrow to view more flight options listed below.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [108, 149]}<image>Thought: Click the dropdown arrow to expand options for flight details.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [123, 162]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll up to view more options or details related to flight inquiries for better search results.\nAction: {\"action\": \"SCROLL UP\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_8.png",
"training_data/AITW/aitw_images/general/17212614498000101786_9.png",
"training_data/AITW/aitw_images/general/17212614498000101786_10.png"
] | 2,969
|
[
{
"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 Seoul to Mexico city<image>Thought: Click the dropdown arrow to expand options for flight details.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [123, 162]}<image>Thought: Scroll up to view more options or details related to flight inquiries for better search results.\nAction: {\"action\": \"SCROLL UP\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"19h\" to view more details about that specific flight option.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [171, 425]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_9.png",
"training_data/AITW/aitw_images/general/17212614498000101786_10.png",
"training_data/AITW/aitw_images/general/17212614498000101786_11.png"
] | 2,970
|
[
{
"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 Seoul to Mexico city<image>Thought: Scroll up to view more options or details related to flight inquiries for better search results.\nAction: {\"action\": \"SCROLL UP\", \"value\": null, \"position\": null}<image>Thought: Click on \"19h\" to view more details about that specific flight option.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [171, 425]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Search on Google Flights\" to view specific flight options from Seoul to Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [114, 494]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_10.png",
"training_data/AITW/aitw_images/general/17212614498000101786_11.png",
"training_data/AITW/aitw_images/general/17212614498000101786_12.png"
] | 2,971
|
[
{
"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 Seoul to Mexico city<image>Thought: Click on \"19h\" to view more details about that specific flight option.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [171, 425]}<image>Thought: Click on \"Search on Google Flights\" to view specific flight options from Seoul to Mexico City.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [114, 494]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully searched for flights from Seoul to Mexico City, as indicated by the flight listings shown.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/17212614498000101786_11.png",
"training_data/AITW/aitw_images/general/17212614498000101786_12.png",
"training_data/AITW/aitw_images/general/17212614498000101786_13.png"
] | 2,972
|
[
{
"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 Helsinki to Seoul<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to find the flight search option on the home screen or app.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_0.png"
] | 2,973
|
[
{
"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 Helsinki to Seoul<image>Thought: Scroll down to find the flight search option on the home screen or app.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the LOCATION button to start searching for flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [238, 984]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_0.png",
"training_data/AITW/aitw_images/general/14126783365089023093_1.png"
] | 2,974
|
[
{
"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 Helsinki to Seoul<image>Thought: Scroll down to find the flight search option on the home screen or app.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click the LOCATION button to start searching for flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [238, 984]}<image>",
"role": "user"
},
{
"content": "Thought: Click CHAT to access the messaging app for flight inquiries or assistance.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [205, 1014]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_0.png",
"training_data/AITW/aitw_images/general/14126783365089023093_1.png",
"training_data/AITW/aitw_images/general/14126783365089023093_2.png"
] | 2,975
|
[
{
"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 Helsinki to Seoul<image>Thought: Click the LOCATION button to start searching for flights.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [238, 984]}<image>Thought: Click CHAT to access the messaging app for flight inquiries or assistance.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [205, 1014]}<image>",
"role": "user"
},
{
"content": "Thought: Type the search query \"flights from Helsinki to Seoul\" in the search bar to find relevant flight information.\nAction: {\"action\": \"TYPE\", \"value\": \"flights from Helsinki to Seoul\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_1.png",
"training_data/AITW/aitw_images/general/14126783365089023093_2.png",
"training_data/AITW/aitw_images/general/14126783365089023093_3.png"
] | 2,976
|
[
{
