Datasets:

---

neulab

/

Mind2Web_train_llava

like 2

Follow

NeuLab @ LTI/CMU 86

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find flights from Chicago to London on 20 April and return on 23 April.\nPrevious actions:\n[a] LON - London, United Kingdom -> CLICK\n[button] 􎀺 View Calendar: To navigate the calendar, use the... -> CLICK\n[link] 20 -> CLICK\n[button] 􎀺 View Calendar: To navigate the calendar, use the... -> CLICK\n[link] 23 -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.636, 0.226, 0.875, 0.246] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book an hourly parking at Atlanta International Airport.\nPrevious actions:\n[tab] Hourly -> CLICK\n[textbox] Search for parking -> TYPE: Atlanta International Airport" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.14, 0.29, 0.86, 0.311] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find an foldable over-ear headphone under 35 USD in certified refurbished condition with lowest price + shipping.\nPrevious actions:\n[link] Electronics -> HOVER\n[link] Headphones -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.194, 0.268, 0.35, 0.317] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Make a playlist and add 3 songs from popular Selena Gomez tracks and name it Love.\nPrevious actions:\n[a] Create a playlist -> CLICK\n[textbox] Playlist title * -> TYPE: Love\n[button] Save -> CLICK\n[button] Close -> CLICK\n[button] More -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.402, 0.482, 0.512, 0.511] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Browse the on demand sales section and rent a film in HD format.\nPrevious actions:\n[link] Visit the On Demand page -> HOVER" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.285, 0.127, 0.488, 0.142] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Play the top track for the top indie artist in the last 30 days\nPrevious actions:\n[link] Music -> CLICK\n[link] indie -> CLICK\n[link] The Smiths -> CLICK\n[button] Sorted by: Last 7 days -> CLICK\n[link] Last 30 days -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.079, 0.276, 0.104, 0.295] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Reserve a lowest priced car for pickup from Birmingham Airport, Alabama, at 11 am on March 22, 2023, and the return location is Montogomery Airport, Alabama, at 5 pm on March 25, 2023.\nPrevious actions:\n[searchbox] Pick-up & Return Location (ZIP, City or Airport) (... -> TYPE: BIRMINGHAM\n[span] Birmingham -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.166, 0.204, 0.188, 0.217] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the schedule between New York and Washington DC sorted by shortest trip.\nPrevious actions:\n[textbox] To -> TYPE: Washington\n[option] Washington, DC - Union Station (WAS) -> CLICK\n[textbox] Depart Date -> CLICK\n[gridcell] Monday, April 17, 2023 -> CLICK\n[button] FIND TRAINS -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.874, 0.192, 0.93, 0.2] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find rental for an economy car in Houston from Mar 10 to Mar 13.\nPrevious actions:\n[button] Ellipsis Icon -> CLICK\n[link] Car rentals -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.143, 0.221, 0.469, 0.243] in the image\nAnd my action is Action: TYPE\nValue: Houston" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: search four door SUVs from Avis while traveling from New Orleans, L, to New York that isn't more than $1000 in total price from Mar 25th to Apr 1st, pickup at 2PM.\nPrevious actions:\n[span] New York, United States -> CLICK\n[generic] Saturday March 25th -> CLICK\n[button] Saturday March 25, 2023 -> CLICK\n[generic] Saturday April 1st -> CLICK\n[button] Saturday April 1, 2023 -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.695, 0.102, 0.74, 0.127] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find tortillas bread in grocery stock near zip 59901 and add 5 pieces of 12 pcs pack in the cart and view the cart.\nPrevious actions:\n[button] Departments -> HOVER\n[button] Grocery & Home -> HOVER\n[link] Grocery -> CLICK\n[link] Bread -> CLICK\n[label] In Stock Today -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.028, 0.459, 0.078, 0.473] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Check the availability of a Nintendo Switch gaming console at the nearest Target store.\nPrevious actions:\n[searchbox] Search: suggestions appear below -> TYPE: Nintendo Switch\n[link] nintendo switch -> CLICK\n[img] Nintendo Switch with Neon Blue and Neon Red Joy-Co... -> CLICK\n[button] pickup - unselected - 1 of 3 - Ready within 2 hour... -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.874, 0.202, 0.97, 0.21] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the flight schedule between London and New York for April 7 and check the upgrade details of the noon flight.