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: Go to ebay <image>Thought: You want to search for items on eBay, so you click on the search bar labeled \"Search for anything\" to initiate your query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [192, 220]}<image>",
"role": "user"
},
{
"content": "Thought: Recognize that the URL shows ebay.com, indicating successful navigation to eBay.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/14567097870190513545.612-647.Go to ebay _0.png",
"training_data/AITW/aitw_images/single/14567097870190513545.612-647.Go to ebay _1.png"
] | 5,000
|
[
{
"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: Add this item to cart<image>",
"role": "user"
},
{
"content": "Thought: Click Add to Cart to select this item for purchase and proceed with my shopping.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [194, 490]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/7319060900754378920.1254-1315.Add this item to cart_0.png"
] | 5,001
|
[
{
"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: Add this item to cart<image>Thought: Click Add to Cart to select this item for purchase and proceed with my shopping.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [194, 490]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully added the item to your cart, as indicated by the subtotal showing $119.99 and \"1 item\" in the cart icon.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/7319060900754378920.1254-1315.Add this item to cart_0.png",
"training_data/AITW/aitw_images/single/7319060900754378920.1254-1315.Add this item to cart_1.png"
] | 5,002
|
[
{
"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: Go to search bar in google chrome then type amazon<image>",
"role": "user"
},
{
"content": "Thought: Tap on the newegg.com link to navigate away from the current web page. This action will take you to a different site than Google Chrome's search bar. You’ll likely be redirected to a product page rather than performing your intended task of searching for \"amazon.\" After this, realize that you are not in the desired place to type your search query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [400, 61]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13042182440874531520.99-179.Go to search bar in google chrome then type amazon_0.png"
] | 5,003
|
[
{
"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: Go to search bar in google chrome then type amazon<image>Thought: Tap on the newegg.com link to navigate away from the current web page. This action will take you to a different site than Google Chrome's search bar. You’ll likely be redirected to a product page rather than performing your intended task of searching for \"amazon.\" After this, realize that you are not in the desired place to type your search query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [400, 61]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"amazon\" into the search bar to initiate a web search for Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13042182440874531520.99-179.Go to search bar in google chrome then type amazon_0.png",
"training_data/AITW/aitw_images/single/13042182440874531520.99-179.Go to search bar in google chrome then type amazon_1.png"
] | 5,004
|
[
{
"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: Go to search bar in google chrome then type amazon<image>Thought: Tap on the newegg.com link to navigate away from the current web page. This action will take you to a different site than Google Chrome's search bar. You’ll likely be redirected to a product page rather than performing your intended task of searching for \"amazon.\" After this, realize that you are not in the desired place to type your search query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [400, 61]}<image>Thought: Type \"amazon\" into the search bar to initiate a web search for Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You typed \"amazon\" in the search bar to find the website. The task is complete as you accessed the site.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13042182440874531520.99-179.Go to search bar in google chrome then type amazon_0.png",
"training_data/AITW/aitw_images/single/13042182440874531520.99-179.Go to search bar in google chrome then type amazon_1.png",
"training_data/AITW/aitw_images/single/13042182440874531520.99-179.Go to search bar in google chrome then type amazon_2.png"
] | 5,005
|
[
{
"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 bestbuy.com in the chrome search bar <image>",
"role": "user"
},
{
"content": "Thought: You clicked on the link to walmart.com/cart, which is not related to bestbuy.com. You are likely confused about the correct URL for Best Buy. You should return to the Chrome search bar to enter \"bestbuy.com\" directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [340, 62]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _0.png"
] | 5,006
|
[
{
"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 bestbuy.com in the chrome search bar <image>Thought: You clicked on the link to walmart.com/cart, which is not related to bestbuy.com. You are likely confused about the correct URL for Best Buy. You should return to the Chrome search bar to enter \"bestbuy.com\" directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [340, 62]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"bestbuy.com\" in the Chrome search bar to locate the website directly.\nAction: {\"action\": \"TYPE\", \"value\": \"bestbuy.com\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _0.png",
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _1.png"
] | 5,007
|
[
{
"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 bestbuy.com in the chrome search bar <image>Thought: You clicked on the link to walmart.com/cart, which is not related to bestbuy.com. You are likely confused about the correct URL for Best Buy. You should return to the Chrome search bar to enter \"bestbuy.com\" directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [340, 62]}<image>Thought: Type \"bestbuy.com\" in the Chrome search bar to locate the website directly.\nAction: {\"action\": \"TYPE\", \"value\": \"bestbuy.com\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to execute the search for bestbuy.com in the Chrome search bar.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _0.png",
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _1.png",
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _2.png"
] | 5,008
|
[
{
"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 bestbuy.com in the chrome search bar <image>Thought: Type \"bestbuy.com\" in the Chrome search bar to locate the website directly.\nAction: {\"action\": \"TYPE\", \"value\": \"bestbuy.com\", \"position\": null}<image>Thought: Press enter to execute the search for bestbuy.com in the Chrome search bar.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully searched for bestbuy.com in the Chrome search bar, confirming the website is loaded.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _1.png",
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _2.png",
"training_data/AITW/aitw_images/single/5274520361556629909.86-289.search bestbuy.com in the chrome search bar _3.png"
] | 5,009
|
[
{
"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: Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor<image>",
"role": "user"
},
{
"content": "Thought: Click the 'Shop' button to navigate to the shopping section where the search bar is located.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [329, 211]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_0.png"
] | 5,010
|
[
{
"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: Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor<image>Thought: Click the 'Shop' button to navigate to the shopping section where the search bar is located.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [329, 211]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"acer nitro\" in the search bar to begin finding the desired monitor.\nAction: {\"action\": \"TYPE\", \"value\": \"acer nitro\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_0.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_1.png"
] | 5,011
|
[
{