"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 Helsinki to Seoul<image>Thought: Click CHAT to access the messaging app for flight inquiries or assistance.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [205, 1014]}<image>Thought: Type the search query \"flights from Helsinki to Seoul\" in the search bar to find relevant flight information.\nAction: {\"action\": \"TYPE\", \"value\": \"flights from Helsinki to Seoul\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"flights from helsinki to seoul\" to select the most relevant search suggestion for flight information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [401, 157]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_2.png",
"training_data/AITW/aitw_images/general/14126783365089023093_3.png",
"training_data/AITW/aitw_images/general/14126783365089023093_4.png"
] | 2,977
|
[
{
"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 Helsinki to Seoul<image>Thought: Type the search query \"flights from Helsinki to Seoul\" in the search bar to find relevant flight information.\nAction: {\"action\": \"TYPE\", \"value\": \"flights from Helsinki to Seoul\", \"position\": null}<image>Thought: Click on \"flights from helsinki to seoul\" to select the most relevant search suggestion for flight information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [401, 157]}<image>",
"role": "user"
},
{
"content": "Thought: Click on News likely to see any articles or updates related to flights from Helsinki to Seoul.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [307, 283]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_3.png",
"training_data/AITW/aitw_images/general/14126783365089023093_4.png",
"training_data/AITW/aitw_images/general/14126783365089023093_5.png"
] | 2,978
|
[
{
"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 Helsinki to Seoul<image>Thought: Click on \"flights from helsinki to seoul\" to select the most relevant search suggestion for flight information.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [401, 157]}<image>Thought: Click on News likely to see any articles or updates related to flights from Helsinki to Seoul.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [307, 283]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully searched for flights from Helsinki to Seoul using the Google search bar.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14126783365089023093_4.png",
"training_data/AITW/aitw_images/general/14126783365089023093_5.png",
"training_data/AITW/aitw_images/general/14126783365089023093_6.png"
] | 2,979
|
[
{
"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 Atlanta<image>",
"role": "user"
},
{
"content": "Thought: Press home to return to the main screen, as the current interface does not support hotel searches.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_0.png"
] | 2,980
|
[
{
"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 Atlanta<image>Thought: Press home to return to the main screen, as the current interface does not support hotel searches.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the NAV BAR CIRCLE to access the app menu for hotel search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [360, 1274]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_0.png",
"training_data/AITW/aitw_images/general/14074740133260658786_1.png"
] | 2,981
|
[
{
"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 Atlanta<image>Thought: Press home to return to the main screen, as the current interface does not support hotel searches.\nAction: {\"action\": \"PRESS HOME\", \"value\": null, \"position\": null}<image>Thought: Click the NAV BAR CIRCLE to access the app menu for hotel search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [360, 1274]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"hotels in Atlanta\" in the search bar to start looking for hotel options.\nAction: {\"action\": \"TYPE\", \"value\": \"hotels in atlanta\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_0.png",
"training_data/AITW/aitw_images/general/14074740133260658786_1.png",
"training_data/AITW/aitw_images/general/14074740133260658786_2.png"
] | 2,982
|
[
{
"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 Atlanta<image>Thought: Click the NAV BAR CIRCLE to access the app menu for hotel search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [360, 1274]}<image>Thought: Type \"hotels in Atlanta\" in the search bar to start looking for hotel options.\nAction: {\"action\": \"TYPE\", \"value\": \"hotels in atlanta\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"atlanta\" to select this suggested search term for hotels in Atlanta.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [248, 180]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_1.png",
"training_data/AITW/aitw_images/general/14074740133260658786_2.png",
"training_data/AITW/aitw_images/general/14074740133260658786_3.png"
] | 2,983
|
[
{
"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 Atlanta<image>Thought: Type \"hotels in Atlanta\" in the search bar to start looking for hotel options.\nAction: {\"action\": \"TYPE\", \"value\": \"hotels in atlanta\", \"position\": null}<image>Thought: Click on \"atlanta\" to select this suggested search term for hotels in Atlanta.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [248, 180]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"260 Hotels in Atlanta, Georgia - Lowest Price Guarantee\" to view available hotel options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [170, 459]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_2.png",