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.167, 0.107, 0.278, 0.14] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find Wheelchair accessible parking for April 09 event in Barclays Center that allows entry after 10am\nPrevious actions:\n[textbox] Search for parking -> TYPE: Barclays Center\n[li] Barclays Center, Atlantic Avenue, Brooklyn, NY, US... -> CLICK\n[textbox] Select Event Date -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.374, 0.278, 0.404, 0.313] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the most popular tours and attractions in Kyoto, Japan from May 25 with free cancellation option, check availability for 2 adults and 1 infant, and book.\nPrevious actions:\n[span] Attractions -> CLICK\n[span] Asia -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.08, 0.273, 0.28, 0.431] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find all outdoor events this month in NYC\nPrevious actions:\n[span] Special events -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.332, 0.199, 0.471, 0.237] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add eleven produce items from Aldi, including 1 pack of blueberry, 6 bananas, 1 grape tomato, 1 roma tomato, 1 cilantro, to my cart and have it delivered tomorrow between 10am and 1pm to the default address.\nPrevious actions:\n[link] Herbs -> CLICK\n[div] Add -> CLICK\n[button] View Cart. Items in cart: 5 -> CLICK\n[button] View Cart. Items in cart: 5 -> CLICK\n[div] Go to checkout -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.113, 0.201, 0.113, 0.22] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Shop 2-5 day cruises to the Caribbean from Miami during August 2023.\nPrevious actions:\n[link] 2-5 DAY CRUISES Get big savings! Check out 2-5 day... -> CLICK\n[button] Sail To -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.042, 0.212, 0.185, 0.257] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a vehicle at New York JFK International Airport May 2-May 5 for an owner above 25 years.\nPrevious actions:\n[searchbox] Pick-up & Return Location (ZIP, City or Airport) (... -> TYPE: New York JFK\n[option] New York JFK International Airport JFK Jamaica, 11... -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.166, 0.227, 0.271, 0.247] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a treehouse in India with free cancellation and below $100 per night.\nPrevious actions:\n[img] -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.356, 0.052, 0.441, 0.083] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Locate the list of movies \"at home\", sorted by most recent.\nPrevious actions:\n[link] MOVIES -> CLICK\n[button] IN THEATERS -> CLICK\n[link] AT HOME -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.156, 0.2, 0.226, 0.215] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book a journey with return option on same day from Edinburg to Manchester on April 2nd and book the best possible option available.\nPrevious actions:\n[span] Manchester -> CLICK\n[radio] Return -> CLICK\n[textbox] Date use format: 29-Mar-23 -> CLICK\n[link] 2 -> CLICK\n[textbox] Date use format: -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.603, 0.283, 0.633, 0.304] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a flight From Dublin To Athens Greece for 1 Adult that leaves on April 1 and returns April 5th.\nPrevious actions:\n[div] Apr -> CLICK\n[generic] 2 -> CLICK\n[div] May 2023 -> CLICK\n[div] Apr -> CLICK\n[div] Choose date -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.134, 0.173, 0.161, 0.192] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find artist Mark Knight from London and follow him and add his most liked track to next up.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.321, 0.007, 0.561, 0.028] in the image\nAnd my action is Action: TYPE\nValue: Mark Knight" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find premium car in Brooklyn for a day.\nPrevious actions:\n[tab] Travel Info -> CLICK\n[link] Rent A Car -> CLICK\n[button] Pick-up -> TYPE: Brooklyn\n[button] Brooklyn New York, United States -> CLICK\n[button] Search -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.031, 0.347, 0.045, 0.357] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Reserve a taxi from O'hare Airport to 123 Main St, West Chicago, IL on Mar 29, 2023 at 12 noon for 2 adults.