"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: Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor<image>Thought: Click the 'Shop' button to navigate to the shopping section where the search bar is located.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [329, 211]}<image>Thought: Type \"acer nitro\" in the search bar to begin finding the desired monitor.\nAction: {\"action\": \"TYPE\", \"value\": \"acer nitro\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Type \"Acer Nitro 27\\\" class FHD gaming monitor\" into the Costco search bar to find the product.\nAction: {\"action\": \"TYPE\", \"value\": \"\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_0.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_1.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_2.png"
] | 5,012
|
[
{
"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: Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor<image>Thought: Type \"acer nitro\" in the search bar to begin finding the desired monitor.\nAction: {\"action\": \"TYPE\", \"value\": \"acer nitro\", \"position\": null}<image>Thought: Type \"Acer Nitro 27\\\" class FHD gaming monitor\" into the Costco search bar to find the product.\nAction: {\"action\": \"TYPE\", \"value\": \"\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Think to find the Acer Nitro 27\" Class FHD Curved Gaming Monitor from the available search suggestions. Click on the item to view more details about it.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [255, 464]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_1.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_2.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_3.png"
] | 5,013
|
[
{
"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: Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor<image>Thought: Type \"Acer Nitro 27\\\" class FHD gaming monitor\" into the Costco search bar to find the product.\nAction: {\"action\": \"TYPE\", \"value\": \"\", \"position\": null}<image>Thought: Think to find the Acer Nitro 27\" Class FHD Curved Gaming Monitor from the available search suggestions. Click on the item to view more details about it.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [255, 464]}<image>",
"role": "user"
},
{
"content": "Thought: The task is complete as the user successfully searched for \"Acer Nitro 27\" class FHD gaming monitor.\"\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_2.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_3.png",
"training_data/AITW/aitw_images/single/13937512818881013854.591-737.Go to costco search bar and search Acer Nitro 27\" class FHD gaming monitor_4.png"
] | 5,014
|
[
{
"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: clear bestbuy search bar<image>",
"role": "user"
},
{
"content": "Thought: Click on the search bar to highlight the text for clearing. Type backspace to delete \"logitech g pro\" from the search bar. Press enter to confirm the empty search query and refresh results. Status task complete as the search bar is cleared successfully.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [305, 207]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2342604083756439960.675-720.clear bestbuy search bar_0.png"
] | 5,015
|
[
{
"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: clear bestbuy search bar<image>Thought: Click on the search bar to highlight the text for clearing. Type backspace to delete \"logitech g pro\" from the search bar. Press enter to confirm the empty search query and refresh results. Status task complete as the search bar is cleared successfully.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [305, 207]}<image>",
"role": "user"
},
{
"content": "Thought: Click the X to remove the text from the search bar and clear the search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [367, 163]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2342604083756439960.675-720.clear bestbuy search bar_0.png",
"training_data/AITW/aitw_images/single/2342604083756439960.675-720.clear bestbuy search bar_1.png"
] | 5,016
|
[
{
"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: clear bestbuy search bar<image>Thought: Click on the search bar to highlight the text for clearing. Type backspace to delete \"logitech g pro\" from the search bar. Press enter to confirm the empty search query and refresh results. Status task complete as the search bar is cleared successfully.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [305, 207]}<image>Thought: Click the X to remove the text from the search bar and clear the search results.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [367, 163]}<image>",
"role": "user"
},
{
"content": "Thought: You likely cleared the search bar to start a new search without previous suggestions interfering.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2342604083756439960.675-720.clear bestbuy search bar_0.png",
"training_data/AITW/aitw_images/single/2342604083756439960.675-720.clear bestbuy search bar_1.png",
"training_data/AITW/aitw_images/single/2342604083756439960.675-720.clear bestbuy search bar_2.png"
] | 5,017
|
[
{
"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 google search bar<image>",
"role": "user"
},
{
"content": "Thought: Click GOOGLE to access the app for performing searches and using the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [114, 1005]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/11298295673199284589.117-268.Open google search bar_0.png"
] | 5,018
|
[
{
"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 google search bar<image>Thought: Click GOOGLE to access the app for performing searches and using the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [114, 1005]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully opened the Google search bar, as indicated by the displayed search field at the top.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/11298295673199284589.117-268.Open google search bar_0.png",
"training_data/AITW/aitw_images/single/11298295673199284589.117-268.Open google search bar_1.png"
] | 5,019
|
[
{
"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 acer predator on amazon<image>",
"role": "user"
},
{
"content": "Thought: Type \"acer predator\" in the search bar to initiate a search for that item on Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"acer predator\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4890081884448054684.299-318.search acer predator on amazon_0.png"
] | 5,020
|
[
{
"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 acer predator on amazon<image>Thought: Type \"acer predator\" in the search bar to initiate a search for that item on Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"acer predator\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully searched for \"acer predator\" using the search bar on Amazon.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4890081884448054684.299-318.search acer predator on amazon_0.png",
"training_data/AITW/aitw_images/single/4890081884448054684.299-318.search acer predator on amazon_1.png"
] | 5,021
|
[
{
"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: Enter in the chrome search bar: amazon.com and go to the amazon website<image>",
"role": "user"
},
{
"content": "Thought: Type \"amazon.com\" in the search bar to navigate directly to the Amazon website.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon.com\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_0.png"
] | 5,022
|
[
{
"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: Enter in the chrome search bar: amazon.com and go to the amazon website<image>Thought: Type \"amazon.com\" in the search bar to navigate directly to the Amazon website.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon.com\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"amazon.com\" in the search results to access the website directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [190, 239]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_0.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_1.png"
] | 5,023
|
[
{
"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: Enter in the chrome search bar: amazon.com and go to the amazon website<image>Thought: Type \"amazon.com\" in the search bar to navigate directly to the Amazon website.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon.com\", \"position\": null}<image>Thought: Click on \"amazon.com\" in the search results to access the website directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [190, 239]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view the search results for amazon.com beneath the search bar.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_0.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_1.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_2.png"
] | 5,024
|
[
{