"training_data/AITW/aitw_images/general/14074740133260658786_3.png",
"training_data/AITW/aitw_images/general/14074740133260658786_4.png"
] | 2,984
|
[
{
"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 Atlanta<image>Thought: Click on \"atlanta\" to select this suggested search term for hotels in Atlanta.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [248, 180]}<image>Thought: Click on \"260 Hotels in Atlanta, Georgia - Lowest Price Guarantee\" to view available hotel options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [170, 459]}<image>",
"role": "user"
},
{
"content": "Thought: Click the Search button to initiate the hotel search for Atlanta.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [309, 998]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_3.png",
"training_data/AITW/aitw_images/general/14074740133260658786_4.png",
"training_data/AITW/aitw_images/general/14074740133260658786_5.png"
] | 2,985
|
[
{
"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 Atlanta<image>Thought: Click on \"260 Hotels in Atlanta, Georgia - Lowest Price Guarantee\" to view available hotel options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [170, 459]}<image>Thought: Click the Search button to initiate the hotel search for Atlanta.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [309, 998]}<image>",
"role": "user"
},
{
"content": "Thought: The user successfully completed the search for hotels in Atlanta, as indicated by viewing the search results.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/14074740133260658786_4.png",
"training_data/AITW/aitw_images/general/14074740133260658786_5.png",
"training_data/AITW/aitw_images/general/14074740133260658786_6.png"
] | 2,986
|
[
{
"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 eyeshadow on Sephora<image>",
"role": "user"
},
{
"content": "Thought: Click the ENVELOPE to access email notifications that may help find eyeshadow recommendations or sales.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [503, 1140]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_0.png"
] | 2,987
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click the ENVELOPE to access email notifications that may help find eyeshadow recommendations or sales.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [503, 1140]}<image>",
"role": "user"
},
{
"content": "Thought: Click the ENVELOPE icon to access your email for shopping updates or promotions for eyeshadow.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_0.png",
"training_data/AITW/aitw_images/general/1494430438008457652_1.png"
] | 2,988
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click the ENVELOPE to access email notifications that may help find eyeshadow recommendations or sales.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [503, 1140]}<image>Thought: Click the ENVELOPE icon to access your email for shopping updates or promotions for eyeshadow.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}<image>",
"role": "user"
},
{
"content": "Thought: Click the envelope icon to open your email app for potential Sephora updates or sales.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_0.png",
"training_data/AITW/aitw_images/general/1494430438008457652_1.png",
"training_data/AITW/aitw_images/general/1494430438008457652_2.png"
] | 2,989
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click the ENVELOPE icon to access your email for shopping updates or promotions for eyeshadow.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}<image>Thought: Click the envelope icon to open your email app for potential Sephora updates or sales.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}<image>",
"role": "user"
},
{
"content": "Thought: Click the ENVELOPE to open the email app, perhaps for browsing promotions on eyeshadow.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_1.png",
"training_data/AITW/aitw_images/general/1494430438008457652_2.png",
"training_data/AITW/aitw_images/general/1494430438008457652_3.png"
] | 2,990
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click the envelope icon to open your email app for potential Sephora updates or sales.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}<image>Thought: Click the ENVELOPE to open the email app, perhaps for browsing promotions on eyeshadow.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}<image>",
"role": "user"
},
{
"content": "Thought: Click on PLAY to possibly launch the desired app for searching eyeshadow at Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [485, 1139]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_2.png",
"training_data/AITW/aitw_images/general/1494430438008457652_3.png",
"training_data/AITW/aitw_images/general/1494430438008457652_4.png"
] | 2,991
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click the ENVELOPE to open the email app, perhaps for browsing promotions on eyeshadow.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [521, 1126]}<image>Thought: Click on PLAY to possibly launch the desired app for searching eyeshadow at Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [485, 1139]}<image>",
"role": "user"
},
{