\nPrevious actions:\n[button] 29, March 2023 -> CLICK\n[button] pickup time input field -> CLICK\n[button] Confirm -> CLICK\n[combobox] Passengers -> SELECT: 2\n[button] Search -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.354, 0.323, 0.553, 0.518] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the lowest travel time flight for 1 traveler from JFK, NYC to Heathrow, London on April 19, then from Heathrow, London to CDG, Paris on April 21 and then from CDG, Paris to JFK, NYC on April 23\nPrevious actions:\n[textbox] From , required. -> TYPE: Heathrow\n[a] LHR - London Heathrow, United Kingdom -> CLICK\n[textbox] To , required. -> CLICK\n[textbox] To , required. -> TYPE: CDG\n[a] CDG - Paris Charles de Gaulle, France -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.509, 0.267, 0.693, 0.286] in the image\nAnd my action is Action: TYPE\nValue: 04/21/2023" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Request ticket receipt for ticket no. 1234567890123 and last name Davis\nPrevious actions:\n[link] Receipts and refunds -> CLICK\n[link] Receipts for tickets and fees -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.105, 0.376, 0.368, 0.419] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the budget price for the vehicle with most seats availible at Roanoke Regional Airport from April 5-6.\nPrevious actions:\n[textbox] Enter your pick-up location or zip code -> TYPE: Roanoke\n[span] Roanoke Regional Airport -> CLICK\n[button] Select My Car -> CLICK\n[link] Price (Low to High) -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.343, 0.147, 0.453, 0.158] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Browse listings for a used 2006 Honda Civic.\nPrevious actions:\n[link] Price New/Used -> CLICK\n[combobox] Year -> SELECT: 2006\n[combobox] Make -> SELECT: Honda\n[combobox] Model -> SELECT: Civic" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.625, 0.221, 0.71, 0.243] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add two $50 playstation store gift cards to cart\nPrevious actions:\n[button] Main Menu -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.0, 0.617, 0.291, 0.646] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find multi-colored fillable eggs for easter eggs made from polypropylene priced between 5 to 10 dollars for pick-up from Barboursville, zip 25504.\nPrevious actions:\n[button] Material -> CLICK\n[div] -> CLICK\n[button] close -> CLICK\n[heading] Pickup -> CLICK\n[button] Pick up Plastic Eggs in Egg Nesting Easter Eggs Mi... -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.72, 0.442, 0.988, 0.474] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Get the hotel with highest review score having free internet and free cancelation in Chennai for 20/03/23\nPrevious actions:\n[button] Monday March 20, 2023 -> CLICK\n[button] Search -> CLICK\n[svg] -> CLICK\n[tab] Review score -> CLICK\n[checkbox] Free internet -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.071, 0.217, 0.081, 0.226] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Check the movie \"65\" that is digital at the AMC Grove City 14 showing at 7:30 pm on Tuesday, March 28, 2023 and buy four tickets for seats J10 through J7\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.094, 0.051, 0.294, 0.1] in the image\nAnd my action is Action: SELECT\nValue: AMC Grove City 14" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Show the balance sheet and cash flow statement for the fiscal year 2021 of Six Flags.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.583, 0.022, 0.644, 0.036] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Buy a group ticket for 20 people in Six Flags Magic Mountain in Los Angeles on Apr 30, add 5 mega meal deals with it. Name the group Crew and check out as James Johnson, [email protected]. It is a family trip organized by Johnson.\nPrevious actions:\n[button] April 30, 2023 -> CLICK\n[button] Buy Now -> CLICK\n[generic] Increase -> CLICK\n[generic] Increase -> CLICK\n[generic] Increase -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.75, 0.05, 0.777, 0.067] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Request ticket receipt for ticket no. 1234567890123 and last name Davis\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.106, 0.817, 0.194, 0.831] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find Hot Dogs in Oakland, CA that Offers Delivery.\nPrevious actions:\n[textbox] Near -> CLICK\n[textbox] Near -> TYPE: Oakland, CA\n[span] Oakland, CA -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.105, 0.022, 0.268, 0.035] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Search for a flight from Dublin in Ireland to Paris Beauvais in france for 2 adults leaving apr 26 and returning may 2 with the promo code 10000001\nPrevious actions:\n[button] Ireland -> CLICK\n[span] Dublin -> CLICK\n[button] France -> CLICK\n[span] Paris Beauvais -> CLICK\n[generic] 26 -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.355, 0.348, 0.382, 0.373] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Search for Dining Room Sets under furniture and filter the results to show only items availble under the Buy It Now format.