"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: Enter in the chrome search bar: amazon.com and go to the amazon website<image>Thought: Click on \"amazon.com\" in the search results to access the website directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [190, 239]}<image>Thought: Scroll down to view the search results for amazon.com beneath the search bar.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the link labeled \"Amazon.com\" to access the website.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [261, 861]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_1.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_2.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_3.png"
] | 5,025
|
[
{
"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: Enter in the chrome search bar: amazon.com and go to the amazon website<image>Thought: Scroll down to view the search results for amazon.com beneath the search bar.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click the link labeled \"Amazon.com\" to access the website.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [261, 861]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"amazon.com\" in the Chrome search bar to initiate a website visit. Press enter to load the Amazon website successfully. Status task complete confirms that the navigation to amazon.com was successful.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_2.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_3.png",
"training_data/AITW/aitw_images/single/6051526963343414946.186-349.Enter in the chrome search bar: amazon.com and go to the amazon website_4.png"
] | 5,026
|
[
{
"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: go to costco cart<image>",
"role": "user"
},
{
"content": "Thought: Click on the Cart icon to access your shopping cart and view items.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [688, 214]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/5945666479123179296.517-653.go to costco cart_0.png"
] | 5,027
|
[
{
"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: go to costco cart<image>Thought: Click on the Cart icon to access your shopping cart and view items.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [688, 214]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully accessed the Costco cart page to check your shopping items.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/5945666479123179296.517-653.go to costco cart_0.png",
"training_data/AITW/aitw_images/single/5945666479123179296.517-653.go to costco cart_1.png"
] | 5,028
|
[
{
"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: Go to the search bar on target and search for usb- c<image>",
"role": "user"
},
{
"content": "Thought: Click the magnifying glass to access the search functionality on the Target mobile website.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [482, 293]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_0.png"
] | 5,029
|
[
{
"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: Go to the search bar on target and search for usb- c<image>Thought: Click the magnifying glass to access the search functionality on the Target mobile website.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [482, 293]}<image>",
"role": "user"
},
{
"content": "Thought: Click the Search bar to prepare for entering the search term \"usb-c.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [82, 145]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_0.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_1.png"
] | 5,030
|
[
{
"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: Go to the search bar on target and search for usb- c<image>Thought: Click the magnifying glass to access the search functionality on the Target mobile website.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [482, 293]}<image>Thought: Click the Search bar to prepare for entering the search term \"usb-c.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [82, 145]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"usb-c\" in the search bar to initiate a product search on Target's website.\nAction: {\"action\": \"TYPE\", \"value\": \"usb-c\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_0.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_1.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_2.png"
] | 5,031
|
[
{
"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: Go to the search bar on target and search for usb- c<image>Thought: Click the Search bar to prepare for entering the search term \"usb-c.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [82, 145]}<image>Thought: Type \"usb-c\" in the search bar to initiate a product search on Target's website.\nAction: {\"action\": \"TYPE\", \"value\": \"usb-c\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the magnifying glass to initiate the search function for \"usb-c\" in the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [414, 156]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_1.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_2.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_3.png"
] | 5,032
|
[
{
"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: Go to the search bar on target and search for usb- c<image>Thought: Type \"usb-c\" in the search bar to initiate a product search on Target's website.\nAction: {\"action\": \"TYPE\", \"value\": \"usb-c\", \"position\": null}<image>Thought: Click the magnifying glass to initiate the search function for \"usb-c\" in the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [414, 156]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully located the search bar and entered \"usb-c\" to find products.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_2.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_3.png",
"training_data/AITW/aitw_images/single/90201725313951350.765-965.Go to the search bar on target and search for usb- c_4.png"
] | 5,033
|
[
{
"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: go to amazon search bar and type razer huntsman <image>",
"role": "user"
},
{
"content": "Thought: Click the Search Amazon bar to focus on it for typing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [199, 281]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/202222001745002420.273-297.go to amazon search bar and type razer huntsman _0.png"
] | 5,034
|
[
{
"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: go to amazon search bar and type razer huntsman <image>Thought: Click the Search Amazon bar to focus on it for typing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [199, 281]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"razer huntsman\" in the search Amazon bar to find the desired product.\nAction: {\"action\": \"TYPE\", \"value\": \"razer huntsman\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/202222001745002420.273-297.go to amazon search bar and type razer huntsman _0.png",
"training_data/AITW/aitw_images/single/202222001745002420.273-297.go to amazon search bar and type razer huntsman _1.png"
] | 5,035
|
[
{
"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: go to amazon search bar and type razer huntsman <image>Thought: Click the Search Amazon bar to focus on it for typing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [199, 281]}<image>Thought: Type \"razer huntsman\" in the search Amazon bar to find the desired product.\nAction: {\"action\": \"TYPE\", \"value\": \"razer huntsman\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Focus on the search bar to type \"razer huntsman\" for the desired product. After entering the text, recognize a dropdown with relevant suggestions appears. Confirm that you finished entering your search terms by checking that \"razer huntsman\" is displayed correctly in the search bar, signaling that the task is complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/202222001745002420.273-297.go to amazon search bar and type razer huntsman _0.png",
"training_data/AITW/aitw_images/single/202222001745002420.273-297.go to amazon search bar and type razer huntsman _1.png",
"training_data/AITW/aitw_images/single/202222001745002420.273-297.go to amazon search bar and type razer huntsman _2.png"
] | 5,036
|
[
{
"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 Duracell Coppertop AAA Batteries 20 Pack Alkaline Battery in target.com<image>",
"role": "user"
},
{
"content": "Thought: Click the \"Duracell Coppertop AAA\" option to access the product details and complete your search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [79, 490]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15125722558828666538.855-898.open Duracell Coppertop AAA Batteries 20 Pack Alkaline Battery in target.com_0.png"