"content": "Thought: Click the \"X\" to close the current search, clearing the screen for a new query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [670, 174]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_3.png",
"training_data/AITW/aitw_images/general/1494430438008457652_4.png",
"training_data/AITW/aitw_images/general/1494430438008457652_5.png"
] | 2,992
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click on PLAY to possibly launch the desired app for searching eyeshadow at Sephora.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [485, 1139]}<image>Thought: Click the \"X\" to close the current search, clearing the screen for a new query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [670, 174]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"new eyeshadow on Sephora\" in the search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"new eyeshadow on Sephora\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_4.png",
"training_data/AITW/aitw_images/general/1494430438008457652_5.png",
"training_data/AITW/aitw_images/general/1494430438008457652_6.png"
] | 2,993
|
[
{
"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 eyeshadow on Sephora<image>Thought: Click the \"X\" to close the current search, clearing the screen for a new query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [670, 174]}<image>Thought: Type \"new eyeshadow on Sephora\" in the search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"new eyeshadow on Sephora\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to execute the search query for new eyeshadow on Sephora.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_5.png",
"training_data/AITW/aitw_images/general/1494430438008457652_6.png",
"training_data/AITW/aitw_images/general/1494430438008457652_7.png"
] | 2,994
|
[
{
"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 eyeshadow on Sephora<image>Thought: Type \"new eyeshadow on Sephora\" in the search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"new eyeshadow on Sephora\", \"position\": null}<image>Thought: Press enter to execute the search query for new eyeshadow on Sephora.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: The task is complete since relevant new eyeshadow options were found on Google.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/1494430438008457652_6.png",
"training_data/AITW/aitw_images/general/1494430438008457652_7.png",
"training_data/AITW/aitw_images/general/1494430438008457652_8.png"
] | 2,995
|
[
{
"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 price of a 12' ladder at Home Depot?<image>",
"role": "user"
},
{
"content": "Thought: You want to initiate a voice search for the ladder's price by clicking the microphone icon.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [172, 496]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8083769447948585687_0.png"
] | 2,996
|
[
{
"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 price of a 12' ladder at Home Depot?<image>Thought: You want to initiate a voice search for the ladder's price by clicking the microphone icon.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [172, 496]}<image>",
"role": "user"
},
{
"content": "Thought: You clicked the link to access a website, but it's unrelated to Home Depot. You need to navigate back or search for Home Depot's site directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [106, 34]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8083769447948585687_0.png",
"training_data/AITW/aitw_images/general/8083769447948585687_1.png"
] | 2,997
|
[
{
"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 price of a 12' ladder at Home Depot?<image>Thought: You want to initiate a voice search for the ladder's price by clicking the microphone icon.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [172, 496]}<image>Thought: You clicked the link to access a website, but it's unrelated to Home Depot. You need to navigate back or search for Home Depot's site directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [106, 34]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"price of a 12' ladder at Home Depot\" into the search bar to find the information.\nAction: {\"action\": \"TYPE\", \"value\": \"price of a 12' ladder at Home Depot\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8083769447948585687_0.png",
"training_data/AITW/aitw_images/general/8083769447948585687_1.png",
"training_data/AITW/aitw_images/general/8083769447948585687_2.png"
] | 2,998
|
[
{
"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 price of a 12' ladder at Home Depot?<image>Thought: You clicked the link to access a website, but it's unrelated to Home Depot. You need to navigate back or search for Home Depot's site directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [106, 34]}<image>Thought: Type \"price of a 12' ladder at Home Depot\" into the search bar to find the information.\nAction: {\"action\": \"TYPE\", \"value\": \"price of a 12' ladder at Home Depot\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the search bar to input \"12'\" for finding ladder prices.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [89, 71]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/general/8083769447948585687_1.png",
"training_data/AITW/aitw_images/general/8083769447948585687_2.png",
"training_data/AITW/aitw_images/general/8083769447948585687_3.png"
] | 2,999
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.