\nPrevious actions:\n[link] Home & Garden -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.013, 0.398, 0.173, 0.416] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Browse xbox series x consoles that are used and between $200 and $400.\nPrevious actions:\n[combobox] Search for anything -> TYPE: xbox series x console" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.786, 0.032, 0.917, 0.06] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add a dome tent to my shopping cart.\nPrevious actions:\n[link] Camping -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.062, 0.159, 0.206, 0.167] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book a flight + cruise for two flying from Atlanta in October for vacation starting on October 8 for 6 nights with Miami as the departure port, choose the cheapest flight, hotel, and room in the cruise for booking.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.478, 0.202, 0.586, 0.231] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: check flight status with flight number 2819 on 18th of March.\nPrevious actions:\n[tab] FLIGHT STATUS -> CLICK\n[button]  Search by date required selected as 15 March 202... -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.274, 0.221, 0.294, 0.236] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Schedule a virtual consultation for Solar Roof. The name is James Smith. The address is 123 st rd. The emial address is [email protected]\nPrevious actions:\n[link] Schedule a virtual consultation -> CLICK\n[textbox] First Name -> TYPE: James\n[div] -> TYPE: Smith\n[input] -> TYPE: 123st rd\n[textbox] Email Address -> TYPE: [email protected]" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.116, 0.725, 0.135, 0.739] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find value of 2016 Toyota Camry XLE Sedan 4D in black color with number plate AZXA46.\nPrevious actions:\n[div] XLE Sedan 4D -> CLICK\n[button] Next -> CLICK\n[div] Select Your Options -> CLICK\n[generic] Black image Black -> CLICK\n[button] Next -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.355, 0.465, 0.634, 0.539] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book the closest parking spot to the Orlando Magic at Brooklyn Nets event, happening in Barclays Center, to enter after 3 pm and exit after 5 pm\nPrevious actions:\n[textbox] Search for parking -> TYPE: barclays center" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.086, 0.123, 0.914, 0.131] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find Metformin 1000mg tablet price and details.\nPrevious actions:\n[link] See All -> CLICK\n[link] Amazon Health -> CLICK\n[link] Amazon Pharmacy -> CLICK\n[span] Type your medication name -> TYPE: Metformin 1000mg\n[li] metformin 1000mg tablet -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.236, 0.083, 0.378, 0.174] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add a CVS brand covid home test kit to the cart.\nPrevious actions:\n[img] -> CLICK\n[span] Shop all at-home COVID-19 tests -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.222, 0.564, 0.377, 0.601] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a ticket from Abbotsford to Sheboygan on March 26, 2023.\nPrevious actions:\n[button] Explore -> CLICK\n[link] Route map -> CLICK\n[textbox] Pick an origin Pick a destination -> TYPE: Abbotsford\n[span] Abbotsford, WI -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.078, 0.287, 0.389, 0.319] in the image\nAnd my action is Action: TYPE\nValue: Sheboygan" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the cheapest wheelchair accessible parking for the Pittsburgh Pirates at St. Louis Cardinals event in Busch Stadium on Apr 13.\nPrevious actions:\n[textbox] Search for parking -> TYPE: busch stadium\n[li] Busch Stadium, Clark Avenue, St. Louis, MO, USA -> CLICK\n[p] Pittsburgh Pirates at St. Louis Cardinals -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.319, 0.345, 0.384, 0.374] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Search a cheapest homestay with balcony in Bali, Indoneshia from April 1 to 4 for 2 adults and 4 children aged 3, 5, 8, and 12 years, respectively, double beds prefered, including breakfast.