] | 5,037
|
[
{
"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 Duracell Coppertop AAA Batteries 20 Pack Alkaline Battery in target.com<image>Thought: Click the \"Duracell Coppertop AAA\" option to access the product details and complete your search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [79, 490]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully accessed the product page for Duracell Coppertop AAA Batteries on target.com.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15125722558828666538.855-898.open Duracell Coppertop AAA Batteries 20 Pack Alkaline Battery in target.com_0.png",
"training_data/AITW/aitw_images/single/15125722558828666538.855-898.open Duracell Coppertop AAA Batteries 20 Pack Alkaline Battery in target.com_1.png"
] | 5,038
|
[
{
"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: Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart <image>",
"role": "user"
},
{
"content": "Thought: Click on Gillette Fusion5 to access the product details and proceed with adding it to the cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [453, 1288]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _0.png"
] | 5,039
|
[
{
"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: Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart <image>Thought: Click on Gillette Fusion5 to access the product details and proceed with adding it to the cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [453, 1288]}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to locate the Gillette Fusion5 Men's Razor Blade Refills on the webpage.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _0.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _1.png"
] | 5,040
|
[
{
"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: Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart <image>Thought: Click on Gillette Fusion5 to access the product details and proceed with adding it to the cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [453, 1288]}<image>Thought: Scroll down to locate the Gillette Fusion5 Men's Razor Blade Refills on the webpage.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view the Gillette Fusion5 Men's Razor Blade Refills product details and add it to the cart.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _0.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _1.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _2.png"
] | 5,041
|
[
{
"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: Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart <image>Thought: Scroll down to locate the Gillette Fusion5 Men's Razor Blade Refills on the webpage.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Scroll down to view the Gillette Fusion5 Men's Razor Blade Refills product details and add it to the cart.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the red \"Add to cart\" button to proceed with adding the Gillette Fusion5 Men's Razor Blade Refills.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [466, 1060]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _1.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _2.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _3.png"
] | 5,042
|
[
{
"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: Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart <image>Thought: Scroll down to view the Gillette Fusion5 Men's Razor Blade Refills product details and add it to the cart.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click the red \"Add to cart\" button to proceed with adding the Gillette Fusion5 Men's Razor Blade Refills.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [466, 1060]}<image>",
"role": "user"
},
{
"content": "Thought: Click the \"Add to cart\" button to select the Gillette Fusion5 Men's Razor Blade Refills for purchase.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [482, 1063]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _2.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _3.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _4.png"
] | 5,043
|
[
{
"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: Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart <image>Thought: Click the red \"Add to cart\" button to proceed with adding the Gillette Fusion5 Men's Razor Blade Refills.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [466, 1060]}<image>Thought: Click the \"Add to cart\" button to select the Gillette Fusion5 Men's Razor Blade Refills for purchase.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [482, 1063]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully added the Gillette Fusion5 Men's Razor Blade Refills to your cart, confirming the task is complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _3.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _4.png",
"training_data/AITW/aitw_images/single/13641155417452696124.289-445.Go to Gillette Fusion5 Men's Razor Blade Refills then add this item to cart _5.png"
] | 5,044
|
[
{
"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: Show my apps and open chrome<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to reveal the app icons on the screen for easier access.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4066054430731247035.40-264.Show my apps and open chrome_0.png"
] | 5,045
|
[
{
"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: Show my apps and open chrome<image>Thought: Scroll down to reveal the app icons on the screen for easier access.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the Chrome icon to open the browser app directly from the apps screen.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [353, 517]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4066054430731247035.40-264.Show my apps and open chrome_0.png",
"training_data/AITW/aitw_images/single/4066054430731247035.40-264.Show my apps and open chrome_1.png"
] | 5,046
|
[
{
"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: Show my apps and open chrome<image>Thought: Scroll down to reveal the app icons on the screen for easier access.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click on the Chrome icon to open the browser app directly from the apps screen.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [353, 517]}<image>",
"role": "user"
},
{
"content": "Thought: You accessed your apps and chose Chrome to browse the web, completing your task successfully.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4066054430731247035.40-264.Show my apps and open chrome_0.png",
"training_data/AITW/aitw_images/single/4066054430731247035.40-264.Show my apps and open chrome_1.png",
"training_data/AITW/aitw_images/single/4066054430731247035.40-264.Show my apps and open chrome_2.png"
] | 5,047
|
[
{
"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 first result of asus zenbook in amazon.com<image>",
"role": "user"
},
{
"content": "Thought: Click the first result to view details about the ASUS ZenBook product.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [372, 605]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6517601905855105972.595-654.open the first result of asus zenbook in amazon.com_0.png"
] | 5,048
|
[
{
"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 first result of asus zenbook in amazon.com<image>Thought: Click the first result to view details about the ASUS ZenBook product.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [372, 605]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully opened the first result for \"asus zenbook\" on Amazon, confirming the task is complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6517601905855105972.595-654.open the first result of asus zenbook in amazon.com_0.png",
"training_data/AITW/aitw_images/single/6517601905855105972.595-654.open the first result of asus zenbook in amazon.com_1.png"
] | 5,049
|
[
{
"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: Go to chrome search bar and search & open amazon <image>",
"role": "user"
},
{
"content": "Thought: Click the search bar at the top of the screen to prepare for a new search. Type 'amazon' to specify your intended destination clearly. Press enter to execute your search for Amazon's website, intending to open it in Chrome.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [330, 63]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _0.png"
] | 5,050
|
[
{