\nPrevious actions:\n[button] 4 -> CLICK\n[img] Add -> CLICK\n[img] Add -> CLICK\n[combobox] Age of Child 1(Child's age (years)) -> SELECT: 3\n[img] Add -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.343, 0.338, 0.389, 0.35] in the image\nAnd my action is Action: SELECT\nValue: 5" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the cheapest last minute cruise deal\nPrevious actions:\n[link] Last-Minute Deals -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.016, 0.28, 0.328, 0.447] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a cheapest standard train ticket from London to Sheffield on March, 29, arriving by 10:45 am for 2 adults.\nPrevious actions:\n[combobox] Arrival station, none selected. Select a station. -> TYPE: sheffield\n[span] Sheffield -> CLICK\n[textbox] Date use format: 24-Mar-23 -> CLICK\n[link] 29 -> CLICK\n[listbox] Leaving at or Arrive by selector -> SELECT: Leaving at" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.074, 0.291, 0.133, 0.316] in the image\nAnd my action is Action: SELECT\nValue: 10" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Make a reservation for 2 guests on april 21 at noon in a restaurant in Austin that is new on resy\nPrevious actions:\n[button] Sat., Apr. 22 -> CLICK\n[button] April 21, 2023. -> CLICK\n[combobox] Party of 2 -> SELECT: 2 Guests\n[link] Este -> CLICK\n[button] 12:00 PM Dining -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.023, 0.294, 0.523, 0.322] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find plastic bathroom baskets that are gray from low to high prices near zip code 60173.\nPrevious actions:\n[link] Products -> CLICK\n[button] Bathroom -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.355, 0.131, 0.605, 0.147] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add a pack of toilet paper and a bottle of laundry detergent to your Amazon cart with the lowest total price.\nPrevious actions:\n[span] Featured -> CLICK\n[option] Price: Low to High -> CLICK\n[button] Add to Cart -> CLICK\n[textbox] Search Amazon -> TYPE: laundry detergent\n[button] Go -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.832, 0.047, 0.866, 0.051] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the gift registry of Carla Cahill, when asked for password use Michael Cahill as name and [email protected] as email.\nPrevious actions:\n[textbox] Last Name -> TYPE: Cahill\n[link] SEARCH -> CLICK\n[span] View -> CLICK\n[link] I don't have the password -> CLICK\n[textbox] Your Name -> TYPE: Michael Cahill" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.329, 0.336, 0.671, 0.37] in the image\nAnd my action is Action: TYPE\nValue: [email protected]" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Plan a trip from Braintree to Boston Airport on April 14, 12 pm, and see the fair calculator.\nPrevious actions:\n[tab] Trip Planner -> CLICK\n[combobox] From -> TYPE: brain\n[link] T red line bus commuter rail Zone 2 Braintree -> CLICK\n[combobox] To  -> TYPE: boston" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.536, 0.217, 0.812, 0.25] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Search a cheapest homestay with balcony in Bali, Indoneshia from April 1 to 4 for 2 adults and 4 children aged 3, 5, 8, and 12 years, respectively, double beds prefered, including breakfast.\nPrevious actions:\n[gridcell] Sat Apr 01 2023 -> CLICK\n[button] 4 -> CLICK\n[img] Add -> CLICK\n[img] Add -> CLICK\n[combobox] Age of Child 1(Child's age (years)) -> SELECT: 3" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.382, 0.273, 0.393, 0.279] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a flight from Mumbai to New York City on 21st April, 2023.\nPrevious actions:\n[link] NYC New York City Area Airports, NY -> CLICK\n[combobox] Trip Type:, changes will reload the page -> CLICK\n[option] One Way -> CLICK\n[button]  Depart and Return Calendar Use enter to open, es... -> CLICK\n[link] 21 April 2023, Friday -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.798, 0.196, 0.858, 0.22] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book a best business class flight ticket from Colombo to New York with a two-day Qatar stopover for 2 adults and 1 child on June 24.\nPrevious actions:\n[textbox] To 2180 results are available, use up and down arr... -> TYPE: new york\n[strong] New York -> CLICK\n[textbox] Depart date please enter date in the format dd spa... -> CLICK\n[gridcell] 24 June 2023 -> CLICK\n[textbox] Passengers / Class -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.82, 0.488, 0.858, 0.517] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find popular deals to USA national parks in the southeast region of Biscayne national park, and add to the cart the fully refundable economy flight from New York to Miami with a departure date of May 13 and a return date of May 18.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.855, 0.048, 0.9, 0.059] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Search for Limited Time Offers available in the Kids section and filter by size, 11-12Y(150).