"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: Go to chrome search bar and search & open amazon <image>Thought: Click the search bar at the top of the screen to prepare for a new search. Type 'amazon' to specify your intended destination clearly. Press enter to execute your search for Amazon's website, intending to open it in Chrome.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [330, 63]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"amazon\" in the search bar to initiate a web search for Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _0.png",
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _1.png"
] | 5,051
|
[
{
"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: Go to chrome search bar and search & open amazon <image>Thought: Click the search bar at the top of the screen to prepare for a new search. Type 'amazon' to specify your intended destination clearly. Press enter to execute your search for Amazon's website, intending to open it in Chrome.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [330, 63]}<image>Thought: Type \"amazon\" in the search bar to initiate a web search for Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to execute the search for Amazon after typing in the search bar.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _0.png",
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _1.png",
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _2.png"
] | 5,052
|
[
{
"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: Go to chrome search bar and search & open amazon <image>Thought: Type \"amazon\" in the search bar to initiate a web search for Amazon.\nAction: {\"action\": \"TYPE\", \"value\": \"amazon\", \"position\": null}<image>Thought: Press enter to execute the search for Amazon after typing in the search bar.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Access the Chrome search bar to initiate a web search for Amazon. Enter \"Amazon\" in the search field and submit the query. Once the results load, select the Amazon link to access the site. After reviewing, confirm that you've completed your task successfully by checking if you're on Amazon's homepage.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _1.png",
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _2.png",
"training_data/AITW/aitw_images/single/3410773500196879638.157-244.Go to chrome search bar and search & open amazon _3.png"
] | 5,053
|
[
{
"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: Go to bestbuy search bar and search macbook air<image>",
"role": "user"
},
{
"content": "Thought: Click on the search bar to begin entering the query for \"MacBook Air\".\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [130, 90]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_0.png"
] | 5,054
|
[
{
"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: Go to bestbuy search bar and search macbook air<image>Thought: Click on the search bar to begin entering the query for \"MacBook Air\".\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [130, 90]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"macbook air\" in the search bar to initiate a product search on Best Buy.\nAction: {\"action\": \"TYPE\", \"value\": \" macbook air \", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_0.png",
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_1.png"
] | 5,055
|
[
{
"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: Go to bestbuy search bar and search macbook air<image>Thought: Click on the search bar to begin entering the query for \"MacBook Air\".\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [130, 90]}<image>Thought: Type \"macbook air\" in the search bar to initiate a product search on Best Buy.\nAction: {\"action\": \"TYPE\", \"value\": \" macbook air \", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on the search bar to enter your query for \"macbook air\". Type \"macbook air\" to initiate the search for results. Press enter to execute the search and view available options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [54, 102]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_0.png",
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_1.png",
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_2.png"
] | 5,056
|
[
{
"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: Go to bestbuy search bar and search macbook air<image>Thought: Type \"macbook air\" in the search bar to initiate a product search on Best Buy.\nAction: {\"action\": \"TYPE\", \"value\": \" macbook air \", \"position\": null}<image>Thought: Click on the search bar to enter your query for \"macbook air\". Type \"macbook air\" to initiate the search for results. Press enter to execute the search and view available options.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [54, 102]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully accessed the search results for \"macbook air\" in the Best Buy app.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_1.png",
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_2.png",
"training_data/AITW/aitw_images/single/9774042394394560011.1378-1835.Go to bestbuy search bar and search macbook air_3.png"
] | 5,057
|
[
{
"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 dell xps on amazon.com then open third product from search results on amazon.com<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Basics\" to explore product categories available on Amazon for a more focused search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 240]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_0.png"
] | 5,058
|
[
{
"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 dell xps on amazon.com then open third product from search results on amazon.com<image>Thought: Click on \"Basics\" to explore product categories available on Amazon for a more focused search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 240]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"dell xps\" in the Amazon search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"dell xps\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_0.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_1.png"
] | 5,059
|
[
{
"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 dell xps on amazon.com then open third product from search results on amazon.com<image>Thought: Click on \"Basics\" to explore product categories available on Amazon for a more focused search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 240]}<image>Thought: Type \"dell xps\" in the Amazon search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"dell xps\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to execute the search for \"dell xps\" on Amazon.com and view the results.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_0.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_1.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_2.png"
] | 5,060
|
[
{
"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 dell xps on amazon.com then open third product from search results on amazon.com<image>Thought: Type \"dell xps\" in the Amazon search bar to find relevant products.\nAction: {\"action\": \"TYPE\", \"value\": \"dell xps\", \"position\": null}<image>Thought: Press enter to execute the search for \"dell xps\" on Amazon.com and view the results.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Scroll down to view more products in the search results for 'dell xps'.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_1.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_2.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_3.png"
] | 5,061
|
[
{
"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 dell xps on amazon.com then open third product from search results on amazon.com<image>Thought: Press enter to execute the search for \"dell xps\" on Amazon.com and view the results.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>Thought: Scroll down to view more products in the search results for 'dell xps'.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on 8GB to narrow down the laptop specifications and highlight relevant options for my search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 980]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_2.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_3.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_4.png"
] | 5,062
|
[
{
"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 dell xps on amazon.com then open third product from search results on amazon.com<image>Thought: Scroll down to view more products in the search results for 'dell xps'.\nAction: {\"action\": \"SCROLL DOWN\", \"value\": null, \"position\": null}<image>Thought: Click on 8GB to narrow down the laptop specifications and highlight relevant options for my search.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [505, 980]}<image>",
"role": "user"
},
{
"content": "Thought: The task is complete as the user successfully searched for \"dell xps\" and opened the desired product listing.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_3.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_4.png",
"training_data/AITW/aitw_images/single/2981609658862329662.254-436.Search for dell xps on amazon.com then open third product from search results on amazon.com_5.png"
] | 5,063
|
[
{
"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: Add first product from results to the cart <image>",
"role": "user"
},
{
"content": "Thought: Click \"Add to Cart\" on the PNY product to include it in my shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [499, 1201]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/10811287213745751856.832-884.Add first product from results to the cart _0.png"
] | 5,064
|
[
{
"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: Add first product from results to the cart <image>Thought: Click \"Add to Cart\" on the PNY product to include it in my shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [499, 1201]}<image>",
"role": "user"
},
{
"content": "Thought: The operation indicates that the product has been successfully added to the cart, confirming task completion.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/10811287213745751856.832-884.Add first product from results to the cart _0.png",
"training_data/AITW/aitw_images/single/10811287213745751856.832-884.Add first product from results to the cart _1.png"
] | 5,065
|
[
{
"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 chrome<image>",
"role": "user"
},
{
"content": "Thought: Click the CHAT icon to access apps where Chrome might be listed.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [164, 485]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3642686345159404511.2255-2295.Open chrome_0.png"
] | 5,066
|
[
{
"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 chrome<image>Thought: Click the CHAT icon to access apps where Chrome might be listed.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [164, 485]}<image>",
"role": "user"
},
{
"content": "Thought: You would open Chrome to access the browser, and now the task is marked complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3642686345159404511.2255-2295.Open chrome_0.png",
"training_data/AITW/aitw_images/single/3642686345159404511.2255-2295.Open chrome_1.png"
] | 5,067
|
[
{
"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 Google and go to the search bar on Google and clear the search bar<image>",
"role": "user"
},
{
"content": "Thought: Click on the Google icon to open the application. Once in Google, locate the search bar to interact with it for clearing any existing text.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [41, 354]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_0.png"
] | 5,068
|
[
{
"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 Google and go to the search bar on Google and clear the search bar<image>Thought: Click on the Google icon to open the application. Once in Google, locate the search bar to interact with it for clearing any existing text.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [41, 354]}<image>",
"role": "user"
},
{
"content": "Thought: Click the search bar to focus on it and prepare for clearing any existing text.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [109, 106]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_0.png",
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_1.png"
] | 5,069
|
[
{
"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 Google and go to the search bar on Google and clear the search bar<image>Thought: Click on the Google icon to open the application. Once in Google, locate the search bar to interact with it for clearing any existing text.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [41, 354]}<image>Thought: Click the search bar to focus on it and prepare for clearing any existing text.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [109, 106]}<image>",
"role": "user"
},
{
"content": "Thought: Tap the X icon in the search bar to clear the existing search query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [244, 28]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_0.png",
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_1.png",
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_2.png"
] | 5,070
|
[
{
"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 Google and go to the search bar on Google and clear the search bar<image>Thought: Click the search bar to focus on it and prepare for clearing any existing text.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [109, 106]}<image>Thought: Tap the X icon in the search bar to clear the existing search query.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [244, 28]}<image>",
"role": "user"
},
{
"content": "Thought: Open Google to access the search interface and display the existing search bar contents. Tap on the search bar to activate it for editing. Use the keyboard's clear function to eliminate any previous input in the search bar. Confirm that the content is cleared and indicate that the task is complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_1.png",
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_2.png",
"training_data/AITW/aitw_images/single/15189390281239989831.320-603.Open Google and go to the search bar on Google and clear the search bar_3.png"
] | 5,071
|
[
{
"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: Go to bestbuy<image>",
"role": "user"
},
{
"content": "Thought: Press enter to execute the search for \"bestbuy\" and retrieve relevant results.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2061996763641510551.218-264.Go to bestbuy_0.png"
] | 5,072
|
[
{
"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: Go to bestbuy<image>Thought: Press enter to execute the search for \"bestbuy\" and retrieve relevant results.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: The user successfully accessed Best Buy's website from the search results. Completing the task demonstrates they wanted to visit Best Buy.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2061996763641510551.218-264.Go to bestbuy_0.png",
"training_data/AITW/aitw_images/single/2061996763641510551.218-264.Go to bestbuy_1.png"
] | 5,073
|
[
{
"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 razer blackwidow in the amazon search bar<image>",
"role": "user"
},
{
"content": "Thought: Click Basics to access a specific category that may list the Razer BlackWidow items.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [95, 103]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_0.png"
] | 5,074
|
[
{
"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 razer blackwidow in the amazon search bar<image>Thought: Click Basics to access a specific category that may list the Razer BlackWidow items.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [95, 103]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"razer blackwidow\" in the Amazon search bar to find the product.\nAction: {\"action\": \"TYPE\", \"value\": \"razer blackwidow\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_0.png",
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_1.png"
] | 5,075
|
[
{
"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 razer blackwidow in the amazon search bar<image>Thought: Click Basics to access a specific category that may list the Razer BlackWidow items.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [95, 103]}<image>Thought: Type \"razer blackwidow\" in the Amazon search bar to find the product.\nAction: {\"action\": \"TYPE\", \"value\": \"razer blackwidow\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"razer blackwicow\" to select the item from the search suggestions in the dropdown menu.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [83, 118]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_0.png",
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_1.png",
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_2.png"
] | 5,076
|
[
{
"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 razer blackwidow in the amazon search bar<image>Thought: Type \"razer blackwidow\" in the Amazon search bar to find the product.\nAction: {\"action\": \"TYPE\", \"value\": \"razer blackwidow\", \"position\": null}<image>Thought: Click on \"razer blackwicow\" to select the item from the search suggestions in the dropdown menu.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [83, 118]}<image>",