\nPrevious actions:\n[link] KIDS -> CLICK\n[link] Limited-Time Offers -> CLICK\n[gridcell] Size -> CLICK\n[label] 11-12Y(150) -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.454, 0.316, 0.473, 0.329] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Create a playlist and name it \"Dirty\" and add 5 Doja Cat songs.\nPrevious actions:\n[img] james9091 -> CLICK\n[link] Playlists -> CLICK\n[link] NEW PLAYLIST -> CLICK\n[button] START FROM SCRATCH -> CLICK\n[h1] My New Playlist -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.351, 0.253, 0.613, 0.3] in the image\nAnd my action is Action: TYPE\nValue: Dirty" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the newest movies for rent in the Redbox kiosk at any Winn Dixie in Daytona Beach.\nPrevious actions:\n[button] Navigate to Find a kiosk nearby -> CLICK\n[textbox] City or ZIP Code -> TYPE: Daytona" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.508, 0.234, 0.709, 0.279] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find rail for Oyster Bay, Provincetown.\nPrevious actions:\n[link] commuter rail Commuter Rail -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.13, 0.189, 0.395, 0.213] in the image\nAnd my action is Action: TYPE\nValue: Oyster Bay" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find hotel deals in Las Vegas for four adults starting on May 17 and ending on May 20, and if deal is not available, set an alert for the same.\nPrevious actions:\n[div] Las Vegas, NV -> CLICK\n[textbox] When? -> CLICK\n[link] 17 -> CLICK\n[link] 20 -> CLICK\n[combobox] How many guests? -> SELECT: 4 Guests" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.918, 0.057, 0.964, 0.082] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the cheapest 2017 Honda Civic within 100 miles of 07470 which has a sun roof and black exterior color while excluding out of market vehicles\nPrevious actions:\n[button] Back to all categories -> CLICK\n[button] Features -> CLICK\n[listitem] Sunroof(s) (2) Sunroof(s) (2) -> CLICK\n[button] Back to all categories -> CLICK\n[button] Exterior Color -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.006, 0.27, 0.045, 0.292] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: search the most recent full time Accounting & Finance Job available in Richmond, Virginia.\nPrevious actions:\n[button] Open helpful links menu -> CLICK\n[link] Search jobs at CarMax -> CLICK\n[textbox] Search for Job title -> TYPE: Accounting\n[link] Accounting & Finance 7 jobs -> CLICK\n[button] State -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.04, 0.282, 0.123, 0.292] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the US box office revenue for the highest tomatometer rated movie that the actress playing Sam Carpenter in the most recent Scream movie has been in.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.239, 0.072, 0.525, 0.091] in the image\nAnd my action is Action: TYPE\nValue: scream" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add documentary \"The Elephant Whisperers\" and \"The Magician's Elephant\" to the watchlist.\nPrevious actions:\nNone" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.0, 0.22, 0.156, 0.236] in the image\nAnd my action is Action: TYPE\nValue: The Elephant Whisperers" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a Ford Mustang with lowest price and save it.\nPrevious actions:\n[link] Cars for Sale -> HOVER\n[link] Cars for Sale -> CLICK\n[textbox] Search by Make, Model, Body Style or Keyword -> TYPE: Mustang\n[div] Ford -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.775, 0.124, 0.888, 0.141] in the image\nAnd my action is Action: SELECT\nValue: Price - Lowest" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Show latest documentaries on Netflix and add the one with 70+ metascore.\nPrevious actions:\n[link] What to Watch on Netflix -> CLICK\n[button] Filter -> CLICK\n[button] Documentary -> CLICK\n[button] Release Date -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.36, 0.131, 0.638, 0.157] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book a SUV rental from Mar 31 (noon) to April 7, 2023 (noon) from Chicago O'hare Intl Airport, ORD and returning to same place with the lowest pay now price.\nPrevious actions:\n[textbox] Search by Airport, City, Zip, Address or Attractio... -> TYPE: Ohare, Chicago\n[span] , United States -> CLICK\n[span] , Chicago -> CLICK\n[link] Make a Reservation -> CLICK\n[button] Select My Car -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.134, 0.145, 0.238, 0.156] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a comedy movie streaming on Netflix and add to watchlist.\nPrevious actions:\n[link] MOVIES -> HOVER" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.253, 0.216, 0.424, 0.231] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add two pairs of women's Pink sweet pants in XXL size to your cart and view your cart.\nPrevious actions:\n[link] WOMEN -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.041, 0.106, 0.257, 0.126] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the most popular guitar tab for Absolute Beginners.\nPrevious actions:\n[link] Tabs -> CLICK\n[svg] -> CLICK\n[link] Most popular of all time -> CLICK\n[link] Absolute Beginner 91,138 -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.38, 0.237, 0.523, 0.246] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Make a playlist and add 3 songs from popular Selena Gomez tracks and name it Love.\nPrevious actions:\n[searchbox] Search -> TYPE: Selena Gomez\n[button] Search -> CLICK\n[link] Selena Gomez -> CLICK\n[link] Popular tracks -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.413, 0.308, 0.462, 0.322] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: search new books from Jk Rowling available for kids within the age from 3 to 5 that are below $20\nPrevious actions:\n[textbox] Search for books by keyword / title / author / ISB... -> TYPE: Jk rowling\n[link] Jk rowling -> CLICK\n[select] Most popular -> SELECT: Publication date, new to old" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.027, 0.174, 0.196, 0.204] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book a roundtrip on July 1 from Mumbai to London and vice versa on July 5 for two adults and a 12-year-old in premium economy and if a flight is not available, search through the calendar for available flights.\nPrevious actions:\n[combobox] Enter your departing city, airport name, or airpor... -> TYPE: MUMBAI\n[button] Mumbai, IN (BOM) -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.298, 0.187, 0.481, 0.214] in the image\nAnd my action is Action: TYPE\nValue: LONDON" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Show car guide info for economy car and reserve.\nPrevious actions:\n[button] Cars & Services -> CLICK\n[link] Car Guides -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.394, 0.373, 0.469, 0.387] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Filter women's sports bras that are purple with high support that are small size.\nPrevious actions:\n[RootWebArea] Women's Sports Bras | Under Armour -> CLICK\n[button] Show more -> CLICK\n[div] Color -> CLICK\n[link] Refine by color: Purple -> CLICK\n[div] Support Level -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.044, 0.351, 0.194, 0.366] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find Hair Loss Evaluation and Treatment in 10018 and show directions.\nPrevious actions:\n[button] Health -> CLICK\n[button] load Health Services Menu -> CLICK\n[link] Skin, Hair & Nails -> CLICK\n[button] Hair Loss Evaluation & Treatment -> CLICK\n[link] Hair Loss Evaluation & Treatment -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.04, 0.343, 0.34, 0.371] in the image\nAnd my action is Action: TYPE\nValue: 10018" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Add to my wishlist the top rated JRPG game.\nPrevious actions:\n[link] Categories -> CLICK\n[link] JRPG -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.365, 0.462, 0.432, 0.474] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find the longest skiing activity in August 10\nPrevious actions:\n[textbox] Where to? -> TYPE: skiing\n[button] When -> CLICK\n[button] Next -> CLICK\n[path] -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.764, 0.433, 0.777, 0.446] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find trip for 123456 last name Smith.\nPrevious actions:\n[button] MY TRIP -> CLICK\n[textbox] Reservation/Ticket Number -> TYPE: 123456\n[div] -> CLICK\n[option] Last Name -> CLICK\n[textbox] Last Name -> TYPE: Smith" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.512, 0.333, 0.623, 0.366] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Search for rental cars in Seattle, WA and filter the results by the Best score.\nPrevious actions:\n[link] Search for cars -> CLICK\n[textbox] Pick-up location -> CLICK\n[textbox] Pick-up location -> TYPE: Seattle\n[span] Seattle, Washington, United States -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.945, 0.085, 0.964, 0.094] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a car to book with miles to pick up on Caldwell (Ohio) on april 8 at 8 am and return on april 13 on noon at the same location\nPrevious actions:\n[heading] CAR -> CLICK\n[label] Book with miles -> CLICK\n[LabelText] Enter pick up city, airport name, or airport code. -> TYPE: caldwell" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.09, 0.282, 0.329, 0.294] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Rent a 12-inch available moving truck with a complete protection package for approx 250 miles, pick it up at 4:00 pm, and drop it off at the same location on June 20 at 4:30 pm from zip 07718.