"role": "user"
},
{
"content": "Thought: You searched for \"razer blackwidow\" in the Amazon search bar. The results displayed successfully confirmed that the search was completed.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_1.png",
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_2.png",
"training_data/AITW/aitw_images/single/16612232925580199452.816-1293.search for razer blackwidow in the amazon search bar_3.png"
] | 5,077
|
[
{
"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 dell xps 13 in the walmart search bar<image>",
"role": "user"
},
{
"content": "Thought: Click the search bar to edit the current search term and input \"dell xps 13.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [364, 161]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_0.png"
] | 5,078
|
[
{
"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 dell xps 13 in the walmart search bar<image>Thought: Click the search bar to edit the current search term and input \"dell xps 13.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [364, 161]}<image>",
"role": "user"
},
{
"content": "Thought: Type \"dell xps\" in the Walmart search bar to find the specific laptop model.\nAction: {\"action\": \"TYPE\", \"value\": \"dell xps\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_0.png",
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_1.png"
] | 5,079
|
[
{
"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 dell xps 13 in the walmart search bar<image>Thought: Click the search bar to edit the current search term and input \"dell xps 13.\"\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [364, 161]}<image>Thought: Type \"dell xps\" in the Walmart search bar to find the specific laptop model.\nAction: {\"action\": \"TYPE\", \"value\": \"dell xps\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click in the search bar to focus on it and prepare for typing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [134, 325]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_0.png",
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_1.png",
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_2.png"
] | 5,080
|
[
{
"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 dell xps 13 in the walmart search bar<image>Thought: Type \"dell xps\" in the Walmart search bar to find the specific laptop model.\nAction: {\"action\": \"TYPE\", \"value\": \"dell xps\", \"position\": null}<image>Thought: Click in the search bar to focus on it and prepare for typing.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [134, 325]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully searched for \"dell xps 13\" in the Walmart search bar, and results are displayed.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_1.png",
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_2.png",
"training_data/AITW/aitw_images/single/15773043732842599476.211-302.search for dell xps 13 in the walmart search bar_3.png"
] | 5,081
|
[
{
"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: Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it<image>",
"role": "user"
},
{
"content": "Thought: You clicked on a link instead of accessing the search bar. You need to return to the home screen first. Then, open Chrome and find the search bar. After that, type in \"bestbuy.com\" to see suggestions. Choose bestbuy.com from the list of search suggestions once it appears. Finally, click on it to open the website. This sequence is necessary since you are currently on a different webpage and can't access your intended destination directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [270, 102]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/7942862312392961691.112-185.Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it_0.png"
] | 5,082
|
[
{
"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: Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it<image>Thought: You clicked on a link instead of accessing the search bar. You need to return to the home screen first. Then, open Chrome and find the search bar. After that, type in \"bestbuy.com\" to see suggestions. Choose bestbuy.com from the list of search suggestions once it appears. Finally, click on it to open the website. This sequence is necessary since you are currently on a different webpage and can't access your intended destination directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [270, 102]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Best Buy\" in the search suggestions to open the website. This is necessary to access Best Buy’s page directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [239, 628]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/7942862312392961691.112-185.Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it_0.png",
"training_data/AITW/aitw_images/single/7942862312392961691.112-185.Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it_1.png"
] | 5,083
|
[
{
"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: Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it<image>Thought: You clicked on a link instead of accessing the search bar. You need to return to the home screen first. Then, open Chrome and find the search bar. After that, type in \"bestbuy.com\" to see suggestions. Choose bestbuy.com from the list of search suggestions once it appears. Finally, click on it to open the website. This sequence is necessary since you are currently on a different webpage and can't access your intended destination directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [270, 102]}<image>Thought: Click on \"Best Buy\" in the search suggestions to open the website. This is necessary to access Best Buy’s page directly.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [239, 628]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully navigated to the search bar in Chrome and selected bestbuy.com from the suggestions.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/7942862312392961691.112-185.Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it_0.png",
"training_data/AITW/aitw_images/single/7942862312392961691.112-185.Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it_1.png",
"training_data/AITW/aitw_images/single/7942862312392961691.112-185.Go to the search bar of Chrome, then choose bestbuy.com from the search suggestion and open it_2.png"
] | 5,084
|
[
{
"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: Go to bestbuy search bar<image>",
"role": "user"
},
{
"content": "Thought: Click the \"Deal\" button to access deals and discounts, possibly leading to the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [451, 262]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6085570407850972074.844-973.Go to bestbuy search bar_0.png"
] | 5,085
|
[
{
"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: Go to bestbuy search bar<image>Thought: Click the \"Deal\" button to access deals and discounts, possibly leading to the search bar.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [451, 262]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully accessed the search bar at the top of the Best Buy website.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/6085570407850972074.844-973.Go to bestbuy search bar_0.png",
"training_data/AITW/aitw_images/single/6085570407850972074.844-973.Go to bestbuy search bar_1.png"
] | 5,086
|
[
{
"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: Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart<image>",
"role": "user"
},
{
"content": "Thought: Type \"ebay.com\" in the search bar to navigate to eBay's website. Click on the link for the eBay website to open it. Look for the cart icon or option on the webpage and click on it to view your cart items. Locate the ASUS - Zenbook 14' item in your cart. Finally, select the option to remove this item from your cart.\nAction: {\"action\": \"TYPE\", \"value\": \"ebay.com\", \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_0.png"
] | 5,087
|
[
{
"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: Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart<image>Thought: Type \"ebay.com\" in the search bar to navigate to eBay's website. Click on the link for the eBay website to open it. Look for the cart icon or option on the webpage and click on it to view your cart items. Locate the ASUS - Zenbook 14' item in your cart. Finally, select the option to remove this item from your cart.\nAction: {\"action\": \"TYPE\", \"value\": \"ebay.com\", \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Press enter to navigate to the eBay website and access my shopping cart.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_0.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_1.png"