\nPrevious actions:\n[input] -> CLICK\n[button] Find Your Truck -> CLICK\n[button] Select Truck -> CLICK\n[input] -> TYPE: 250\n[button] Continue to Location -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.739, 0.153, 0.915, 0.168] in the image\nAnd my action is Action: SELECT\nValue: 04 00 PM" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Reserve a train ticket with seat from Berlin to Zurich on March 31 after 8 am in aisle quite zone in side by side seat formation.\nPrevious actions:\n[textbox] Date use format: 25-Mar-23 -> CLICK\n[link] 31 -> CLICK\n[listbox] hour -> SELECT: 08\n[listbox] minutes -> SELECT: 00\n[button] Get cheapest tickets -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.562, 0.509, 0.886, 0.583] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Browse movie trailers currently in theaters with \"fresh\" rating.\nPrevious actions:\n[link] MOVIES -> HOVER\n[link] Trailers -> CLICK\n[link] VIEW ALL -> CLICK\n[span] Tomatometer -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.781, 0.318, 0.802, 0.333] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a cheapest flight from London to New York on 9th May.\nPrevious actions:\n[tab] Flights -> CLICK\n[combobox] Flying from -> TYPE: London\n[option] Destination London -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.548, 0.259, 0.764, 0.273] in the image\nAnd my action is Action: TYPE\nValue: New York" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a black sleeping pad that is under $40.\nPrevious actions:\n[textbox] Search -> TYPE: black sleeping bag\n[button] Search -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.019, 0.236, 0.244, 0.252] in the image\nAnd my action is Action: TYPE\nValue: 40" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a car to book with miles to pick up on Caldwell (Ohio) on april 8 at 8 am and return on april 13 on noon at the same location\nPrevious actions:\n[heading] CAR -> CLICK\n[label] Book with miles -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.075, 0.487, 0.165, 0.493] in the image\nAnd my action is Action: TYPE\nValue: caldwell" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: find offer for a black Honda with vin number 1HGCM66543A064159 with 155000 millage near 49102. The car should be in good condition with no damage, 2+ keys and paid off.\nPrevious actions:\n[button] No -> CLICK\n[button] No -> CLICK\n[button] 2+ -> CLICK\n[radio] Owned -> CLICK\n[button] Next -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.012, 0.465, 0.111, 0.494] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Find a park in Alaska that is also a National Heritage Area.\nPrevious actions:\n[button] Find a Park by State -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.212, 0.371, 0.788, 0.392] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Confirm my vip tour at the six flags Discovery Kingdom\nPrevious actions:\n[button] Browse the Parks Below -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.527, 0.61, 0.846, 0.639] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: Book a table for 2 at a restaurant that was featured on the latest Hit List\nPrevious actions:\n[link] The Hit List -> CLICK\n[a] -> CLICK\n[link] Book Now -> CLICK" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.192, 0.335, 0.354, 0.361] in the image\nAnd my action is Action: CLICK" } ]

[ { "from": "human", "value": "<image>\nImagine that you are imitating humans doing web navigation for a task step by step. At each stage, you can see the webpage like humans by a screenshot and know the previous actions before the current step decided by yourself through recorded history. You need to decide on the first following action to take. You can click an element with the mouse, select an option, or type text with the keyboard. (For your understanding, they are like the click(), select_option() and type() functions in playwright respectively) One next step means one operation within the three.\nPlease pinpoint the bounding box coordinates of the element’s region that I should click in order to complete the following instruction: find the score of the latest nfl game played by the las vegas raiders.\nPrevious actions:\n[combobox] Search query -> TYPE: las vegas raiders" }, { "from": "gpt", "value": "For my next step, the target element of my action is bounded by the box [0.234, 0.079, 0.259, 0.093] in the image\nAnd my action is Action: CLICK" } ]