] | 5,088
|
[
{
"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: Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart<image>Thought: Type \"ebay.com\" in the search bar to navigate to eBay's website. Click on the link for the eBay website to open it. Look for the cart icon or option on the webpage and click on it to view your cart items. Locate the ASUS - Zenbook 14' item in your cart. Finally, select the option to remove this item from your cart.\nAction: {\"action\": \"TYPE\", \"value\": \"ebay.com\", \"position\": null}<image>Thought: Press enter to navigate to the eBay website and access my shopping cart.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>",
"role": "user"
},
{
"content": "Thought: Click the SHOPPING CART icon to access your saved items for easier management.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [612, 187]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_0.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_1.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_2.png"
] | 5,089
|
[
{
"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: Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart<image>Thought: Press enter to navigate to the eBay website and access my shopping cart.\nAction: {\"action\": \"PRESS ENTER\", \"value\": null, \"position\": null}<image>Thought: Click the SHOPPING CART icon to access your saved items for easier management.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [612, 187]}<image>",
"role": "user"
},
{
"content": "Thought: You click \"Remove\" to delete the ASUS - Zenbook 14' from your shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [652, 789]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_1.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_2.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_3.png"
] | 5,090
|
[
{
"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: Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart<image>Thought: Click the SHOPPING CART icon to access your saved items for easier management.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [612, 187]}<image>Thought: You click \"Remove\" to delete the ASUS - Zenbook 14' from your shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [652, 789]}<image>",
"role": "user"
},
{
"content": "Thought: Click Remove to confirm the action of removing the ASUS - Zenbook from the shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [653, 792]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_2.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_3.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_4.png"
] | 5,091
|
[
{
"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: Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart<image>Thought: You click \"Remove\" to delete the ASUS - Zenbook 14' from your shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [652, 789]}<image>Thought: Click Remove to confirm the action of removing the ASUS - Zenbook from the shopping cart.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [653, 792]}<image>",
"role": "user"
},
{
"content": "Thought: You successfully typed ebay and accessed the website. You then viewed your cart to check for items. Upon finding the ASUS - Zenbook 14' listed, you removed it to complete your task of clearing the cart. The status indicates that the task is now complete with no items left in your cart.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_3.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_4.png",
"training_data/AITW/aitw_images/single/4870230774127142814.325-1133.Type ebay and go to the ebay website, then show the cart and remove the ASUS - Zenbook 14' from the cart_5.png"
] | 5,092
|
[
{
"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: Add item to cart<image>",
"role": "user"
},
{
"content": "Thought: Click the \"Add to cart\" button to include the item in your shopping cart for future purchase.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [359, 946]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4975014716477036792.707-735.Add item to cart_0.png"
] | 5,093
|
[
{
"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: Add item to cart<image>Thought: Click the \"Add to cart\" button to include the item in your shopping cart for future purchase.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [359, 946]}<image>",
"role": "user"
},
{
"content": "Thought: You have successfully added the item to your cart as shown by the status task complete.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/4975014716477036792.707-735.Add item to cart_0.png",
"training_data/AITW/aitw_images/single/4975014716477036792.707-735.Add item to cart_1.png"
] | 5,094
|
[
{
"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: go to acer nitro and open first search result <image>",
"role": "user"
},
{
"content": "Thought: Click \"acer nitro\" to select the first search result from the displayed list.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [103, 544]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3885375936410226917.501-544.go to acer nitro and open first search result _0.png"
] | 5,095
|
[
{
"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: go to acer nitro and open first search result <image>Thought: Click \"acer nitro\" to select the first search result from the displayed list.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [103, 544]}<image>",
"role": "user"
},
{
"content": "Thought: Click on \"Win 11 Laptop\" to open the first search result for Acer Nitro.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [393, 674]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3885375936410226917.501-544.go to acer nitro and open first search result _0.png",
"training_data/AITW/aitw_images/single/3885375936410226917.501-544.go to acer nitro and open first search result _1.png"
] | 5,096
|
[
{
"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: go to acer nitro and open first search result <image>Thought: Click \"acer nitro\" to select the first search result from the displayed list.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [103, 544]}<image>Thought: Click on \"Win 11 Laptop\" to open the first search result for Acer Nitro.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [393, 674]}<image>",
"role": "user"
},
{
"content": "Thought: You completed the task of finding the Acer Nitro by successfully opening the first search result on the screen.\nAction: {\"action\": \"STATUS TASK COMPLETE\", \"value\": null, \"position\": null}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/3885375936410226917.501-544.go to acer nitro and open first search result _0.png",
"training_data/AITW/aitw_images/single/3885375936410226917.501-544.go to acer nitro and open first search result _1.png",
"training_data/AITW/aitw_images/single/3885375936410226917.501-544.go to acer nitro and open first search result _2.png"
] | 5,097
|
[
{
"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 Chrome app and search bestbuy.com in chrome search bar<image>",
"role": "user"
},
{
"content": "Thought: Click the Chrome icon to launch the app from the home screen.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [366, 913]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2342604083756439960.138-402.Open Chrome app and search bestbuy.com in chrome search bar_0.png"
] | 5,098
|
[
{
"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 Chrome app and search bestbuy.com in chrome search bar<image>Thought: Click the Chrome icon to launch the app from the home screen.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [366, 913]}<image>",
"role": "user"
},
{
"content": "Thought: You mistakenly clicked the link to Costco's catalog instead of opening the Chrome app. You need to return to your home screen or re-open Chrome.\nAction: {\"action\": \"CLICK\", \"value\": null, \"position\": [230, 85]}",
"role": "assistant"
}
] |
[
"training_data/AITW/aitw_images/single/2342604083756439960.138-402.Open Chrome app and search bestbuy.com in chrome search bar_0.png",
"training_data/AITW/aitw_images/single/2342604083756439960.138-402.Open Chrome app and search bestbuy.com in chrome search bar_1.png"
] | 5,099
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.